An investigation into the influence of stent strut thickness on in-stent restenosis using the finite element method

by

Fariza Fida Basir, BEng

A thesis submitted to Dublin City University in partial fulfilment of the requirement for the degree of

Masters in Engineering

Supervisors Dr. Caitriona Lally, BEng, MEng, PhD Professor M.S.J. Hashmi, Phd, DSc

School of Mechanical and Manufacturing Engineering Dublin City University

November 2005

DECLARATION
I hereby certify that this material, which I now subm it for assessm ent on the program m e o f study leading to the award o f M asters in Engineering, is entirely my own work and has not been taken from the w ork o f others save and to the extent that such work has been cited and acknow ledged w ithin the text o f my work.

JD N o: Fariza Fida Basir

97194735

Date:

8lh N ovem ber 2005

II

ACKNOWLEDGEMENTS
I w ould like to take this opportunity to thank m y supervisors, Dr. C aitriona Lally and Prof. M .S.J. H ashm i, for their help and guidance throughout the research study. I w ould like to extend m y gratitude especially tow ards Dr. C aitriona Lally who w ent the extra m ile and past her responsibilities to ensure that I achieve a good standard in m y w ork; I thank her, for her endless support, guidance and encouragem ent throughout the study. I am forever grateful. I w ould like to thank m y technician, K eith H ickey for his assistance in

accom m odating the technical resources the project needed. A nd to m y friends; m y best friend Jaycey, for h er constant support and for having faith in me. Shubom a w ho shared m any things w ith me, A neela, m y gym friend when studying got so overwhelm ing. To D arragh, M ary and Barbara, I

thank them for their friendship over the years. To the lads in the lab, John, D eclan and D ave, for m aking m y life in college a good one w ith their little stories and humour. To A ngah and Amy, I w ould like to extend m y appreciation for their kindness and generosity. A nd to Jam eel, I w ould like to thank him for listening patiently to the pains and laughters, let it be study or personal throughout the years. M ost o f all, m y love and thank to the m ost im portant people in m y life, to m y m other who has done so much, from cooking m y favourite m eal to supporting m e financially and em otionally. I thank her for having constantly praying for m y success and w ell-being. To m y sister and also m y best friend Farrah, who has It w ould not be possible to return her endless

never failed to be there for me.

support and sacrifices. Let us look forw ard to the new beginning that lies ahead.

Ill

An investigation into the influence of stent strut thickness on in-stent restenosis using the finite element method
Fariza Fida Basir A bstract In-stent restenosis represents the m ajor lim itation for stenting procedures. In-stent restenosis is the renarrow ing o f the artery lum en w ithin a stent predom inantly due to excessive growth o f neointim al hyperplasia. Clinical studies have found that stent design is a key determ inant in the propensity o f stents to cause restenosis, indicating a vital link between the biom echanics o f stents and the developm ent o f the disease. The ISAR-STEREO Trial specifically assessed the effect o f strut thickness on restenosis outcom e and found that for the sam e stent design, a thinner strut stent was associated w ith a significant reduction o f angiographic and clinical restenosis com pared to the sam e stent w ith a thicker strut. The main objective o f this study is to use the finite elem ent m ethod to simulate these stenting procedures, and to exam ine the stresses induced w ithin the stented arterial vessel w alls by the stents, thus enabling the m echanical stim uli for in-stent restenosis to be identified. Finite elem ent m odels o f thin and thick strut stents w ere developed and the stents w ere deployed in various stenosed vessel geom etries such that the stresses induced w ithin the stented vessels by the two stents could be compared. The stresses w ere exam ined at the end o f stent deploym ent, to determ ine the m echanical stim uli for acute dam age, and again at stent unloading, to determ ine the long term stim uli for in-stent restenosis. The stress analyses w ere used to determ ine the level o f vascular injury caused to the artery by different strut thickness stents. The finite elem ent studies successfully identified differences between the m echanical loading o f the arterial tissue in the vessels stented with the two different stents. The higher restenosis rate o f the thicker strut stent, reported in the ISA R -STEREO clinical study, was found to be the result o f the higher luminal gain achieved by the thicker strut stent, due to the low er recoil o f the stent structure w hen both stents were expanded to the same initial lum en diameter. Further stenting analyses, however, found that the thicker strut stent resulted in a low er percentage o f volum e stressed at high levels com pared w ith the thinner strut stent w hen it was expanded to the same final lum en diameter. This suggests that a thicker strut stent m ay in fact have the potential to be expanded to an optimal diam eter w hereby the in-stent restenosis is m inim ised. Therefore, it is proposed that the use o f preclinical testing tools, such as finite elem ent m odelling, could be used to predeterm ine the deploym ent protocol and optim um lum inal gain o f a particular stent design in order to m inim ise the m echanical stim uli for in-stent restenosis.

IV

TABLE OF CONTENTS

D E C L A R A T IO N

II

AC K N O W LE D G E M EN T S

HI

A BSTRA C T

IV

N O M EN C L A T U R E

IX

LIST O F FIGURES

L IST O F TA BLES

XVI

INTR O D U CTIO N

1.1
1.2 1.3 1.4 1.5 1.6

Cardiovascular Disease Coronary H eart Disease Percutaneous Translum inal Coronary A ngioplasty Intravascular Stenting Clinical and Finite Element Studies o f Stents O bjectives o f the Study

1
2 5 6 8 9

LITER ATU R E REVIEW 2.1 2.2 2.3 2.4 Introduction Structure o f the Arterial Wall Constitutive M odels o f the Arterial Wall M echanism o i Restenosis 2.4.1 2.4.2 2.4.3 2.4.4 Throm bosis Inflam mation Neointim al Proliferation V ascular Rem odelling

11 11 12 13 20 22 23 23 24

2.5

Clinical Trials on Intravascular Stents E xam ining Restenosis O utcom e 2.5.1 2.5.2 2.5.3 Influence o f Stent D esign on R estenosis O utcom e Influence o f Stent M aterial on R estenosis O utcom e Influence o f Strut Thickness on R estenosis O utcom e 25 25 30 31 32 37

2.6 2.7

Finite Elem ent Studies o f Stenting Procedures Conclusion

M A T E R IA L S AND M E T H O D 3.1 3.2 Introduction M odelling o f Stent G eom etry 3.2.1 Case Study 1: Free Expansion o f M linkthin and M linkthiC k 3.2.2 Case Study 2 and Case Study 3: Expansion o f M1inkthjn and M linkthiC k Inside V arious Stenosed Coronary A rtery V essels 3.2.3 Case Study 4: Expansion o f Mlinkthin and M linkthjC k Inside Localised Stenotic C oronary A rtery V essels 3.3 3.4 3.5 3.6 Stent M aterial: 316L Stainless Steel A rtery M aterial: Porcine Coronary A rtery Plaque M aterial: H ypocellular H um an A therosclerotic Plaques Sum m ary

39 39 40

47

49

52 54 56 58 59

RESU LTS 4.1 4.2 Introduction Case Study 1: Free E xpansion o f Mlinkthin and M linkthiC k 4.3 Case Study 2:

60 60

60

Expansion o f Mlinkthin and M linkthiC k Inside a Stenotic C oronary A rtery A chieving the Same Initial Expanded Lum en D iam eter 4.4 Case Study 3: Expansion o f Mlinkthin and M linkthiC k Inside a Stenotic C oronary A rtery A chieving the Sam e Final Expanded Lum en D iam eter 84 67

VI

4.5

Case Study 4: Expansion o f M linkthm and M linkthiC k Inside a L ocalised Stenotic Coronary A rtery A chieving the Sam e Initial E xpanded Lum en D iam eter 101 117

4.6

Sum m ary

D IS C U S S IO N 5.1 5.2 5.3 Introduction Lim itations o f this Study C om parison o f Pressure D eploym ent 5.3.1 5.3.2 C om parison o f Thinner Strut Stent vs. Thicker Strut Stent Com parison o f E ach Stent D esign U nder Free E xpansion and U nder the Influence o f V arious Stenosed V essels 5.4

118 118 121 121 122

122

C om parison o f the Mlinkthm and Mlinkthick Stents A fter D eploym ent 123 5.4.1 5.4.2 5.4.3 von M ises Stress and Plastic Strain Foreshortening D ogboning 123 125 126 128 129

5.5 5.6

R adial R ecoil Stresses Induced W ithin the Stented V essels 5.6.1 Stress Analysis: Case Study 2 Expansion o f Mlinkthm and M linkthiC k Inside a Stenotic A rtery V essel G eom etry w ith A rterial M aterial Properties, A chieving the Same Initial E xpanded Lum en D iam eter, 0 i thm = 0 i thick 5.6.2 Stress Analysis: Case Study 3 Expansion o f Mlinkthm and Mlinkthick Inside a Stenotic A rtery V essel G eom etry w ith A rterial M aterial Properties as in Case

130

Study 2, H ow ever, A chieving the Sam e Final Lum en D iam eter,

0 F thin = 0 F thick 133

5.6.3

Stress Analysis: Case Study 4 Expansion o f M linkthm and Mlinkthick Inside a L ocalised Stenotic Artery, A chieving the Sam e Initial Expanded Lum en D iam eter,
0 i thin 0 1 thick 135

5.7

Preclinical Testing o f Stents and Stent D esign

137

V II

C O N CLU SIO N S 6.1 6.2 M ain Findings Future W ork

139 139 141

REFERENCES

142

A PPE N D IX A: SIM U L A TIO N O F A FU LL M O D E L O F M L IN K thick IN SID E A STEN O TIC STR AIG H T V ESSEL A .l A .2 A. 3 A .4 Introduction M aterials and M ethod R esults and D iscussion Conclusions 152 152 152 154 156

V IE

NOMENCLATURE
Roman Letters

aij E L R S W

Material constants G reen-Lagrangian strain tensor Length Radius Com ponent o f the second Piola-K irchoff stress tensor Strain energy

G reek Letters

Ii, I2, I3 ^ 1, ^ 2 ,
^3

Stretch invariants Principal stretches Diam eter von M ises stress Principal stresses Plastic strain rate tensor Equivalent plastic strain

0
a
a i , a 2, a 3

IX

LIST OF FIGURES
1.1
1.2
1.3 1.4 1.5 G raph o f prem ature death rates from cardiovascular disease across EU countries, [3]. 2 Illustration o f artherosclerosis in the coronary arteries [adapted from M edline], [7]. 3 G raph o f the frequency o f the use o f coronary stenting and coronary artery bypass grafts in 1998 and 1999, [9]. 4 Illustration o f PTC A procedure [adapted from H eart Centre], [4]. 5 Pictorial representation o f a flouroscopic im age o f the heart [adapted from Adam ], [13]. 6 Illustration o f stenting procedure [adapted from H eart Centre], [4]. 7 Illustration o f three layers m ade up an artery [adapted from [34]]. 12 Cycle stress response o f a d o g s carotid artery under preconditioning showing hysteresis loop over a num ber o f cycles [adapted from [47]]. 15 The four phases o f vascular repair after stenting, leading to the form ation o f in-stent restenosis [14], 22 A n illustration o f response to injury from stent strut expansion in norm al porcine coronary artery after 28 days. L=lum en; N =neointim a; IEL/EEL =intem al/extem al elastic lam ina [adapted from Lowe et al., [65]. 24 Illustration o f selected stent designs: (A) self-expanding stent, W allstent (Schneider), (B) coil stent, G ianturco-R oubin stent (Cook), (C) slotted tube stent, Palm az-Schatz stent (Johnson & Johnson), (D) closed cell structure, N IR stent (M edinol/Boston Scientific), (E) open cell structure, M ultiLink stent (Guidant), (F) BX stent (Cordis) and (G) Radius N itinol (Scim ed), [adapted from [10, 73]]. 29 Representative M ovat pentachrom e-stained sections from the four stent designs: (A) BX stent, (B) M ultilink, (C) R adius N itinol, and (D) Palm azSchatz, [adapted from Taylor et al., [81]]. 29 Expanded ACS M ultiLink R X D uet [adapted from [90], 42 Schem atic illustrating the radial dim ensions o f the M linkthin and M linkthkk models. 42 V olum es o f (a) Mlinktilin and (b) M linkthiC k in the Cartesian coordinate system and the m eshed m odel o f (c) Mlinkulin and (d) M linkthiC k43 Schem atic illustrating the stents expansion at (A) stent before expansion, (B) stent at loading w hen pressure is applied and (C) stent at unloading w hen pressure is rem oved and (D) illustrates the division o f interest. 46 Finite elem ent m eshes o f unexpanded Mlinkthin and Mlinkthick, used for case study 1. 48 Plot o f pressure path o f M linkthjn and Mlinkthick48 Finite elem ent m esh o f unexpanded M linkthm and M linkthiC k inside a stenotic coronary artery, w ith an inner radius o f 1 m m and outer radius o f 1.8 mm, used for case study 2 and case study 3. 51 Illustration o f discontinuities created as local m esh adaptivity occurs. 51 Finite elem ent m esh o f unexpanded M linkthm and M linkthiC k inside a localised stenotic vessel, w ith an inner radius o f 1 m m and outer radius o f 1.8, used for case study 4. 52 Plot o f stress-strain relationship o f 316L stainless steel for M linkthin (0.05 m m ) and M linkthick (0.14 mm ), [adapted from M urphy et al., [88]]. 55

1.6 2.1 2.2

2.3 2.4

2.5

2.6

3.1 3.2 3.3 3.4

3.5 3.6 3.7

3.8 3.9

3.10

3.11 3.12 4.1 4.2

4.3 4.4

4.5 4.6 4.7 4.8

4.9

4.10 4.11

4.12

4.13

4.14

4.15

4.16

4.17

Plot o f stress-strain uniaxial and biaxial data for porcine coronary tissue [41]. 57 Plot o f stress-strain uniaxial and biaxial data for hypocellular hum an atherosclerotic plaque. 59 D eform ed geom etry o f M linkthin and M linkthiC k under free expansion, achieving an initial outer diam eter o f 3 mm. 61 Radial displacem ent distribution throughout the stent structure o f M linkthin and Mlinkthick, subjected to free expansion, achieving an initial outer diam eter o f 3 mm. 62 The proxim al, central and distal points o f interest. 62 Radial displacem ent o f Mlinkthin and Mlinkthick through loading and unloading, subjected to free expansion, achieving an initial outer diam eter o f 3 mm. 64 Com parison o f percentage radial recoil o f Mlinkthin and Mlinkthick, subjected to free expansion, achieving an initial outer diam eter o f 3 m m . 64 The resulting von M ises stress contours throughout the structures o f Mlinkthin and Mlinkthick, subjected to free expansion. 66 The resulting Total Equivalent Plastic strain contours throughout the structures o f Mlinkthin and Mlinkthick, subjected to free expansion. 66 D eform ed geom etry o f the proxim al and distal halves o f Mlinkthin and Mlinkthick, scaffolding a stenotic vessel w ith thickness o f 0.8 mm, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm. 68 Radial displacem ent distribution throughout the stents structure o f Mlinkthin and Mlinkthick, under the influence o f a stenotic vessel, achieving the same initial vessel lum en diam eter o f 0 / = 3 m m . 69 The proxim al, central and distal points o f interest. 69 Radial displacem ent o f Mlinkthin and Mlinkthick through loading and unloading, under the influence o f stenotic vessel, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm. 71 C om parison o f percentage radial recoil o f Mlinkthin and Mlinkthick, subjected to free expansion, achieving the sam e initial vessel lum en diam eter o f 0 / = 3 m m . 71 The resulting von M ises Stress contours throughout the structure o f Mlinkthin and Mlinkthick, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the sam e initial vessel lum en diam eter o f 0 / 3 mm. 72 The resulting Total Equivalent Plastic Strain contours throughout the structure o f Mlinkthin and Mlinkthick, under the influence o f stenotic vessel o f 0.8 mm thickness, achieving the sam e initial vessel lum en diam eter o f 0 / = 3 mm. 73 Tensile circum ferential stresses induced in the arterial w all stented b y M linkthin and Mlinkthick, at loading, achieving th e sam e initial vessel lum en diam eter o f 0 / = 3 mm. 75 Com pressive radial stresses induced in the arterial w all stented by Mlinkthin and M linkthiC k, at loading, achieving the sam e initial vessel lum en diam eter o f 0 / = 3 mm. 76 Tensile and com pressive longitudinal stresses induced in the arterial w all stented by M linkthin and Mlinkthick, at loading, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm. 77

XI

4.18

4.19

4.20

4.21

4.22

4.23

4.24

4.25

4.26

4.27

4.28

4.29

4.30

4.31

4.32

4.33

The percentage stress volum es b y the tensile circum ferential stress (A) and radial com pressive stress (B), within the stented vessel b y M linkthm and Mlinkthick,, at loading, achieving the sam e initial vessel lum en diam eter o f 0/=3m m . 78 The percentage stress volum es b y the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel b y Mlinkthm and Mlinkthick, at loading, achieving the same initial vessel lum en diameter o f 0 / = 3 mm. 78 Tensile circum ferential stresses induced in th e arterial w all stented by Mlinkthin and Mlinkthick, at unloading, resulting in different final lum en diameter. 80 C om pressive radial stresses induced in the arterial w all stented by Mlinkthin and Mlinkthick, at unloading, resulting in different final lum en diameter. 81 Tensile and com pressive longitudinal stresses induced in the arterial w all stented b y M linkthin and Mlinkthick, at unloading, resulting in different final lum en diameter. 82 The percentage stress volum es b y the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel b y Mlinkthm and Mlinkthick, at unloading, resulting in different final lum en diam eter. 83 The percentage stress volum es b y the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel by M linkthm and Mlinkthick, at loading, resulting in different final lum en diameter. 83 D eform ed geom etry o f proxim al and distal h a lf o f Mlinkthin and Mlinkthick, scaffolding a stenotic vessel, thickness o f 0.8 m m , achieving the same final vessel lum en diam eter o f 0 f = 2.28 mm. 85 Radial displacem ent distribution throughout the stents structure o f Mlinkthin and Mlinkthick, under the influence o f a stenotic vessel, achieving the sam e final vessel lum en diameter. 86 R adial displacem ent o f Mlinkthin and Mlinkthick through (a) loading and (b) unloading, under the influence o f stenotic vessel, achieving the same final vessel lum en diameter. 86 C om parison o f percentage radial recoil o f Mlinkthin and Mlinkthick, subjected to free expansion, achieving the sam e final vessel lum en diameter. 88 The resulting von M ises Stress contours throughout the structure o f M linkthin and Mlinkthick, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the sam e final vessel lum en diam eter. 89 The resulting Total Equivalent Plastic Strain contours throughout the structure o f M linkthin and Mlinkthick, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the sam e final vessel lum en diameter. 90 Tensile circum ferential stresses induced in the arterial w all stented by Mlinkthin and M linkthick, at loading, achieving different initial vessel lum en diameter. 92 Com pressive radial stresses induced in the arterial w all stented by Mlinkthm and M linkthiC k, at loading, achieving different initial vessel lum en diameter. 93 Tensile and com pressive longitudinal stresses induced in the arterial wall stented b y Mlinkthin and Mlinkthick, at loading, achieving different initial vessel lum en diameter. 94

X II

4.34

4.35

4.36

4.37

4.38

4.39

4.40

4.41

4.42

4.43

4.44

4.45

4.46

The percentage stress volum es by the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel b y M linkthin and Mlinkthick, at loading, achieving different initial vessel lum en diameter. 95 The percentage stress volum es by the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel b y M linkthin and Mlinkthick, at loading, achieving different initial vessel lum en diameter. 95 Tensile circum ferential stresses induced in the arterial w all stented b y Mlinkthin and Mlinkthick, at unloading, achieving the sam e final vessel lum en diameter. 97 Com pressive radial stresses induced in the arterial w all stented b y Mlinkthin and Mlinkthick, at unloading, achieving the sam e final vessel lum en diameter. 98 Tensile and com pressive longitudinal stresses induced in the arterial w all stented b y Mlinkthin and Mlinkthick, at unloading, achieving the sam e final vessel lum en diameter. 99 The percentage stress volum es b y the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel b y Mlinkthin and Mlinkthick, at unloading, achieving the same final vessel lum en diameter. 100 The percentage stress volum es by the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel b y Mlinkthin and Mlinkthick, at unloading, achieving the sam e final vessel lum en diameter. 100 D eform ed geom etry o f distal halves o f Mlinkthin and Mlinkthick, scaffolding a stenotic vessel, thickness o f 0.8 mm, achieving the sam e initial vessel lum en diam eter o f 0 / = 3.18 mm . 101 Radial displacem ent distribution throughout the stents structure o f Mlinkthin and Mlinkthick, under the influence o f a localised stenotic vessel, thickness o f 0.8 mm, achieving the sam e initial vessel lum en diameter. 102 C om parison o f radial displacem ent o f Mlinkthin and Mlinkthick inside a localised stenotic vessel throughout the structure at (a) loading and (b) unloading, achieving an initial lum en diam eter, 0 / o f 3.18 m m and a final lum en diam eter, 0 F o f 2.44 m m for M linkthm and 0 F o f 2.52 m m for Mlinkthick105 Com parison o f percentage recoil o f M linkthin and Mlinkthick inside a localised stenotic vessel, achieving an initial lum en diam eter, 0 / o f 3.18 mm. 105 The resulting von M ises stress contours throughout the structure o f Mlinkthin and Mlinkthick, under the influence o f a localised stenotic vessel, achieving an initial lum en diam eter, 0 / o f 3.18 m m and a final lumen diam eter, 0 F o f 2.44 m m for Mlinkthin and 0 F o f 2.52 m m for Mlinkthick106 The resulting Total Equivalent Plastic strain contours throughout the structure o f Mlinkthin and M linkthiC k, under the influence o f a localised stenotic vessel, achieving an initial lum en diam eter, 0 / o f 3.18 m m and a final lum en diameter, 0 F o f 2.44 m m for Mlinkthin and 0 F o f 2.52 m m for Mlinkthick107

XIII

4.47

4.48

4.49

4.50

4.51

4.52

4.53

4.54

5.1

5.2

5.3

Tensile circum ferential stresses induced in the arterial w all stented by Mlinkthin and M linkthiC k, at loading, achieving the sam e initial lum en diam eter, 0 / o f 3.18 mm. 109 C om pressive radial stresses induced in the arterial w all stented by M linkthin and M link^ck, at loading, achieving the sam e initial lum en diam eter, 0 / o f 3.18 mm. 110 Tensile and com pressive longitudinal stresses induced in the arterial w all stented by M linkthjn and Mlinkthick, at loading, achieving the sam e initial lum en diameter, 0 / o f 3.18 mm. Ill The percentage volum e stress b y the tensile circum ferential stress (A), radial com pressive stress (B), tensile longitudinal stress (C) and com pressive longitudinal stress (D), at loading, w ithin the stented vessel b y Mlinkthin and Mlinkthick, achieving an initial lum en diam eter, 0 / = 3.18 m m at loading. 112 Tensile circum ferential stress induced in the localised stenotic vessel w all stented b y Mlinkthin and Mlinkthick, at unloading, achieving a final lum en diam eter, 0/.- o f 2.44 m m for M linkthin and 0 F o f 2.52 m m for Mlinkthick114 C om pressive radial stress induced in the localised stenotic vessel w all stented by Mlinkthin and Mlinkthick, at unloading, achieving a final lum en diam eter, 0 F o f 2.44 mm for M linkthin and 0 F o f 2.52 m m for Mlinkthick115 Tensile and com pressive longitudinal stress induced in the localised stenotic vessel w all stented b y Mlinkthin and Mlinkthick, at unloading, achieving a final lum en diam eter, 0 F o f 2.44 m m fo r M linkthin and 0 F o f 2.52 m m for Mlinkthick116 The percentage volum e stress b y the tensile circum ferential stress (A), radial com pressive stress (B), tensile longitudinal stress (C) and com pressive longitudinal stress (D), at unloading, w ithin the stented vessel b y Mlinkthin and Mlinkthick, achieving a final lum en diam eter, 0 F o f 2.44 mm for Mlinkthin and 0 F o f 2.52 m m for Mlinkthick117 C om parison o f pressure deploym ent o f M linkthin and Mlinkthick; for 1) free expansion (Case study 1); 2) w ithin the straight stenotic vessel achieving the sam e initial lum en diam eter (Case study 2); 3) w ithin the straight stenotic vessel achieving the same final lum en diam eter (Case study 3); and 4) w ithin the localised stenotic vessel achieving the sam e initial lum en diam eter (Case study 4). 123 C om parison o f m ax von M ises stress found in the structures o f M linkthin and Mlinkthick, upon deploym ent for 1) free expansion (Case study 1); 2) in a straight stenotic vessel achieving the sam e initial lum en diam eter (Case study 2); 3) in a srtaight stenotic vessel achieving the sam e final lum en diam eter (Case study 3); and 4) in a localised stenotic vessel achieving the same initial lum en diam eter (Case study 4). .124 C om parison o f radial recoil o f Mlinkthin and Mlinkthick; 1) free expansion (Case study 1); w ithin a straight stenotic vessel achieving the same initial lum en diam eter (Case study 2); w ithin a straight stenotic vessel achieving the sam e final lum en diam eter (Case study 3); and w ithin a localised stenotic vessel achieving the same initial lum en diam eter (Case study 4). 129

XIV

5.4

5.5

5.6

A.1 A.2 A.3 A .4

Com parison o f tensile circum ferential stresses (A), com pressive radial stresses (B), tensile longitudinal stresses (C) and com pressive longitudinal stresses (D), at loading (1) and unloading (2), in vessels b y M linkthin and Mlinkthick, w hereby the stents w ere expanded to achieve the sam e initial lum en diam eter o f 3 m m , and at unloading M linkthin and Mlinkthick achieved a final lum en diam eter o f 2.28 m m and 2.54 m m respectively. 132 C om parison o f (A) tensile circum ferential stresses; (B) com pressive radial stresses; (C) tensile longitudinal stresses; and (D) com pressive longitudinal stresses, at loading (1) and unloading (2), in stented vessels by Mlinkthin and Mlinkthick, w hereby they w ere expanded to achieve the same final lum en diam eter o f 2.28 m m , w hereby at loading M linkthin achieved the inital lum en diam eter o f 3 m m and 2.48 m m for Mlinkthick134 Com parison o f (A) tensile circum ferential stresses; (B) com pressive radial stresses; (C) tensile longitudinal stresses; and (D) com pressive longitudinal stresses at loading (1) and unloading (2), in stented vessels by Mlinkthin and M linkthjC k The stents w ere expanded to achieve the same initial lum en diam eter o f 3.18 m m , and at unloading M linkthin and Mlinkthick achieved a final lum en diam eter o f 2.44 m m and 2.52 mm, respectively. 136 Finite elem ent m esh o f mlinkthick inside a stenotic vessel. 153 R adial displacem ent distribution throughout m linkthick scaffolding a stenotic vessel. 155 Circum ferential distribution throughout the m iddle plane o f the vessel. 155 Longitudinal distribution throughout the m iddle plane o f the vessel. 156

XV

LIST OF TABLES
4.1 4.2 Geom etric data o f Mlinkthin and M linkthiC k through loading and unloading, subjected to free expansion, achieving an initial outer diam eter o f 3 m m .63 G eom etric data o f Mlinkthin and M linkthiC k through loading and unloading, under the influence o f stenotic vessel o f 0.8m m thickness, achieving the sam e initial vessel lum en diam eter o f 0 / = 3 m m . 70 G eom etric data o f Mlinkthin and Mlinkthick through loading and unloading, under the influence o f stenotic vessel o f 0.8m m thickness, achieving the same final vessel lum en diameter. 87 G eom etric data o f Mlinkthin and Mlinkthick through loading and unloading, under the influence o f localised stenotic v essel o f 0.8m m thickness, achieving the sam e initial vessel lum en diam eter o f 0 / o f 3.18 m m and a final lum en diam eter, 0 F o f 2.44 m m for Mlinkthin and 0 F o f 2.52 m m for Mlinkthick103 Foreshortening o f Mlinkthin and M linkthickafter deploym ent, for all o f the case studies evaluated in this thesis. 126 D ogboning o f M linkthin and Mlinkthick after deploym ent, for all o f the case studies evaluated in this thesis. 127

4.3

4.4

5.1 5.2

X VI

Chapter 1

INTRODUCTION

1.1

Cardiovascular Disease

C ardiovascular disease (CVD) is a general description for a variety o f disorders and conditions that can affect the circulatory system w here the arteries o f the body becom e blocked due to a build up o f plaque. CVD is a m ajor health and econom ic burden throughout the world, especially in m ost developed countries w here it is the leading cause o f death. In 1997, cardiovascular disease caused 43% o f deaths w orldw ide [1]. In Ireland, CVD is the highest single cause o f death and it caused betw een 38-40% o f deaths in 2004 [2]. In general, there has been a drop in prem ature deaths from CVD across alm ost all EU countries (w ith the exception o f Latvia) over the years from 2000 to 2002, see Figure 1.1 [3]. This drop in death rate is partly due to advances in

m edical care such as the developm ent o f alternative non-invasive treatm ents for cardiovascular disease treatm ent, such as balloon angioplasty and stenting. It is also as a result o f increased health awareness. H ow ever, it is im portant to note that this drop still constitutes less than 10% o f total deaths from CVD and it is therefore still vital that further m easures to reduce the m ortality rate from CVD be employed. Throughout the world, private and funded research organisations have been playing a vital role in researching the causes, preventive m easures, treatm ents and diagnostic procedures for CVD, and in particular heart disease. There are several treatm ents for CVD w ith the m ost com m on being artery bypass

Prem ature D eath Rates from C VD in EU

2000 12001 D2002 180 150 Q .

120

& 90
q

0 )
E 01

60 30
_

lim ili n
<r G raph o f prem ature death rates from cardiovascular disease across EU countries, [3].

Figure 1.1

surgery,

balloon

angioplasty,

stent

im plantation,

atherectom y

and

laser

vaporisation techniques [4]. Since their introduction in 1994, the use o f stents has dram atically increased over the years for the treatm ent o f CVD. In 1997, the use o f stents in m any countries ranged from 30% -65% o f the total cases o f coronary heart disease, w here France was found to have the highest, and B razil the lowest, use o f stents [5].

1.2 Coronary Heart Disease

Coronary heart disease is a condition w hich develops due to a build up o f deposits in the wall o f the coronary arteries, know n as plaque. Plaque builds up on the walls o f the coronary arteries and causes hardening and narrow ing o f the arteries, see Figure 1.2. This process, called artherosclerosis, m ay significantly reduce or block blood flow to the heart, ultim ately leading to a m yocardial infarction, see Figure 1.2. H eart disease represents the m ajority o f CVD cases in Ireland and it was responsible for 51% o f total CVD cases in 2004 [2].

Build up o f plaque

R ight coronary artery

B lockage L eft anterior descending artery

Figure 1.2

Illustration o f artherosclerosis in the coronary arteries [adapted from M edline], [7].

Coronary heart disease m ay be treated in a num ber o f w ays, w hich depends on the severity o f the disease. The m ain purpose o f treatm ent is to

restore blood flow to the heart. E arly stages o f the disease m ay be im proved by lifestyle changes, e.g. cessation o f sm oking, exercise program , reduction o f stress. In the m ost severe cases bypass surgery is required to restore blood flow. Coronary artery bypass grafting (CABG) is a surgical procedure that involves detouring blood around a blockage in a coronary arlery by taking a segm ent o f a blood vessel from another part o f the body or an artificial graft. A ccording to AHA, CA BG is the m ost com m only perform ed surgery in the U nited States and around the w orld, w ith over 571,000 CA BG procedures perform ed in 1999 [6]. This procedure is generally very successful, w ith less than 5% chance o f heart dam age and less than 2% m orbidity in m ost cases [4]. H ow ever, it is the m ost invasive vascular treatm ent, and consequently it is lengthy and costly. As a result, CA BG is lim ited to treatm ent o f the m ost severe cases o f coronary heart disease.

The introduction o f Percutaneous Translum inal C oronary A ngioplasty (PTCA) in 1977 [8], was an attem pt to non-invasively treat CYD. The frequent occurrence o f restenosis post angioplasty led to the developm ent o f other catheterbased techniques, including atherectom y, excim er laser and balloon angioplasty com bined w ith stenting [4]. A therectom y is a catheter-based procedure to rem ove plaques w ith a rotating blade device that cuts aw ay the plaques and it is generally used in conjunction w ith balloon angioplasty or stenting w hen plaque is exceptionally hard due to calcification. The use o f an excim er laser for CVD

treatm ent involves the use o f a laser tipped catheter w hich is used to vaporise the arterial plaque. A therectom y and excim er laser treatm ents are rarely used today, how ever, balloon angioplasty and stenting are com m only preform ed procedures. The use o f stents has proven to be m ore effective, efficient, and less invasive in com parison to other treatm ents o f heart disease. In Ireland, recent studies have found that the use o f coronary stents in 1998 w as 2.4 tim es higher than surgical bypass procedures w hilst their usage increased by 31% the follow ing year, see Figure 1.3 [9].
Statistics On Stenting Vs Bypass
Stenting Bypass Graft

Figure 1.3

G raph o f the frequency o f the use o f coronary stenting and coronary artery bypass grafts in 1998 and 1999, [91.

1.3 Percutaneous Transluminal Coronary Angioplasty

Percutaneous translum inal coronary angioplasty (PTCA) w hich is also know n as plain old balloon angioplasty (POBA ) is a widely used catheter-based procedure for the treatm ent o f heart disease. The first PTC A w as perform ed in 1977 by the Swiss physician A ndreas Gruntzig, as an alternative form o f C A B G [8]. PTCA requires a catheter, w ith a small inflatable balloon on the end, w hich is positioned w ithin the narrow ed section o f the affected artery. Inflation o f the balloon catheter causes the balloon to push outw ard com pressing the plaque and expanding the surrounding w all o f the artery. This results in a w ider arterial lum en and hence im proved blood flow, see Figure 1.4. W hile PTC A is an effective alternative to bypass surgery, there is a risk o f early acute closure o f the coronary artery and a high rate o f reoccurance o f the obstruction o f blood flow, i.e. restenosis. Restenosis is the renarrow ing o f the artery lum en induced by elastic recoil, vascular rem odelling and excessive grow th o f neointim al hyperplasia [10], Several balloon dilation techniques were

investigated to m inim ize the onset o f restenosis post PTC A w ithout success [11]. The frequent occurrence o f restenosis post angioplasty led to the developm ent o f intravascular stents.

D eflated balloon

Inflated balloon

Plaque

Com pressed plaque

Figure 1.4

Illustration o f PTC A procedure [adapted from H eart Centre], [4],

1.4 Intravascular Stenting

A n intravascular stent is a small, latticed, expandable w ire m esh tube, w hich is inserted into a blocked artery to restore norm al blood flow [4]. Stenting is also a catheter-based and m inim ally invasive procedure, w hich w as first perform ed in the m id 1980s and first approved by the FD A in the m id 1990s [10]. It is

perform ed, m ost often, in conjunction w ith other catheter-based procedures, such as balloon angioplasty and atherectom y. These adjunct procedures are used to

partially reduce the narrow ing caused by the plaque prior to stenting, w hilst the stent rem ains in the vessel scaffolding the arterial w all to ensure norm al blood flow is m aintained in the long-term. Stenting procedures are m ost com m only perform ed via the femoral artery in the groin. H owever, they can also be delivered via the brachial artery at the elbow or the radial artery in the w rist [12], A guide catheter is initially

inserted into the artery and it is navigated to the location o f the blockage. A dye, w hich acts as a contrast m edium , is then injected through the catheter. The dye can be seen using fluoroscopy and provides a visual aid for the interventional cardiologist throughout the procedure [12], see Figure 1.5.

Dye is injected into the coronary arteries

Coronary a r te ry ------ 1 blockage site

X-ray image

Figure 1.5

Pictorial representation o f a flouroscopic im age o f the heart [adapted from Adam ], [13].

The stent is generally expanded by inflation o f an angioplasty balloon w hich plastically deform s the stent structure. The balloon is then deflated and the catheter is rem oved while the stent rem ains perm anently in the artery, keeping the artery open and hence restoring norm al blood flow, see Figure 1.6 [12]. W ithin one month, the stent becom es em bedded in the w all o f the artery due to restenotic grow th [14]. There are two basic types o f stents, balloon expandable stents and self expanding stents. Balloon expandable stents are m ost com m only used due to their ability to be precisely positioned. They also have the advantage o f allow ing

expansion to a controlled predeterm ined diameter. Self-expanding stents are made from shape m em ory m aterials that can recover their predeterm ined diameter. They are elongated and constrained on a catheter by a rolling m em brane sheath and retraction o f the sheath at the blockage allows the stent to expand and dilate the vessel [15]. The usage o f stents is m ost com m only know n for the treatm ent o f coronary artery disease. H owever, stents are also used for the treatm ent o f

peripheral vascular disease [16], renal vascular hypertension [17], and carotid artery disease [18]. A ll o f these conditions have one aspect in com m on - blocked arteries. All stents are used to restore blood flow w ithin arteries by scaffolding them open.

stent

Figure 1.6

Illustration o f stenting procedure [adapted from H eart Centre], [4].

All forms o f percutaneous coronary interventions cause a certain level o f injury to the vessel w all and it is this injury that is thought to be the m ain cause o f restenosis w ithin arteries. The m ost successful o f all the available catheter-based techniques are balloon angioplasty and intravascular stenting. P ost-angioplasty

restenosis is predom inantly due to elastic recoil and vascular rem odelling [19]. Com parative studies have shown the benefit o f coronary stenting versus balloon angioplasty alone since stenting significantly reduces elastic recoil and vascular rem odelling [20]. However, excessive neointim al hyperplasia has been found to occur post-stenting and, as a result, in-stent restenosis also represents the m ajor lim itation to stenting procedures. It occurs due to adaptation o f the arterial

w hereby the intim a cells begin to proliferate due to the presence o f an im planted foreign object. This grow th m ay aggravate depending on the severity o f the injury to the arterial w all, thus leading to excessive neointim al hyperplasia [21]. Two m ain approaches have been adopted that have been successful in reducing in-stent restenosis; (i) optim ising the stent design to reduce vascular injury and hence subsequent neointim al hyperplasia, and (ii) using novel stent m aterials and coatings to inhibit the growth o f neointim al hyperplasia through the release o f anti-proliferative drugs. It is believed, therefore, that b y com bining

optim al stent design, and both the optim um drug and drug elution method, m ore effective and successful intravascular stents m ay be developed [22], It is therefore vital to address the issue o f the exact m echanical stim ulus for in-stent restenosis. This inform ation w ould be invaluable in term s o f m inim ising the occurrence o f restenosis since it w ould enable stent designs to b e optim ised to lower this stim ulus and also to ensure drug delivery to the areas m ost affected by the stim ulus for restenosis. It is generally believed that the stim ulus for restenosis is m echanical injury but a m easure for the injury has not yet been conclusively determined.

1.5 Clinical and Finite Element Studies of Stents

N um erous clinical studies have been carried out to evaluate the influence o f stent design on in-stent restenosis outcome. H ausleiter et al. [23] dem onstrated that

stent design is the strongest predictor o f in-stent restenosis. Escaned et al. [24],

H ausleiter et al. [23] and H offm ann et al. [25] dem onstrated that the M ultiLink design resulted in the low est in-stent restenosis rate in com parison to som e other com m ercially available stent designs. A further clinical study, the ISA R

STEREO Trial by K astrati et al. [26] used two variations in the M ultiLink stent design to assess the effect o f strut thickness on in-stent restenosis outcom e. The clinical result indicated the use o f a thinner-strut stent is associated w ith a significant reduction in angiographic and clinical restenosis. The finite elem ent m ethod is a num erical m odelling technique that can be used to provide inform ation on the m echanical behavior o f com plex structures. It has therefore been used extensively in recent years for research and developm ent into stents as an integral part o f the design process. Several

num erical m odelling sim ulation studies have been used to address the issue o f optim izing stent designs. The m ajority o f these studies have concentrated on

analysing the expansion o f different stent designs [27, 28, 29] w hilst only recently have the stenting sim ulations included the arterial w all [30, 31, 32], Com plex

finite elem ent analyses o f this nature, sim ulating stent expansion w ithin arteries, can be used to preclinically test and evaluate the perform ance and effectiveness o f stents as w ell as their propensity to cause in-stent restenosis, thereby providing valuable inform ation on the optim um stent design. O ptim isation o f stent designs cannot be perform ed if the m echanical stim ulus for restenosis is unknown. It therefore still rem ains to identify m ore

conclusively the m echanical stim uli for in-stent restenosis in order to know w hat m echanical variables that need to be m inim ised to reduce in-stent restenosis. This can be achieved b y using num erical m odelling techniques to sim ulate the conditions o f previous clinical trials that have clearly identified particular stent designs to have a greater propensity for restenosis than others.

1.6 Objectives of the study

The m ain objective o f this study is to identify the m echanical stim uli for restenosis. The ISA R -STEREO Trial by K astrati et al. [26] exam ined the clinical restenosis outcom e o f the ACS RX M ultilink and the ACS M ultilink RX Duet. These M ultilink stents are sim ilar in design but w ith strut thicknesses o f 0.05 mm

and 0.14 m m , respectively.

This study focuses on exam ining the role o f strut

thickness on restenosis outcome and the hypothesis that a thicker strut stent induces higher stresses on the arterial w all than a thinner strut stent, using the finite elem ent m ethod (FEM). These higher stresses induced in the arterial w all are believed to cause different injury levels to the arterial w all and subsequently influence the variation o f restenosis outcom e found in the clinical study b y K astrati et al. [26]. To achieve this objective, finite elem ent m odelling w as used to exam ine the expansion o f both stents in various stenosed vessel geom etries and the results o f these sim ulations are presented here. The stresses induced w ithin the vascular w all by the stents w ere estim ated for the tw o different stent designs. The

differences in the levels o f stresses in the radial, circum ferential and longitudinal direction w ere exam ined for both stent designs. A com parison to clinical data was used to determ ine the m ost likely m echanical stim uli for in-stent restenosis. Overall, this com putational study aims to provide valuable inform ation to aid the optim isation o f coronary intravascular stent designs.

10

Chapter 2

LITERATURE REVIEW
2.1 Introduction
In-stent restenosis represents the m ajor lim itation for stenting procedures. In-stent restenosis is the renarrow ing o f the artery lum en w ith in a stent predom inantly due to excessive grow th o f neointim al hyperplasia. Stent design has been found to be a m ajor factor in determ ining restenosis outcome. This study focuses on exam ining the role o f strut thickness on the grow th o f neointim al hyperplasia. It is therefore vital to understand the restenosis process and the progression o f this grow th that subsequently leads to in-stent restenosis. Finite Elem ent m odelling can be used to investigate the influence o f stent strut thickness on in-stent restenosis b y sim ulating the biom echanical interaction betw een a stent and an artery w hereby stresses w ithin the arterial vessel wall, caused by the stent stm ts, are com puted and exam ined. The reliability o f the

com puted results in any finite elem ent study, how ever, depends upon m any param eters; the m ost essential is the accuracy o f the m aterial m odels for the stent and the artery. It is therefore necessary to establish an appropriate constitutive m odel for the artery by exam ining the structure and general characteristics o f arterial wall. This chapter therefore review s the follow ing:-

1. The structure o f arterial the w all 2. C onstitutive m odels o f the arterial w all

3. The m echanism o f restenosis 11

4. Clinical factors identified to influence the progression o f restenosis 5. Finite elem ent studies o f stenting procedures

2.2 Structure of the Arterial Wall

B efore num erical m odels o f stenting procedures can be carried out, it is necessary to understand the artery into w hich the stent is expanded. A rteries are blood

vessels that supply oxygenated blood to the organs and cells throughout the hum an body. Coronary arteries are vessels that carry oxygenated blood to the

heart. A rteries are m ade o f three distinct layers; the tunica intim a, tunica media, and tunica adventitia [33], see Figure 2.1. The intim a consists o f a m onolayer o f endothelial cells (sm ooth m uscle cells) and an underlying thin basal lamina. The basal lam ina consists largely collagen, the adhesion m olecules lam inin and fibronectin, and of some

Endoitettal cell

Jnrtiroa Media Esrtcrnai elastic

memljrafie

elastic membrane

A d v e n titia

F ig u re 2.1

Illustration o f three layers m ade up an artery [adapted from [34]].

12

proteoglycans, w here its prim ary function is to provide a base for the endothelial cells to grow [33]. This layer is crucial since the stent com es into direct contact w ith the lum en o f the arterial wall. The m edia consists o f beds o f sm ooth m uscle cells that are em bedded in an extracellular m atrix o f small am ounts o f elastic tissue, collagen and proteoglycans. The com position o f the arterial w all becom es stiffer as it reaches the outer layer [33]. The ratio o f collagen and elastin determ ines the rigidity o f the arterial w all [35]. The internal elastic lam ina separates the intim a and media, w hilst the external elastic lam ina separates the m edia and the adventitia. These elastic lam inae allow the transport o f w ater, nutrients, and electrolytes across the w all as w ell as direct trasm ural cell-to-cell com m unication betw een the individual layers o f the artery [33]. The adventitia consists o f a dense netw ork o f collagen fibres and elastic tissues. In coronary arteries, the adventitia layer takes up only approxim ately

10% o f the arterial wall. D ue to its stiff com position, it is believed to serve as a protective sheath, w here it limits acute overdistension o f the vessel. U nder

norm al physiological load, the adventitia does not contribute extensively to the m echanical behaviour, it prim arily provides tethering to the connective tissue and provides additional structural support [36]. The size o f coronary arteries range from 1.59 m m to 4.15 m m in inner diam eter [35] and the thickness o f atherosclerotic hum an coronary arteries have been reported to range from 0.56 mm to 1.25 m m , depending upon the location o f the coronary arteries on the surface o f the heart [37]. surrounding

2.3

Constitutive Models of the Arterial Wall

The constitutive m odel used to describe the arterial w all is very im portant in finite elem ent analyses to investigate the interaction betw een a stent and an artery. M any researchers have investigated and experim entally determ ined the

m echanical behaviour o f arterial tissue. The increase in studies to characterise the m echanical properties o f arterial tissue is due to the advances in intravascular procedures, such as balloon angioplasty and stenting that m echanically deform the

13

vessel wall.

Finite elem ent analysis techniques have been used to preclinically

investigate the perform ance o f stents and they require suitable constitutive laws to represent the m echanical behaviour o f arterial and plaque tissue, and the stent m aterial. A constitutive m odel is the m athem atical expression o f the relationship betw een the stress and the strain o f a m aterial and it can be established from experim ental m easurem ents. The m echanical behaviour o f an artery, and hence the constitutive law to represent it, is dependent upon the com position o f the artery and it m ay vary w ith the type o f vessel, disease, age and gender [38, 39] and the deform ation m ode the sample is subjected to upon testing [40],

A rteries are found to exhibit the following characteristics

1. V iscoelasticity Taking into account the structural constituents o f an artery, arterial behaviour is m ore accurately described as viscoelastic, since arteries have both a solid and fluid com ponent and hence elastic and viscous behaviour. Arteries

exhibit hysteresis under cyclic loading, stress relaxation under constant extension and creep under constant load. H ysteresis is observed on loading and unloading curves do not coincide, see Figure 2.2 [47], W hen a segm ent o f an artery is tested for a tensile test w ith cyclic varying strain, a process called preconditioning is depicted w hereby the stress response shows the hysteresis loop decreases w ith succeeding cycle and lead to a steady state after a num ber o f cycles, see Figure 2.2 [47]. The viscoelasticity o f arterial tissue is m ost evident in the tissu es strain-rate dependency, w hereby there is a developm ent o f additional strains in the arterial tissue over a period o f tim e w hen subjected to stress. A lthough arterial tissue is truly viscoelastic, purely elastic m odels have been used to describe arterial tissue. The elastic m odels have been used in num erical m odels w hereby the experim ental data used to describe the constitutive m odel was obtained at strain rates typical o f that applied in the application o f the m aterial m odel [41].

14

F ig u re 2.2

Cycle

stress response o f a d o g s carotid

artery under

preconditioning showing hysteresis loop over a num ber o f cycles [adapted from [47]].

2. H eterogeneous A n artery is com posed o f endothelial cells, elastin and collagen, and the distribution o f these constituents varies from one layer to the other, see Figure 2.1. The varied orientation o f these structural constituents determ ines the m echanical properties o f the artery. M any constitutive m odels used in the literature to

describe arterial tissue assum e hom ogeneity for sim plification o f the constitutive m odel [40], The assum ption o f hom ogeneity m ay be ju stified because the m edia constitutes the m ajority o f the vessel w all since the adventitia and intim a generally constitute only 10% o f the vessel wall, and as a result the m edia bears m ost o f the load [33]. M oreover, the experim ental data used to describe constitutive m odels generally includes the overall m echanical behaviour o f the arterial tissue, where uniaxial and equibiaxial data are generally used to describe the m odel, [40]. Therefore, for the purpose o f investigating the stresses and strains induced in the w all o f a coronary artery b y a stent, the artery properties m ay be adequately represented b y a m aterial m odel that assum es hom ogeneity.

15

3. Incom pressibility Tests on arteries have shown very sm all changes in volum e o f the arterial w all over a w ide range o f deform ation. Sim ilar to m any other soft tissues the arterial w all is regarded as almost incom pressible, w here it preserves its volum e under load [42],

4. Residual stress R esidual stresses in arteries are the stresses that exists in the arterial wall in the absence o f an externally applied load. R esidual stress can be characterized b y the level o f opening angle w hen a sample segm ent is radially cut [43]. The inclusion o f residual stress into an arterial m odel w as found to result in a m ore uniform distribution o f stress, particularly a decrease in stress at the intim al as opposed to a m odel w ithout residual stress [40].

5. A nisotropy Sm ooth m uscle cells are found to be oriented circum ferentially in arterial tissue and w hen they are subjected to an internal pressure, constriction and dilation o f the sm ooth m uscle cells alters their orientation. Therefore the arterial w all m ay be considered anisotropic, w here it exhibits different properties in different directions [40]. H olzapfel and O gden [44] proposed a com bination o f exponential and polynom ial type strain energy functions w hich include the anisotropic behaviour exhibited by arterial tissue. They have suggested dividing the strain energy function into the addition o f the isotropic strain energy function and an anisotropic strain energy function. A lthough the constitutive m odel used by H olzapfel successfully captures the anisotropic characteristics o f arterial tissue it is unavailable in com m ercially available softw are codes and hence com plex and highly com putationally expensive to im plem ent. In addition, the degree o f anisotropy o f tissue is highly patient specific and w ould therefore be m ost valuable for describing the m echanical properties o f arterial tissue from in vivo m easurem ents.

16

For the purpose o f preclinically identifying k ey characteristics o f stent designs that affect restenosis outcome, it is proposed that an isotropic m odel m ay be sufficient. A n isotropic m odel that includes the m ain features o f arterial tissue such as its non-linear stress stiffening behaviour and its incom pressibility could identify the influence o f stent design variables on restenosis outcome

preclinically. M any researchers have proposed various constitutive equations, to describe the m echanical behaviour o f arterial tissue [40, 44, 45], The elasticity o f an artery is norm ally quantified in term s o f a strain energy density function, W, where the recoverable energy is stored in the m aterial as it deform s. W is defined in term s o f suitable strain com ponents and the description o f W is used to describe a constitutive equation for the arterial tissue. This is due to the fact that a hyperelastic m aterial is one that, by definition, experiences large deform ations that are com pletely recoverable and has m echanical behaviour that m ay be defined by its strain energy density function, W , such that differentiating W w ith respect to a strain com ponent gives the corresponding stress com ponent. For example, w here W is described in term s o f the G reen-Lagrangian strain tensor, E, then each com ponent o f the 2nd P iola-K irchoff stress tensor, S, is given by the

differentiation o f the strain energy density function, W , w ith respect to the corresponding com ponent o f the G reen-Lagrangian strain tensor, E [33]. This may be described b y the follow ing equation:-

s = aw
3E

Ideally artery w all m aterials should be represented by a viscoelastic model. Holzapfel et al. [42] have proposed a viscoelastic m odel, describing the strain rate dependent characteristics o f the tissue and also including hysteresis. The stress-strain relationship was found to be considerably different at loading and unloading w ith each stress-strain relationship highly non-linear and unique to each cycle in the process. The strain energy function proposed b y H olzapfel et al. [42] was decoupled into a viscoelastic and an elastic part. In addition, H olzapfel

17

et al. [42] divided the artery into two layers, the m edia and the adventitia, describing each w ith a unique strain energy density function. T he m odel also included anisotropy through the inclusion o f fibre orientations in each layer o f the model. In the m ost recent com putational study o f assessing stent design by H olzapfel et al. [32], an elastic anisotropic constitutive m odel is used where the strain energy density function w as given b y the sum o f the isotropic and anisotropic strain energy functions. The m aterial data w as taken from an external iliac atherosclerotic artery o f a 65 year old w om en and the m odel w as therefore patient-specific. The m aterial properties o f arterial tissue is know n to be dependent upon the com position o f the artery and it m ay vary w ith the type o f vessel, disease, age and gender [37, 38, 39], It m ay therefore still be useful to investigate stent design param eters in com parative FE studies using sim plified isotropic constitutive m aterial m odels that are derived based on experim ental data from tests on arterial tissue. Therefore, for the purpose o f com parative analyses, the isotropic M ooney-R ivlin m odel is proposed to lim it the com plexity o f m odelling the process o f stent expansion w ithin an artery vessel. The m ain objective o f this

thesis is to estim ate the stresses induced in the arterial w all o f stented stenotic vessels by different thickness stent struts, and to relate these stresses to the different restenosis outcom es observed w ith these different stents. The non

linearity o f the stent m aterial and the com plex stent geom etry alone result in a highly com putationally expensive sim ulations. The key m echanical property o f arterial tissue is its stress stiffening behaviour and uniaxial and equibiaxial data are sufficient to adequately represent this behaviour in the m ain deform ation modes that stented arteries experience [46]. The strain energy density function, W, o f an isotropic hyperelastic m aterial can be described in polynom ial or exponential form. O ne o f the general polynom ial form s o f the strain energy density function in term s o f stretch invariants that has been extensively used to m odel incom pressible isotropic hyperelastic m aterials is know n as the M ooney-R ivlin m odel and is given b y [47]:

18

W ( I,,I2,l3,)= i > s ( /, - 3 ) (72 - 3 ) " ( / , - 3 ) , a 000 = 0 1.7=0

W here ay are m aterial constants, and m and n are exponents w hich determ ine the order o f the model. The stretch invariants for the m aterial are defined through the principal stretches (A x, A 2, A3) o f the m aterial as:

Ii Af + A2 + Aj I2 Aj A^ + Af Aj + A\ Aj I3 = A2 , AI A]

The M ooney-R ivlin m odel is w idely available in a range o f form s in m any FE codes. Petrini el al. [48] and Lally et al. [31] have used form s o f the M ooney-R ivlin m odel in their sim ulations investigating the interaction betw een a stent and an artery, w here Petrini et al. [48] com pares the perform ance o f a selfexpandable stent w ith a balloon-expandable stainless steel stent and L ally et a l [31] com pares the perform ance o f two different stent designs. B oth Petrini et al. [48] and Lally et al. [31] used a third order form o f the M ooney-R ivlin m odel to describe the m echanical behaviour o f the artery. The M ooney-R ivlin m odel can be used to define a suitable constitutive equation to describe arterial tissue in this study by fitting to published experim ental data from uniaxial and equibiaxial tension tests on porcine coronary tissue [41]. Ideally, num erical studies on intravascular stents should use data from fresh hum an coronary arteries, how ever, data on hum an coronary tissue is very difficult to obtain. Recently, van A ndel et al. [37] carried out a com parative study o f the inflation response o f a lim ited num ber o f hum an and porcine arteries. The study found that the porcine arteries have three tim es greater elasticity than the hum an arteries. However, the published test data cannot easily be used to define the constitutive m odel param eters w ithin the available m aterial m odel in M arc/M entat (M scSoftware, Santa Ana, CA, U SA). In addition, another study

19

has show n that the porcine coronary artery has very sim ilar m echanical properties to the hum an coronary artery [49]. The m ain goal o f FE studies on stenting procedures is to understand the effects o f stent deploym ent on the arterial wall. These num erical studies are

highly concerned with the large deform ation o f the arterial w all caused by the stents structure. Some studies, including recent studies, have m odelled the artery using a sim ple linear-elastic m odel [50, 51, 52]. O thers have used isotropic

hyperlastic constitutive m odels based on uniaxial data [48, 53]. It is proposed that the use o f the experim ental data describing the m ain deform ation m odes that arterial tissues physiologically experience, nam ely uniaxial and biaxial tension, is the best w ay to describe an isotropic material w hich can represent the properties o f arterial tissue. This data is available in the literature fo r porcine coronary arteries [41]. The uniaxial and equibiaxial data are found sufficient to represent the key feature o f the non-linear stress-stiffening behaviour o f the arterial tissue. A therosclerotic vessels are also com posed o f stenotic plaque w hich m ay also be described by a hyperelastic constitutive m odel and used to represent the plaque tissue in num erical m odels. A suitable constitutive m odel for plaque can be described by fitting to published data on the stress-strain behaviour o f hum an atherosclerotic plaque. Experim ental data from uniaxial tension tests on hum an aortic atherosclerotic plaques has been published b y Loree et al. [54]. To the

authors know ledge no other uniaxial tensile test data on hum an atherosclerotic plaques have been published and no data on the biaxial tensile behaviour o f hum an atherosclerotic plaques could be found in the literature.

2.4

Mechanism of Restenosis

Restenosis is the reoccurance o f a stenosis w ithin a vessel after an intravascular procedure. Reduction o f restenosis is the m ain advantage o f stenting compared w ith conventional PTCA. B alloon angioplasty is less successful com pared to

stenting due to elastic w all recoil and vascular rem odelling, w hich results in early lum en loss [19]. Stenting was introduced to act as a perm anent scaffold following PTC A and thereby prevent negative rem odelling and elastic recoil o f the arterial wall.

20

C om parative studies between stenting and angioplasty have shown that stenting had a higher rate o f procedural success than angioplasty. One study

shows a larger im m ediate increase in the diam eter o f the vessel lum en (1.72 mm vs. 1.23 mm, P<0.001), and a larger lum inal diam eter im m ediately after the procedure (2.49 m m vs. 1.99 mm, PO .O O l) and a low er rate o f binary restenosis (31.6% vs. 42.1% , P=0.046) than vessels treated w ith balloon angioplasty [20]. B EN ESTEN T and GISSO C trials show the benefit o f coronary stenting versus balloon angioplasty, w here the results reported the effectiveness o f stenting in treating stenosed coronary arteries. The results o f these studies dem onstrate the suitability o f stents to achieve large lum en cross-sectional areas and their ability to prevent negative vascular rem odelling w hilst scaffolding the vessel open [55, 56,

The factors w hich cause restenosis after balloon angioplasty are acute vessel recoil, throm bus form ation, chronic constrictive rem odelling and

neointim al grow th [58]. In-stent restenosis occurs due to adaptation o f the arterial w all to the presence o f an im planted foreign object and it is predom inantly due to grow th o f neointim al hyperplasia [59, 60], All o f these studies have identified

sm ooth m uscle cell hyperplasia as the m ajor com ponent o f in-stent restenosis [61]. The m echanism o f in-stent restenosis is im portant in understanding the stim uli for the disease. M any researchers have observed and identified the m ain steps leading to the form ation o f in-stent restenosis as follows, see Figure 2.3:-

1. 2.

Initial reaction to the foreign body leading to throm bosis. This grow th o f throm bosis leads to an inflam m atory reaction o f the arterial wall.

3.

Inflam m ation is follow ed by proliferation o f intim al cells. This phase is found m ore prom inent after a stenting procedure com pared w ith balloon angioplasty, w here the stent is perm anently fixed as it scaffolds the arterial wall.

4.

Finally rem odelling o f the artery w all occurs due to the abundant grow th o f sm ooth m uscle cells.

21

Ttwombus deposition

Days after stenting

Figure 2.3

The four phases o f vascular repair after stenting, leading to the form ation o f in-stent restenosis [14].

2.4.1 Thrombosis
Throm bus form ation begins at the site o f injury caused b y the stent struts. The throm botic phase usually takes place w ithin 1-3 days after the intervention [14, 59], As the stent struts are expanded to scaffold open the arterial w all, the stent struts m ay penetrate and lacerate the arterial wall. This is defined as the acute

arterial injury to the arterial w all that initiates the initial healing process. Throm bosis is the form ation o f a blood clot, a sealing m echanism to restore norm al blood flow w hen injury is induced in the vessel wall. The

form ation o f a throm bus is governed b y the severity o f dam age to the arterial w all and can obstruct blood flow [62],

22

2.4.2 Inflammation

The throm bus at arterial injury sites develops an inflam m atory reaction w here leukocytes m ake their w ay to the tightly stretched internal elastic m em brane o f the artery. This process occurs approxim ately 3-7 days after the intervention and

results in a layer o f inflam m atory cells at the throm botic sites at and betw een the stent struts [14].

2.4.3 Neointimal proliferation

Proliferation is the third phase o f vascular response to stenting.

Intim al cells

begin to proliferate and sim ultaneously the inflam m atory cells m igrate from the vessel surface to the neointim a. This process begins approxim ately 7 days after the stent im plantation and m ay continue for m any m onths [14]. A t this stage the throm bus and inflam m atory cells are replaced b y neointim al cells that

progressively proliferate tow ard the injured m edia. This leads to the activation o f sm ooth m uscle cells (SM C) in the m edia. The activation o f the sm ooth muscle cells results in both SM C proliferation and SM C m igration into the intima. Sim ultaneously large am ounts o f extracellular m atrix are produced b y the SMCs. This grow th response to the arterial injury leads to the developm ent o f neointim al hyperplasia or the thickening o f the intima. As the grow th o f the neointim al

hyperplasia progresses, it results in a narrow ed lum en, thus blocking the flow o f the blood to the heart m uscle, see Figure 2.4 [21, 63, 64, 65].

23

Stent struts

Figure 2.4

A n illustration o f response to injury from stent strut expansion in norm al porcine coronary artery after 28 days. N =neointim a; IE L/EEL =intem al/extem al elastic L=lum en; lam ina

[adapted from Low e et al., [65]].

2.4.4 Vascular remodelling

The last phase in the response to vascular injury is the vascular rem odelling. In response to the neointim al hyperplasia growth, geom etric rem odelling occurs and causes the artery to contract [14]. The artery is dynam ic and responsive and it adapts to the high strain im posed by the stent. It fights against the resulting strain o f the arterial w all due to neointim al hyperplasia by increasing levels o f collagen and stiffening the vessel wall. H owever, in the case o f a stented vessel, the high radial strength o f stent prevents vascular rem odelling after stenting and it has therefore been found to be an alm ost insignificant contributor to in-stent restenosis [66]. The grow th o f neointim al hyperplasia has been found to be the m ain lim itation in the long-term success o f intravascular stenting procedures. V ascular injury caused by the stent structure is the m ajor determ inant o f the grow th o f neointim al hyperplasia [67]. D rug eluting stents have em erged in the recent years, w here the drug inhibits SM Cs proliferation [68]. W ith or w ithout drug elution,

24

stent design has been extensively reported to p lay a vital role in determ ining restenosis [69].

2.5 Clinical Trials on Intravascular Stents Examining Restenosis Outcome

Cellular response is the key to the developm ent and progression o f in-stent restenosis. It is still as yet unknow n the exact feature o f a stenting procedure that stim ulates cell proliferation. Some researchers have considered the influence o f the altered fluid flow around the stent and hypothesised that it m ay govern the developm ent o f neointim al proliferation [70]. Som e other studies concluded that the m aterial o f the stent and/or stent coating played an im portant role in governing this proliferation [71]. H owever, a large num ber o f studies have dem onstrated that stent design is the m ajor key factor in determ ining restenosis outcom e [22, 69, 72]. These studies on stent design have also investigated the influence o f the m echanism o f expansion, the strut thickness, the m etal-to-artery ratio, the cross sectional area and the length o f the stent structure have on restenosis outcome.

2.5.1 Influence of stent design on restenosis outcome

There are tw o basic designs o f stent; (i) self-expanding and (ii) balloonexpandable stents. Self-expanding stents are m ade from shape m em ory materials that can recover their predeterm ined diam eter w hen placed at the site o f treatm ent [15]. sheath. They are elongated and constrained in a catheter by a rolling m em brane R etraction o f the sheath results in stent self-expansion. The most

com m on self-expanding stent is the W allstent (Schneider), see Figure 2.5, w hich has a m ultiple w ire braid structure and is com m only m ade from a cobalt alloy [73]. The m ain advantage o f the W allstent is its flexibility. It is considered one o f the m ost desirable features in a coronary stent w here it eases deploym ent through tortuous vessels and consequently results in a less traum atised vessel w hen subjected to bending [74]. Self-expanding stents, such as the W allstent,

have an advantage in term s o f lum inal gain since their inherent expanding force

25

enables them to continue to expand after im plantation [24].

D espite the

W allstents advantage o f flexibility and higher lum inal gain, assessm ent o f its clinical outcom e has show n that this design is inferior in the long-term w here it caused the highest restenosis rate o f 49% in com parison to balloon expandable stents [24], The high lum inal gain and the large m etal surface area o f this stent design led to a greater grow th o f neointim a. As discussed above, m any studies have identified sm ooth m uscle cell hyperplasia as the m ajor com ponent o f in-stent restenosis. It can b e concluded that high lum inal gain can serve as a disadvantage w here it challenges the original b elief that bigger is better [75]. However,

recent studies show prom ising results b y com bining the use o f intracoronary beta radiation w ith the W allstent, w hereby the beta radiation reduced neointim al proliferation and thus m aintained a w ider lum en diam eter [76]. B alloon expandable stents are com m only m ade from stainless steel m aterials. They are expanded by a balloon and plastically deform ed to a desired diameter. These are m ost com m only used because they can be precisely

positioned and have the advantage o f being able to accurately achieve a predeterm ined diameter. There are tw o m ajor designs o f balloon expandable

stent; coil stent designs and slotted tube stent designs. The form er is fabricated out o f a continuous single strand o f w ire form ed into a repeating pattern w hile the latter is m ade o f a stainless steel tube b y laser m achining the tube into the desired pattern [73]. The m ost com m on coil stent is the G ianturco-R oubin (Cook), see Figure 2.5, w hich has a coiled structure and a continuous spine that serves as longitudinal support, and prevents separation o f the coils. This coil like design has a very flexible structure and is therefore often favoured for com plex bifurcation lesions [10]. However, clinical studies have reported restenosis rates o f 43% [77], 46% [24], and 57% [10] for the G ianturco-R oubin stent. These

studies have dem onstrated that the coil stent design has less initial gain and significantly large lum inal loss during follow-up. A high level o f recoil

contributed largely to the large lum inal loss found in these studies. C arrozza et al. [78] have dem onstrated that the G ianturco-R oubin stent recoiled 30% o f its initial gain cross sectional area. The coil stent designs have shown poor radial strength and allow tissue prolapse betw een the w idely spaced w ire elem ents [24], In

26

addition, neointim al proliferation also contributed largely to the long-term luminal loss. The earliest com m ercially available stent design w as a slotted tube stent made from a stainless steel tube called the Palm az-Schatz stent (Johnson & Johnson), see Figure 2.5, [73]. The P alm az-Schatz stent has been found to have low er recoil (15% ) w hen com pared to the G ianturco-R oubin stent [78]. It has

been proposed that this low er recoil that subsequently lead to low er lum inal loss as recorded in the clinical study b y Escaned et al., [24], The Palm az-Schatz stent structure provides good surface coverage o f a stented lesion and subsequently provides good scaffolding against the collapsing force o f the arterial wall. Clinical studies have reported 98% o f procedural success w ith the Palm az-Shatz stent [10] w ith restenosis rates o f 20% reported by E scaned et al. [24] and 36.8% b y H ausleiter et al. [23]. The slightly higher restenosis rate reported b y H ausleiter et al. [23] was expected as the stents w ere im planted in sm all coronary arteries. These restenosis rates are consistent w ith the findings o f Colom bo and Tobis [10] w here the restenosis rate associated w ith the Palm az-Schatz stents w ere reported to be 19% in large vessels and 30% for small vessels. Stankovic et al. [79] have also reported a significant reduction in angiographic restenosis for the sam e group o f stents used in large vessels com pared w ith those used in sm all vessels. The early slotted tubular stent design becam e the basic structure o f the m odem stent designs. Studies show ed that flexibility is a critical requirem ent o f a stent [10]. D uring deploym ent, the stent m ay be subjected to bending at certain areas and therefore flexibility is essential to ease deploym ent through tortuous vessels to the target area. Advances in m anufacturing m ethods have enabled more flexible geom etries to be generated in recent years. The em ergence o f laser

m achining has also created diversity in stent designs [73], R ecent generation o f tubular stents can be classified into slotted tube and m odular stent designs [69]. The recent generation o f slotted tube stent design has a closed cell structure, a com m on exam ple is the N IR stent (M edinol/B oston Scientific), see Figure 2.5, w hilst a m odular stent design has an open cell structure, a com m on exam ple being the M ultiLink stents (Guidant), see Figure 2.5. Escaned et al. [24] exam ined both o f these afore m entioned stent designs in a clinical study and found them to have sim ilar scaffolding properties, sim ilar radial strength and sim ilar total m etal area.

27

This clinical study also com pared the restenosis rate for m ulticellular, slotted tube, coil and self-expandable stent designs. The M ultiL ink and N IR stents w ere in the m ulticellular design group and these stents w ere found to have the low est restenosis rate o f 10% in com parison to the other stents. C arrozza et al. [78] reported sim ilar percentage diam eter recoils for these m ulticellular stents, w ith 9.6% observed in the M ultiL ink stent and 10.4% in the N IR stent. C osta et al. [80] carried out a com parison study betw een the

M ultiLink stent and N IR stent. They reported that the M ultilink stent resulted in a slightly higher degree o f neointim al form ation at 6 m onths o f 46 m m 3 w hilst the N IR stent resulted in 39.9 m m 3 o f neointim al formation. This w as expected as the length o f the M linkL ink stent w as greater than the N IR stent and stent length has been reported as one o f the factors affecting neointim al proliferation [23]. H offm ann et al. [25] reported the intim al hyperplasia thickness caused by the M ultiLink stent and the N IR stent to be 0.20 mm and 0.31 m m respectively, w hile the Palm az-Schatz caused 0.26 m m thickness o f intim al hyperplasia. The same pattern neointim al form ation was observed in a clinical study b y H ausleiter et al. [23] w hereby the restenosis rate w as reported to be 29.6% in the M ultiLink, 36.8% in the Palm az-Schatz and 44.3% in the N IR stent. These recent stent designs addressed m ore difficult and challenging tasks such as achieving flexibility w ithout com prom ising radial strength and lesion coverage. The developm ent o f new er stent designs generally revolves around

varying the m ulti-ring design such as varying the connection betw een the rings. There has been a vast grow th o f com m ercially available endovascular stents in recent years [72]. It is clear that stent design plays a m ajor role in determ ining restenosis outcome. Taylor et al., [81] studied the stent-artery interactions o f four different stent designs; the B X stent, M ultiLink stent, R adius N itinol stent and Palm az-Schatz stent, see Figure 2.5. There w as a significant difference betw een the vascular response o f each o f these stent designs, as show n in Figure 2.6.

28

Figure 2.5

Illustration o f selected stent designs: (A) self-expanding stent, W allstent (Schneider), (B) coil stent, G ianturcoRoubin stent (Cook), (C) slotted tube stent, Palm az-Schatz stent (Johnson & Johnson), (D) closed cell structure, N IR stent (M edinol/B oston Scientific), (E) open cell structure, M ultiLink stent (G uidant), (F) B X stent (Cordis) and (G) Radius N itinol (Scim ed), [adapted from [10, 73]].

A.

B.

C.

D.

Figure 2.6

Representative M ovat pentachrom e-stained sections from the four stent designs: (A) B X stent, (B) M ultilink, (C) Radius N itinol, and (D) Palm az-Schatz, [adapted from Taylor et al., [81]].

29

Follow ing studies o f in vitro and in vivo testing, anim al studies, clinical evaluation and clinical application over the years, the ideal stent design is characterised to com pose the following critical requirem ents [10]:-

1. 2.

G ood flexibility - to navigate through tortuous vessel A dequate radiopacity to aid visual visibility through stenting

procedure for accurate m anipulation and positioning 3. Low recoil - to m aintain the desired resulting lum en diam eter, w ithout com prom ising lesion coverage. 4. Sufficient radial strength - to w ithstand the collapsing o f the arterial vessel and to prom ote adequate scaffolding ability 5. A low profile - to have m inim al w all thickness and m etallic surface area as they w ere found to contribute w idely to throm bosis form ation and neointim al proliferation. 6. G ood biocom patibility - to ensure low adverse reaction o f the stent m aterial w hen im planted inside the hum an body

2.5.2 Influence of stent material on restenosis outcome

C linical studies have shown the superiority o f one stent design over another. C linical trials have evolved to exam ine the specific characteristics o f a stent that cause the differences in the level o f restenosis outcome. Stent coating has been identified as one o f the predictors o f restenosis outcome. Endovascular stents are frequently m ade o f stainless steel [12]. It is durable and has a relatively good

radiopacity. H owever, the thinner the struts o f a stainless steel stent, the low er the level o f radiopacity [82], Therefore, som e gold-coated stents w ere m ade available for the purpose o f increasing radiopacity. Gold has a low m elting point and high m alleability, w hich leaves it ideal to w ork w ith [82], A lthough gold-coated stents have a high radiopacity, they have relatively poor biocom patibility in com parison to stainless steel stents [23, 25, 82], These clinical studies by Edelm an et al. [82], H ausleiter et al. [23], and H offm ann et al. [25] exam ined the restenosis outcom e

30

o f gold coated stents versus stainless steel stents and dem onstrated that goldcoated stents trigger significantly m ore intim al proliferation.

2.5.3 Influence of strut thickness on restenosis outcome

R andom ised clinical studies [23, 26, 79, 83] have dem onstrated that strut thickness affects the long-term clinical outcomes. Stankovic et al. [79] also

evaluated the effect o f stent strut thickness o f different stent types (slotted tube and ring design) on angiographic restenosis rates. The stents w ere divided into 2 groups o f thin-strut (<0.1 m m ) and thick-strut (>/=0.1 mm ). The clinical study w as divided according to the coronary vessel sizes (sm all (<2.8 m m ) and large (>/=2.8 m m )) and lesion lengths (short lesion (<15 m m ) and long lesion (>/=15mm )). The clinical result indicated that there is a significant reduction in angiographic restenosis w ith the im plantation o f a thin-strut stent. K astrati et al. [26] investigated the effect o f strut thickness on restenosis outcom e w ith sim ilar stent designs. They random ly assigned a total o f 651

patients in ISAR-STEREO Trial to receive one o f tw o com m ercially available stents o f sim ilar design but w ith different strut thickness. The thinner strut stent used w as the ACS R X M ultiLink (Guidant) w ith a strut thickness o f 0.05 m m and the A CS M ultiLink RS D uet (G uidant) w ith a strut thickness o f 0.14 mm. The clinical result indicated that the use o f a thinner strut stent w as associated w ith a significant reduction o f angiographic restenosis (15% in the thin-strut group and 25.8% in the thick-strut group) and clinical restenosis (8.6% am ong the thin-strut patients and 13.8% in the thick-strut group). A ngiographic restenosis is the

percentage o f cases w here the diam eter stenosis is greater than 50% found during angiographic evaluation and the clinical restenosis is the percentage o f cases that required reintervention because o f restenosis-induced ischem ia. This result is

consistent w ith the findings b y H ausleiter et al. [23] w here the thinner strut stent o f the M ultiLink was associated w ith a low er restenosis rate o f 29.6% as com pared to 34.2% restenosis rate w ith the thicker strut M ultiL ink D uet stent. The ISAR-STEREO -2 Trial by Pache et al. [83] w as carried out after the ISA R-STER EO Trial, w hich concluded that for two stents w ith sim ilar design, the

31

risk for restenosis is dependent on the stm t thickness.

The ISA R-STR EREO -2

Trial assessed the im pact o f stent strut thickness o f different stent designs on restenosis outcome. A total o f 611 patients w ere random ly assigned to receive

either the thin-strut ACS RX M ultiLink stent (G uidant) (strut thickness 0.05 mm, interconnected ring design) or the thick-strut B X V elocity stent (Cordis Corp.) (strut thickness 0.14 mm, closed cell design). The clinical result indicated that the use o f thinner struts is associated w ith a significant reduction o f angiographic restenosis (17.9% in the thin-strut group and 31.4% in the thick-strut group) and clinical restenosis (12.3% o f the thin-strut group and 21.9% o f the thick-strut group), after intervention. C linical studies have assessed the perform ance o f com m ercially

available intravascular stents w ith exam ination o f the restenosis outcom e set as a m easure o f success o f the intravascular stents. It is clear that stent design is a key factor determ ining restenosis outcome.

2.6 Finite Element Studies of Stenting Procedures

Im plantation o f balloon expandable coronary stents in a hum an artery to treat coronary artery disease is a com plicated process. The behaviour o f the stent

expansion during stent deploym ent in an artery is com plicated and difficult to m odel using the finite elem ent m ethod. D espite a num ber o f softw are tools

available, published w orks on the behaviour o f stent expansion using com puter sim ulations are lim ited and the current know ledge is m ostly based on the results o f clinical trials. Stent deploym ent involves the interaction o f a balloon, stent and artery and therefore som e researchers have investigated balloon and stent interaction [50, 84]. M ost o f FE studies, how ever, have investigated free stent expansion w hereby the stent is expanded w ithout any interaction w ith a balloon or an artery [27, 28, 29, 48, 85]. Som e recent studies have investigated the stent and artery interaction [31, 32, 48, 52, 86, 87]. D um oulin and C ochelin [27] sim ulated a balloon-expandable stent that had a sim ilar geom etrical design to the Palm az Schatz stent by using the com m ercial code A BAQ U S. The expansion o f the stent structure and its

32

m echanical properties w ere investigated.

Stent expansion w as characterised by

radial recoil, longitudinal recoil and foreshortening for different final diam eters o f stent expansion. Stress-strain fields in the stent w ere exam ined. It w as found that the m ajor equivalent plastic strains w ere localised in the com ers w hile the m ajor stresses w ere located in the m iddle o f the stent cells junctions. The sensitivity o f the critical pressure to geom etric im perfections was also studied and the fatigue life was tested. The results show ed that the P308 Palm az stent studied was

insensitive to geom etric im perfections and that the stent could w ithstand an indefinite num ber o f cardiac cycles. The m ain lim itation o f this study was the

analysis o f the stent behaviour b y only analysing one unit o f the stent structure. The external force applied to com press the expanded stent structure w as also based on an assum ption and m ay have had a significant effect on the findings o f this study. M igliavaca et al, [29] investigated the m echanical behaviour o f several coronary stents using the finite elem ent m ethod. The effects o f different

geom etrical param eters such as the thickness and the m etal-to-artery surface ratio w ere evaluated in term s o f radial recoil, longitudinal recoil, foreshortening and dogboning. Three m odels w ere analysed, a typical diam ond-shaped Palm az

Schatz stent, a stent resem bling a M ultiL ink Tetra stent (G uidant, Indianopolis, IN, U SA ) and a stent resem bling a Carbostent (Sorin Biom edica, Saluggia, Italy). The stents w ere assum ed to be m ade o f 316L stainless steel. The inelastic

constitutive response w as described through a V on M ises-H ill plasticity m odel w ith isotropic hardening. A large deform ation analysis w as perform ed using the finite elem ent com m ercial code A BAQUS. The results o f the sim ulation showed the influence o f the stent geom etry on the m echanical behaviour o f the stent. A stent w ith a low m etal-to-artery surface ratio w as show n to have a higher radial and longitudinal recoil, but low er dogboning. The stent stm t thickness was found to influence the stents perform ance in term s o f foreshortening, longitudinal recoil and dogboning. The von M ises stress distribution along the stent in the loaded configuration showed that stresses w ere concentrated in the area o f the connection betw een slots, w here the plastic hinges allow ed the stent to expand. Again, the geom etries o f the stent stm ctures w ere based on some assum ptions and the influence o f the arterial w all w as ignored.

33

A recent study o f stent free expansion b y M cG arry et al. [85] incorporated a m icroscale deform ation m echanism into the com putational m odels o f stent expansion. A nother recent study by this research group has show n that the stressstrain behaviour o f stent struts is size dependent and as a result it has been suggested that certain areas o f thin struts could experience local failure w hich w ould not be described if m acroscale stress strain curves w ere used to describe the m aterial behaviour [88]. The finite elem ent analyses o f the stent expansion found non-uniform and localised deform ation occurred in the structure due to the m aterial inhom ogenity and the discontinuity o f crystal lattice orientation w ithin the m icrostructure. The stent geom etry used w as b ased on the N IR stent,

represented b y a tw o-dim ensional m odel o f a single u n it structure. W ith such a m icroscale representation o f the stent, it w ould b e extrem ely difficult to investigate the behaviour o f the entire stent structure and the influence o f the arterial wall. C hua et al. [28] carried out a finite elem ent sim ulation o f free expansion o f a balloon-expandable stent that has a sim ilar geom etrical design to the Palm az Schatz stent. Later, Chua et a l [84] incorporated the influence o f a balloon into their analyses. In the m ost recent study by C hua et al. [52] the deform ation

characteristics o f the stent w ere investigated as a stent w as expanded inside an artery. These analyses w ere perform ed using ANSY S (version 5.5) as a pre- and post-processor and LS-D Y N A for solution. They reported the von M ises stresses in the stent structure and the distribution o f the von M ises stress in the vessel wall. Som e lim itations o f the analyses include the geom etry o f the stent structure, w hich was scaled up, and the use o f 304 stainless steel m aterial properties as opposed to 316L. represented by a In addition, the non-linear m aterial o f the artery was simple linear-elastic model. These assum ptions and

sim plifications in the m aterial m odel w ould affect stent expansion and the resulting stress distribution in the arterial wall. Recent studies o f stent expansion sim ulation studies have included the artery and plaque in the analyses. Petrini et al. [30, 48, 86] have reported

num erous com putational studies exam ining the expansion o f m etallic stents and the expansion o f a m etallic stent in a stenotic coronary artery by using the finite elem ent method. The num erical analyses w ere perform ed using the com m ercial

34

code ABAQUS (H ibbit K arlsson & Sorenses, Inc., Paw tucket, RI, U SA ). Each sim ulation consisted o f three parts: the artery, the artherosclerotic plaque and the stent. The sim ulation consisted o f a stent m odel resem bling the Palm az Schatz stent, and the m aterial w as assum ed to be m ade o f 316L stainless steel. The 316L stainless steel stents w as described using the von M ises-H ill plasticity m odel w ith kinem atic hardening from the ABAQUS m aterial library. The Y oungs M odulus o f the 316L was 193 GPa, the Poisson ratio w as 0.3 and the yield stress w as 205 M Pa. The m aterial o f the arterial w all and the plaque w ere m odelled as non

linear, isotropic hyperelastic m aterials using a third order M ooney-R ivlin m odel obtained from experim ental data. The effects o f the expansion o f the stent w ere analysed in term s o f contact pressures, stresses and radial displacem ents. The

sim ulation results showed that the stresses w ere concentrated in the contact areas o f the plaque w ith the stent. They also reported that the pressure required to

expand the stent increases w ith the stiffness o f the plaque. Therefore the am ount o f pressure required is dependent upon the m aterial properties o f the plaque. Petrini et al. [48] also evaluated the effect o f different materials, com paring a 316L stainless steel stent w ith a self-expandable Shape M em ory A lloy (SM A) stent (N ickel-Titanium alloy). The sim ulations consisted o f a stent m odel resem bling the Palm az Schatz stent. SM As have the unique ability to

recover their original shape (superelastic effect) w hen m echanically deform ed to a m uch greater extent than conventional metals. The com parison o f the two

analyses showed that the stainless steel stent induced a 50% higher m agnitude o f von M ises stress at the m axim um expansion com pared w ith the SM A stent. The low er s ffhess o f the SM A, how ever, resulted in the stent having a low er ability to expand and to contrast the vessel elastic recoil. Petrini et al. [30] continued their num erical analyses to analyse the behaviour o f new generation stents. Two different stent m odels w ere sim ulated resem bling the Cordis B X -V elocity (CV) (Johnson & Johnson) and the C arbostent Sirius (SC) (Sorin B iom edica) coronary stents. B oth m odels consisted o f sim ilar tubular-like rings and the rings w ere connected w ith different bridging links. The stents w ere m ade o f 316L stainless steel and m odelled using a von M ises-H ill plasticity m odel w ith isotropic hardening. The m eshes w ere generated by

G AM BIT com m ercial code (Fluent Inc., Lebanon, N H , USA ) and large

35

deform ation analyse w as perform ed using the A B A Q U S com m ercial code.

com parison o f the num erical simulations indicated that the links o f the stent m odel resem bling the Cordis B X -V elocity show ed a greater capability to deform independently from the rings during bending. The stresses w ere found to be

concentrated in the links and increase in a continuous w ay as the material plastically deformed, w hile the rings rem ained unstressed. This is due to the w ay the link m em bers are connected to the ring structure. In the stent m odel

resem bling the C arbostent Sirius stent, the stresses developed both in the links and in the segm ents o f the rings connected to the links. The C arbostent Sirius stent structure was found to be stiffer than the Cordis B X -V elocity stent. Recently, Lally et al. [31] investigated the m echanical behaviour o f two stent designs, N IR stent and S7 stent, on the biom echanical interaction between the stents and the artery during stenting procedure. These analyses w ere A third

perform ed using M arc/M entat (M scSoftware, Santa A na, CA, USA).

order M ooney-R ivlin m odel w as used to m odel the atherosclerotic coronary vessel geom etry for the artery, incorporating arterial properties from uniaxial and equibiaxial tensile test o f hum an femoral arterial tissue and calcified plaque tissue properties from uniaxial tensile test on cellular hum an calcified plaque. The

exam ination o f the arterial w all stresses and the com puted am ount o f volum e stress in the stented arteries by the N IR and the S7 indicated that the m ore flexible S7 stent design w ould cause lower stresses than the N IR stent design. This result w as found to correlate w ith the findings o f clinical studies, w hereby the S7 has been reported to be less likely to cause restenosis than the NIR. The m ain

limitations o f this study w ere the idealised geom etrical representation o f the stenosed coronary artery and the fact that the stents w ere m odelled in their expanded shape, w hereby the process o f stent expansion w as not m odelled. H olzapfel et al. [32] evaluated the difference in the m echanical environm ent w ithin the arterial w all by a vessel stented w ith three different stent designs; the M ulti-Link-Tetra stent, the N IRO Y A L E lite stent and the InFlowG old-Flex stent. The com putational m odels included a com plex stenosed arterial vessel geom etry obtained from high resolution m agnetic resonance images (hrM RI) o f a hum an iliac stenotic artery. A n anisotropic constitutive m odel was used to represent the arterial tissue properties. T hey exam ined the perform ance o f

36

three

different

stent

designs in

term

o f lum inal

gain,

intim al pressure

concentration and the circum ferential stress changes w ithin the arterial wall caused by the stent. G eneral observation found that as lum inal gain increased, the intim al pressure concentration and the stresses w ithin the vessel rapidly increased. The M ulti-Link-Tetra stent was found to be the least likely to cause restenosis. It w as found to cause h a lf the m agnitude o f stress induced w ithin the arterial w all by the InFlow -G old-Flex stent w hen the M ulti-L ink-T etra stent gained 90% o f the lum inal gain o f the InFlow -Gold-Flex stent. This analysis is probably the m ost com plete to-date since it includes the anisotropy o f the tissue and the realistic vessel geometry. H owever, the m ore sim plistic representation o f the arterial

tissue and artery geom etry in other studies have also successfully identified key characteristics o f stents that m ay influence restenosis outcome.

2.7 Conclusion
It m ay be concluded that stent design is the k ey determ inant in the propensity o f stents to cause restenosis. Stent design is responsible for the different levels o f vascular dam age w ithin stented vessels leading to the growth o f neointim al hyperplasia and restenosis. Clinical studies have dem onstrated that different stent designs cause different levels o f restenosis. Researchers have carried out FE

studies in an attem pt to understand the process o f stenting procedures by sim ulating the expansion o f stents and their interaction w ith arterial vessels by sim ulating their expansion inside stenotic coronary artery m odels. M ost o f FEM studies focused on the expansion characteristics o f stents after deploym ent in free expansion, w ithout the influence o f the stenotic coronary artery. The results

focused on the influence o f the stent geom etry on the stent expansion behaviour. In sim ulations o f the free expansion o f stents, the absence o f the external force on the stent structure caused by the arterial w all structure presents a m ajor lim itation o f these studies. W hilst these studies have concentrated on the geom etrical

characteristics o f specific stent designs, the resulting geom etrical results, such as the structure recoil, from these free expansion sim ulations are likely to be different w ith the influence o f an arterial vessel. H ow ever, there are lim ited FEM studies incorporating the influence o f stenotic coronary artery due to the

37

com plexity o f m odelling the procedure.

A nalyses that have incorporated the

arterial w all have generally exam ined the contact pressures and stresses w ithin the vessel induced by the stent and the radial displacem ents w ithin the artery. The

stress concentrations w ithin the stented vessels have generally been interpreted as a m easure for vascular injury. This study focuses on determ ining the am ount o f vascular injury caused b y different stent designs. The ISA R -STEREO clinical Trial [26] indicated that the use o f a thinner-strut stent is associated w ith a significant reduction o f angiographic and clinical restenosis than the thicker-strut stents. Therefore it is hypothesised that the thicker-strut stents induces higher stresses on the arterial wall and that these high stresses m ay be responsible for the injury stim uli for growth o f neointim al hyperplasia. The stresses induced in the arterial w all will therefore be determ ined for the different stent designs and a com parison to clinical data w ill be used to determ ine m ore conclusively the m echanical stim ulus for in stent restenosis.

38

Chapter 3

MATERIALS AND METHOD

3.1 Introduction

The purpose o f this study is to identify the m echanical stim uli for in-stent restenosis. This m ay be achieved by investigating the m echanical cause o f the difference in restenosis outcom e b y thinner and thicker strut stents following the ISA R-STEREO clinical trial. Finite elem ent m odels o f these stents were

developed and the stents w ere deployed in various stenosed vessel geom etries in order to achieve this objective. The stents w ere expanded and deployed w ithin the arteries to scaffold open stenosed arteries, in line w ith that w hich w as carried out in the clinical trial. To generate these num erical sim ulations, the finite elem ent method requires a num ber o f inputs; the geom etry o f the stents and the various stenosed coronary arterial vessels, the m aterial properties o f the stents, the plaque and the artery, and the appropriate application o f loading and boundary conditions. The finite elem ent m odels o f the stents w ere developed in A N SY S (Canonsburg, PA, U SA ), a finite elem ent software package. The properties o f 316L stainless steel were assigned to the stents material. The m aterial behaviour o f the artery was

obtained from available published data on porcine coronary arterial tissue. The m aterial m odel for the plaque was obtained from available published data on hypocellular hum an atherosclerotic plaques. The finite elem ent analyses w ere

carried out to com pare the stresses induced w ithin the stented artery vessels by

39

these stents, and hence used to determ ine the level o f vascular injury caused to the artery b y different strut thickness stents.

3.2 Modelling of Stent Geometry

In finite elem ent sim ulations, it is extrem ely im portant to realistically m odel the problem . To realistically represent stent deploym ent, full three-dim ensional

m odels o f the stents had to be developed. The m odels w ere generated and m eshed using A NSY S and the elements, nodes and their connectivities w ere transferred into M arc/M entat (M scSoftwarc, Santa A na, CA, U SA ) for solving. A num ber o f finite elem ent packages have been used for m odelling the deploym ent o f intravascular stents, such as A BA Q U S [27, 29, 30, 48, 86], A N SY S/LS-D Y N A [28, 52, 84] and M arc/M entat [31]. M arc/M entat was chosen for solving the analyses in the current project because o f its capability o f handling highly non-linear and large deform ation m aterial behaviour. The non-linear

regression routine, available in M arc/M entat, w as used to obtain the hyperelastic constitutive m odel that best fit the uniaxial and equibiaxial published data and was found to be effective in obtaining a stable m aterial. It has a robust contact

algorithm that is suitable for defining the contact betw een the stent, plaque and artery. M arc/M entat has m esh adaptivity capability w hich is highly advantageous for the contact analyses used in this study since it enables autom atic m esh refinem ent o f the arterial element, as contact occurs betw een the stent and the surface o f the artery. This m esh refinem ent w as needed for accurate stress

analysis o f the stenosed artery and this function w as found effective in m inim ising penetration, thus obtaining convergence. Cyclic sym m etry is another option Circum ferential

available in M arc/M entat w hich w as used in these analyses.

sym m etry w as found in the stents design and the cyclic sym m etry capability in M arc/M entat allow ed one third in the circum ferential direction to be m odelled as apposed to the full stent. The num ber o f elem ents used for the analyses was

therefore m inim ised and hence the com putational tim e and resources required w ere reduced.

40

The stent geom etries in the finite elem ent m odels w ere based on the stents used in the ISA R STEREO Trial [26]. They w ere the A CS R X M ultiLink and the ACS M ultiLink RX Duet. B oth stents are m anufactured by the same

com pany, nam ely G uidant/A dvanced C ardiovascular Systems. B oth stents have an interconnected-ring design and different strut thickness. These stents have

sim ilar design w ith a strut w idth o f 0.1 mm. The strut thickness o f the ACS RX M ultiLink (M linkthin) is 0.05 mm and the strut thickness o f the A C S M ultiLink RX D uet (Mlinkthick) is 0.14 mm. They have a square cross-sectional area. Full three-dim ensional m odels o f the stents w ere developed to determ ine the expansion characteristics o f the stents after stent deploym ent. The m ain

m easurem ents o f the stents w ere extrapolated from the handbook o f coronary stents [89]. B oth stents are available in a range o f lengths, w hilst the length o f the stents investigated in this study w as 7.2 mm. For this length, the A CS M ultiLink RX D uet stent consists o f six rings in the longitudinal direction w ith six crow ns in each ring, see Figure 3.1. The rings are connected to one another at three different evenly spaced locations and these locations are offset from one another along the length o f the stent as shown in Figure 3.1. For both o f the finite elem ent m odels o f the stents, all the param eters, such as the length and w idth w ere kept constant and the only variation betw een the tw o designs w as the thickness o f the stent struts. A strut thickness o f 0.05 m m w as assigned to the m odel o f the ACS RX M ultiLink (Mlinkthin) w hilst a strut thickness o f 0.14 m m w as assigned to the ACS M ultiLink R X D uet (M linkth,ck)- B oth stents w ere sim ulated w ith an inner radius o f 0.72 m m and length o f 7.2 m m w ith a corresponding outer radius o f 0.77 mm for the Mlinkthin and 0.86 m m for the M linkthiC k, see Figure 3.2. The stent does not show sym m etry in the longitudinal direction however, sym m etry is observed in the circum ferential direction. D ue to the circum ferential sym m etry, only one third in the circum ferential direction w as m odelled w ith the full length in the longitudinal direction. The stent geom etry was initially

m odelled in three-dim ensional Cartesian coordinate system , representing the stent in an opened-out, planar configuration. To generate the planar geom etry o f the stents, line profiles that represented the skeleton o f the stent w ere created and a square area w as subsequently extruded through the lines profiles to generate the planar geom etry o f the stent volum e, see Figure 3.3.

41

Figure 3.1

Expanded ACS M ultiLink R X D uet [adapted from [90].

r i thin = 0.72 m m r 0 thin = 0.77 m m thickness (s thm ) = 0.05 m m

r i thick = 0.72 mm r o thick = 0.86 m m thickness (s thick) - 0.14 m m

Figure 3.2

Schem atic illustrating the radial dim ensions o f the M linkthin and Mlinkthick m odels.

42

Figure 3.3

Volum es o f (a) M linkthin and (b) M linkthiC k in the C artesian coordinate system and the m eshed m odel o f (c) Mlinkthin and (d) Mlinkthick.

The volum es o f Mlinkthin and Mlinkthick were discretised by m eans o f eight-noded isoparam etric, three-dim ensional brick elem ents. The m odels were

m eshed w ith one elem ent through the thickness, see Figure 3.3. B oth stents were m eshed w ith a total o f 532 elem ents w ith corresponding 2010 total num ber o f nodes. The elem ents w ere checked for their shape to ensure that no distorted

elem ents were generated. Contact analysis for M arc/M entat requires a finer m esh on the m ost deform able body o f the contact bodies, i.e. the stenosed artery in this case. Therefore this m esh density o f the stent was chosen to m inim ise the num ber o f elements and was found reasonable. A finer m esh o f the stent structure w ould lead to a finer m esh density o f the artery and m uch greater heavy com putational tim e and resources. The nodal coordinates o f the m eshed m odel were transferred from a Cartesian coordinate system into a cylindrical coordinate system, using a procedure reported by Lally et al. [31], w hereby the planar configuration was w rapped to represent the cylindrical structure o f the stents.

43

The elements, nodes and their connectivity w ere then transferred into the finite elem ent code M SC M arc M entat and large deform ation analyses w ere solved to sim ulate the expansion o f the stents. A lthough A N SY S was used to

m odel the stents geom etries, it was found that it w as not suitable for solving the analyses w hich included contact betw een the stent, plaque and artery and the use o f hyperelastic m aterial m odels to define the plaque and the arterial wall. Four case studies w ere carried out for each o f the stent designs to investigate the influence o f strut thickness on stent deployment. The influence o f variations in strut thickness on the stresses that w ere induced w ithin arteries during and after stent deploym ent b y the stents struts w ere also investigated. studies carried out w ere as follows:The four case

I. Case study 1: Free expansion o f M linkthm and MlinkthickThis study investigated the

perform ance o f the stents in the absence o f an artery.

II. Case study 2: Expansion o f M linkthin and Mlinkthick inside a straight stenotic vessel geom etry with arterial m aterial properties. The sam e initial expanded lum en diam eter, 0 /, w here 0 / = 0 / thiCk w as achieved for both stents. 0 / is

achieved w hen pressure is applied, see Figure 3.4.

III.

Case study 3: Expansion o f Mlinkthin and M linkthjck inside a straight stenotic vessel geom etry w ith arterial m aterial properties as in Case study 2. However, the sam e final expanded lum en diameter, 0/-, w here 0 F thin = 0 / - thick was achieved for both stents in this case. com pletely rem oved, see Figure 3.4. 0 F is achieved w hen pressure is

IV.

Case study 4: Expansion o f Mlinkthin and Mlinkthick inside a localised stenotic artery vessel geom etry w ith plaque and arterial m aterial properties assigned to the two

44

com ponents o f the arterial wall. The sam e initial expanded lum en diameter, 0 /, w here & ithm = 0 / thick was achieved for both stents.

The output geom etric quantities and calculated param eters o f interest during the analyses are sum m arised in Table 3.1. These values w ere taken from the stents structures for the unexpanded stents configurations, at loading and on rem oval o f the expansion pressure at unloading during stent deploym ent for all four case studies.

Stent before expansion Rorig = original radius Lorig = original length

Stent after loading Rioad proximal = radius o f stent at the proxim al end o f stent Rioad distal = radius o f stent at the distal end o f stent Rioad central = radius o f stent at the centre o f stent Lioad = length at the end o f loading

Stent after unloading Runioad proximal = radius o f stent at the proxim al end o f stent
R u n io a d d ista l

= radius o f stent at the distal end o f stent

Runioad central = radius o f stent at the centre o f stent

L u n io a d

= length at the end o f unloading

Calculated param eters

D _ U

Proxim al radial recoil = ---------------------------------D

load proximal

i-i

-t

1V load proximal

A V unload proximal

Central radial recoil = R ladce."gal 0

1 v load central

45

D istal radial recoil - R]oaddistai " 'n

load distal

Longitudinal recoil = L 'oad ~Lmloai^'load

t-

--------Foreshortening = -----5
^ o rig

^'orig

^ lo a d

D o g b o n in g

R load distal______ - R load central

^ load distal

Table 3.1

List o f geom etric data taken follow ing the expansion o f

Mlinkthin and MlinkthiC k for all four case studies.

0Orig [A] Stent before expansion

[B] Stent after loading

[C]

Stent after unloading

L u n lo a d

Central Figure 3.4 Schem atic illustrating the stents expansion at (A) stent before expansion, (B) stent at loading w hen pressure is applied and (C) stent at unloading w hen pressure is rem oved and (D) illustrates the division o f interest.

46

The von M ises stress, cr , and the equivalent plastic strain, eqv, w ere evaluated for the expanded stent structures. These quantities are defined as follows:-

a =

[(cr,-aJ2+(a2-a-3 )2+(0-3- 0,)2r

V2

where, cr,, cr2 , cr3 are the principal stresses.

w here, e pl is the plastic strain rate tensor.

3.2.1

Case study 1: Free expansion of M linkthi and Mlinkthick*

Stent deploym ent involves expansion o f the stent by th e application o f an internal pressure (loading), and rem oval o f the internal pressure (unloading). Initially in this study, the stents w ere subjected to free expansion to investigate their expansion under loading w ithout the influence o f the artery, see Figure 3.5. A uniform , linearly increasing radial pressure (P) w as applied as a surface load to the internal surface o f the stent until a radius o f 3 m m w as achieved in the central region o f the stent. The pressure required to expand M linkthin was 0.437 M Pa and the pressure required to expand Mlinkthick w as 1.3 M Pa, see Figure 3.6. The pressure w as ram ped for a tim e step size o f 0.01. This w as an iteration process w hereby the static analyses involve the application o f pressure to achieve the desired stents expansion. The pressure w as unloaded by h a lf the tim e it took to load it. Too high o f a tim e step size results in too much pressure being applied in one increm ent over tim e and this m ay cause too m uch deform ation resulting in non convergence o f the solution.

47

Cyclic sym m etry boundary conditions w ere im posed on the nodes o f the stent in the circum ferential plane o f symmetry. The top and bottom ends o f the stent were free from any constraints. Two nodes in the centre o f the stent w ere constrained in the circum ferential direction to prevent rigid body rotation o f the stent.

Figure 3.5

Finite elem ent m eshes o f unexpanded M linkthm and Mlinkthick, used for case study 1.

P re s s u re P a th o f M lin k th ln a n d M lin k th ic k

I--m llnkthin

m lin kth ic k

Figure 3.6

P lot o f pressure path o f M linkthin and Mlinkthick-

48

3.2.2

Case study 2 and case study 3: Expansion of Mlinkthin and Mlinkthick inside various stenosed coronary artery vessels.

Three-dim ensional sim ulations o f stent deploym ent w ithin various stenosed vessel geom etries w ere developed to determ ine m ore realistically the expansion o f Mlinkthin and Mlinkthick and to quantify the stresses induced w ithin stented stenotic coronary arteries by Mlinkthm and MlinkthickFor case study 2 and case study 3, the sam e finite elem ent m odels w ere simulated, w ith the only difference being the am ount o f pressure applied. For

case study 2, 8.2 M Pa was applied to expand M linkthin and 10.5 M Pa w as applied to expand Mlinkthick, to achieve the same initial vessel lum en diameter. For case study 3, the pressure required to expand Mlinkthin and Mlinkthick w ere 8.2 M Pa and 2.2 M Pa respectively, to achieve the sam e final vessel lum en diam eter. The

pressure w as applied as a uniform, linearly increasing surface load and was applied to the internal surface o f the stent. The pressure w as rem oved linearly in h a lf the tim e o f loading, to the pressure o f 13.3 kPa, w hich corresponds to m ean blood pressure o f 100 mm Hg. A full m odel o f the stent was expanded inside a stenotic straight vessel and it was found to be highly com putationally expensive. The results w ere

analysed and it was found that the m ovem ent in the m idplane w as m inim al, see A ppendix A. Therefore to save com putational tim e and resources, the full m odel w ere divided into two h a lf models; a proxim al h a lf and a distal h a lf o f the stented vessels and each h a lf was analysed separately and com pared for each stent design. As a result o f the necessity to divide the m odel in tw o, four sim ulations had to be carried out to com plete case study 2 and case study 3, see Figure 3.6. Each sim ulation m odel com posed o f two bodies, the stent and the stenotic coronary artery. The geom etrical m odels for the tw o stent designs rem ained the same. The stenotic coronary artery w as m odelled as a straight vessel geom etry o f internal radius o f 1 m m and thickness o f 0.8 m m , see Figure 3.7. The thickness o f atherosclerotic hum an coronary arteries range from 0.56 mm to 1.25 mm, depending upon the location o f the arteries on the surface o f the heart [37]. For

49

this reason, a thickness o f 0.8 m m was chosen to represent the stenosed coronary artery. A rterial properties w ere assigned to the stenosed vessel. D ue to cyclic

symmetry, only one-third o f the artery and stent w ere represented in the model. Five elem ents w ere assigned through the thickness o f the vessel, thirty elements w ere assigned in the circum ferential length and forty-eight elem ents w ere assigned in the longitudinal length. stenotic coronary artery. In addition, adaptive m eshing was used to ensure m inim al penetration occurred as the stent cam e into contact w ith the stenotic artery as the edge o f an elem ent o f one body w ould penetrate the other body i f the m eshing w as not sufficiently fine to prevent penetration. Penetration during stent deploym ent There w ere 7200 elem ents in total for the

sim ulations can be a significant problem since the relatively thin struts o f the stents contact the artery in very localised areas and have to support the w hole arterial structure. Local m esh adaptivity w as assigned to the elem ents w here the stent cam e into contact w ith the stenotic artery, i.e. the inner layer o f the artery wall. As a result, these associated elem ents w ere subdivided b y tw o in all three coordinate directions resulting in a substantial refinem ent o f the m esh and hence an im provem ent in the accuracy o f the solution. W hen adaptive m eshing occurs, to ensure com patibility, the new nodes created are internally tied to the neighbouring nodes. F or exam ple, to ensure

com patibility node E is effectively tied to node A and node B, and node F is effectively ties to node B and C, see Figure 3.8. This occurs internally and does not conflict w ith other use-defm ed ties or contact. These new nodes are

autom atically inherent the sam e boundary conditions as the neighbouring nodes tying the new generated nodes. In the case o f the new nodes generated on the exterior o f a contact body are autom atically treated as potential contact nodes and they are checked to determ ine if they are in contact.

50

Stenotic vessel Stent

Proxim al h alf

D istal h a lf Mlinkihii

Proxim al h a lf

D istal h a lf

Mlinkthick

Figure 3.7

Finite elem ent m esh o f unexpanded M linkthin and M linkthiC k inside a stenotic coronary artery, w ith an inner radius o f 1 m m and outer radius o f 1.8 mm, used for case study 2 and case study 3.

C Figure 3.8

Illustration o f discontinuities created as local m esh adaptivity occurs.

3.2.3

Case study 4: Expansion of Mlinkthjn and Mlinkthjck inside localised stenotic coronary artery vessels.

For case study 4, tw o sim ulations were carried out to expand the distal h a lf o f Mlinkthin and Mlinkthick inside a localised stenotic coronary artery vessel, see Figure 3.9. The pressure used to expand Mlinkthin was 41 M Pa and 45.7 M P a for

Mlinkthick, to achieve the same initial lum en diam eter. The pressure w as applied
as a uniform , linearly increasing surface load on the internal surface o f the stent. The pressure w as rem oved linearly in h a lf the tim e o f loading, to the pressure up to 13.3 kPa, w hich corresponds to m ean blood pressure o f 100 mmHg.

I
Mlinkthin

A rtery

Plaque

Stent Mlinkthick

Figure 3.9

Finite elem ent m esh o f unexpanded Mlinkthin and Mlinkthick inside a localised stenotic vessel, w ith an inner radius o f 1 m m and outer radius o f 1.8, used for case study 4.

52

The vessel incorporated arterial and plaque properties.

The stenosed This was

coronary artery w as m odelled as a vessel w ith a localised stenosis.

represented by a localised crescent-shaped axisym m etric stenosis w ith an internal radius o f 1 mm and outer radius o f 1.8 mm. Six elem ents w ere assigned through the thickness o f the artery, thirty elem ents w ere assigned in the circum ferential length and forty-eight elements w ere assigned in the longitudinal length. The

healthy arterial tissue o f the artery w as m eshed w ith 4320 elem ents, w hilst the plaque was m eshed w ith 3270 elements. In total the localised stenotic coronary artery vessel w ere discretised by 7590 elem ents, see Figure 3.9. F or this case study, adaptive m eshing w as again applied to the inner layer o f the artery to allow element refinem ent upon contact w ith the stent during stent deployment. For case study 2, 3 and 4, the elem ent type used for the stenotic coronary artery w as a full integration, three-dim ensional eight-node isoparam etric arbitrary hexahedral elem ent (Type 7 in M arc/M entat). Cyclic sym m etry boundary

conditions w ere im posed on the nodes o f the stent and stenotic coronary artery in the circum ferential plane o f symmetry. This cyclic sym m etry condition enables the nodes on the two cyclically sym m etrical faces to be coupled or tied. A ll nodes o f the stent w ere constrained in the longitudinal direction along the central plane o f the stent to represent the h a lf sym m etry condition. These nodes w ere the nodes that w ould have connected to the other h a lf o f the m odel. B oth ends o f the

stenotic artery w ere constrained in the longitudinal direction to represent the longitudinal tethering. One node on the outer surface o f the stenotic artery was constrained in the circum ferential direction to prevent rigid body rotations. In cases 2, 3 and 4, surface to surface frictionless contact w as assigned betw een the stent and the stenotic artery. U sing this procedure, the m otion o f the contact bodies are tracked, and w hen contact occurs direct constraints are placed on the m otion using boundary conditions - both kinem atic constraints on transform ed degrees o f freedom and nodal forces. D eform able-deform able

contact was also used to describe the contact betw een the stent and th e stenotic artery. The contacting surfaces o f both contact bodies w ere represented as The description o f the contact bodies using NURB

analytical N U R B surfaces.

surfaces enables recalculation o f these surfaces as the stent and artery deform and

53

as a result this technique represents the deploym ent procedure m ore accurately than the default discrete technique w here the boundary o f the contacted body is described b y the finite elements that the body is m ade up. This can cause

problem s due to the fact that the norm als o f the body are not continuous for a curved boundary. As a result the analytical NURB surfaces technique leads to a better convergence.

3.3

Stent Material: 316L Stainless Steel

The A CS R X M ultiLink and the ACS M ultiLink R X D uet stents are m ade o f 316L stainless steel [89]. 316L stainless steel is a com m on m aterial used for

stents and its bulk properties are w ell known. H ow ever, an experim ental study w as carried out b y M urphy et al. [88] to investigate the m echanical behaviour o f 316L stainless steel stent struts to determ ine the influence o f strut size on the m aterials behaviour. D uring stent

deploym ent, the stent m ust undergo plastic deform ation to ensure that it rem ains in its expanded configuration and can therefore scaffold open the artery. U pon deploym ent, stents m ay undergo as m uch as 20% -30% plastic strain. The study by M urphy et al. [88] dem onstrated that a size dependent stress-strain relationship m ust be used to describe the tensile behaviour o f the 316L stainless steel for the size scale o f coronary stent struts particularly as the breaking strength o f 316L stainless steel is clearly size dependant, see Figure 3.10. The stress-strain

relationship o f 316L stainless steel for M linkthin (strut thickness o f 0.05 m m ) and

Mlinkthick (strut thickness o f 0.14 m m ) in the finite elem ent m odels in this study
m ay therefore be described by the m echanical behaviour o f the struts tested by M urphy et al. [88] as shown in Figure 3.10. The m aterial was described as an isotropic m aterial w ith the linear elastic region o f the curve defined through the m aterial values for 316L stainless steel; Y oungs M odulus o f 196 GPa, P oissons Ratio o f 0.3. W hen the stress reaches the yield stress, the m aterial no longer exhibits elastic behaviour and the stress-strain behaviour becom es non-linear. A piecew ise linear function was used to represent this non-linearity through a von M ises

54

--Thick (0.14 mm) --Thin (0.05 mm)

Engineering Strain

F ig u re 3.10

Plot o f stress-strain relationship o f 316L stainless steel for Mlinkthin (0.05 m m ) and M linkthiC k (0.14 m m ), [adapted from M urphy et al. [88]].

plasticity m odel w ith isotropic hardening. In a uniaxial test, the w orkhardening slope is defined as the slope o f the stress-plastic strain curve. The w orkhardening slope relates the increm ental stress to increm ental plastic strain in the inelastic region and dictates the conditions o f subsequent yielding. The V on M ises yield criterion ensures that yield occurs w hen the equivalent stress equals the yield stress initially defined. The uniaxial stress-plastic strain curves o f 316L stainless steel for the two strut thickness stents were represented by a piecew ise linear function, as described by the curves show n in Figure 3.10. The yield stress and the w orkhardening data input into the finite elem ent code had to be com patible w ith the procedure used in the analysis. In M arc

M entat, w hen LA RG E D ISPLA CEM EN T, U PD A TE, FIN ITE param eters or PLA STIC ITY are used, the yield stress m ust be defined in term s o f Cauchy stress, and the w orkhardening data w ith respect to logarithm ic plastic strains. The LA RG E D ISPLA C EM EN T option was used for this large

deform ation non-linear analysis. A LA RG E STRAIN U PD A TED LA G RA N G E

55

form ulation was used w hereby the elem ent quantities w ere evaluated w ith respect to the current updated deform ed configuration. The FIN ITE param eter w as

allow ing the effects o f the change in m etric due to large inelastic deform ation to be included. This resulted in a different stiffness o f the structure as w ell as in a m odified calculation o f stresses and inelastic strains. In analyzing the m echanical perform ance o f the stents, the m aterial behaviour is o f critical im portance in acquiring accurate sim ulated studies. Once the stent is delivered to the site o f lesion, it is plastically expanded. plastically deform ed structure rem ains as it supports the arterial wall. The It is

therefore essential to accurately m odel the plastic region o f the stress-strain response o f the m aterial after the yield stress is reached. V alidation tests w ere carried out to ensure accurate sim ulations o f the m aterial behaviour o f 316L stainless steel. This w as achieved by sim ulating a uniaxial test o f a rectangular 316L specim en and the results o f the simulation confirm ed that the stress-strain behaviour o f the m aterial in the finite elem ent m odels w as the same as that represented in Figure 3.10.

3.4 Artery Material: Porcine Coronary Artery

A second order M ooney-R ivlin hyperelastic constitutive equation w as used to represent the non-linear stress-strain relationship o f the arterial wall. A M ooneyR ivlin hyperelastic m aterial is defined b y a strain energy density function, W given in Eq. (1). The hyperelastic constitutive equation w as determ ined by fitting to available published data from uniaxial and equibiaxial tension tests o f porcine coronary tissue [41]. A non-linear regression routine, available in M arc/M entat, w as used to obtain the hyperelastic constitutive m odel that best fit the uniaxial and equibiaxial data. The least-squares error w as m inim ised during data fitting, and in this case was based on absolute error defined as follows:-

A bsolute error =

data m easured(i) - data calculated(i)]2

56

The uniaxial and biaxial data fits w ere checked for positive definiteness and the least squares fit yielded a strain energy density function describing the stress-strain curves show n in Figure 3.11. The regression routine fitted both the uniaxial and biaxial data sim ultaneously that led to the slight deviation in the stress-strain curves. This was found as a good fit com paratively w ith other A

m odels, w ith the low est error value o f 0.487657 associated w ith this fit.

Signiorini m odel (second order m odel) was found to best fit to the data as given by the follow ing hyperelastic constitutive equation:

W = 708.416 (Ii - 3) - 620.042 (I2 - 3) + 2827.33 (Ii - 3)2

(kPa)

w here W is the strain energy density function o f the m aterial and Ii and h are the strain invariants. The second order m odel was found to best fit the chosen

uniaxial and biaxial datas despite the third order model that was used by Lally et al. [41]. The resulting hyperelastic constitutive equation is specific to the stressstrain curves from a specific sample.

| uniaxial signiorini uniaxial exp biaxial exp

biaxial signiorini

Strain

F ig u re 3.11

Plot o f stress-strain uniaxial and biaxial data for porcine coronary tissue [41].

57

3.5

Plaque Material: Hypocellular Human Atherosclerotic Plaques

A third order M ooney-R ivlin hyperelastic constitutive equation was used to represent the non-linear stress-strain relationship o f hum an atherosclerotic plaque. The hyperelastic constitutive equation w as determ ined b y fitting to available published data on uniaxial tensile tests carried out on hum an atherosclerotic plaques [54]. A gain, using the non-linear regression routine in M arc/M entat, the m odel that best fit the uniaxial data w as determined. The least-squares error was m inim ised during data fitting, and in this case was based on relative error defined as follows:-

R elative error = > [1 -

data calculated (z)

The uniaxial data fit w as checked to ensure that it yielded curves that w ere positive definite and that w ere a best fit to the data w ith a low error value. A third order deform ation m odel w as found to best fit to the data and the hyperelastic constitutive equation is given by:

W = -677.134 (Ii - 3) + 977.734 (I2 - 3) - 296180 (Ii - 3)2 + 326979 (Ii - 3) 02 - 3) + 148343 (Ii - 3)3 (kPa)

w here W is the strain energy density function o f the m aterial and I] and h are the strain invariants. The uniaxial and biaxial behaviour described by this strain The fit to the data was

energy density function are shown in Figure 3.12. achieved w ith an error value o f 0.123461.

58

--uniaxial exp --uniaxial 3rd order def

biaxial exp

Strain

F ig u re 3.12

Plot o f stress-strain uniaxial and biaxial data for hypocellular hum an atherosclerotic plaque.

3.6

Summary

This chapter has detailed all o f the m ethods used in this study to investigate the perform ance o f different strut thickness stents using the finite elem ent method. The num erical m odels outlined enable the influence o f stent design to be analysed in term s o f the expansion characteristics o f the stent. The m odels also include the im portant property that the vessel w all is non-linearly elastic and they m ay therefore be used to investigate the loading that different stents im pose on different arterial geom etries during their expansion. In this way, the m odels

enable the stresses induced in stented arteries to be estim ated and a greater insight into the m echanical stim uli for restenosis to be gained.

59

Chapter 4

RESULTS

4.1

Introduction

The following chapter outlines all o f the results obtained for the four case studies exam ined during this thesis to investigate the influence o f stent strut thickness on stent expansion and vessel w all stresses. B oth stents, M linkthin and M linkthiC k,

w ere subjected to free expansion and later to expansion inside tw o different vessel geom etries. The vessel w all stresses w ere exam ined w here stents w ere expanded to achieve the same initial expanded lum en diam eter, 0 / and also for cases w here the stents were expanded to the same final expanded lum en diameter, 0 F. The stresses induced w ithin the vessels w ere studied for all o f the load cases and for both stents.

4.2

Case Study 1: Free Expansion of Mlinkth in and Mlinkth ick

Mlinkthin and Mlinkthick w ere expanded to achieve an outer diam eter o f 3 mm, w hich corresponds to a typical diam eter o f a coronary artery, see Figure 4.1. The pressure required to expand Mlinkthin was 0.437 M P a and the pressure required to expand Mlinkthick was 1.3 M Pa. The radial expansion distribution w as found to be sim ilar for both Mlinkthin and MlinkthickThe radial displacem ent distribution throughout both

60

F ig u re 4.1

D eform ed geom etry o f M linkthin and M linkthiC k under free expansion, achieving an initial outer diam eter o f 3 mm.

m odels was found to be highly non-uniform w ith the low est radial displacem ent observed at the proxim al end o f the stents structures, see Figure 4.2. Radial displacem ent m easurem ents were obtained at key geom etrical m arkers along the lengths o f both stents in order to allow direct com parison o f their radial expansion, see Figure 4.3 and Table 4.1. The proxim al end

experienced the least radial displacem ent, followed by distal end 2, w hilst the central region and distal end 1 w ere found to undergo sim ilarly high radial displacem ents, see Figure 4.4.

mm 0.78
0.68

0.57 0.46 0.36 0.25

M linkthin

mm
0.68

0.59 0.50 0.40 0.31 0.22 M lin kthick

Figure 4.2

R adial displacem ent distribution throughout the stent structure o f Mlinkthin and M linkthiC k, subjected to free expansion,

achieving an initial outer diam eter o f 3 mm.

C entral

Distal_2

Figure 4.3

The proxim al, central and distal points o f interest.

62

Mlillktl.in

M lin k t],ick

Stent before expansion

R o rig Long

(m m ) 0.77 7.2

(m m )

0.86
7 .2

Stent after loading

R -load p ro x im al R -lo ad ccn lral R lo a d d ista l 1 R lo a d d istal_ 2 L lo a d

1.06 1.43 1.42 1.23 7.31

1.1
1.48 1.47 1.37 7.35

Stent after unloading

R u n load p ro x im al R u n load c cn lra l ^ u n l o a d d ista l 1 R u n lo a c l d isla l_ 2 L im k o a d

0.9 9 1.36 1.38 1.18 7 .2 9

1.05 1.43 1.42 1.33 7.33

Calculated parameters L o n g itu d in a l re co il F o re sh o rte n in g D o g b o n in g

p roxim o!

(% ) 0.3 -1.5 -3 4 .9 -0.7 -16.3

(% ) 0.3 -2.1 -34.5 -0 .7 -8 .0

D o g b o n in g d istau D o g b o n in g disla| 2

T a b le 4.1

G eom etric data o f M linkthin and M linkthiC k through loading and unloading, subjected to free expansion, achieving an initial outer diam eter o f 3 mm.

63

M linkthin
I loading unloading

Mlinkthick
M loading u nloading!

R proximal

R central

R distal 1 R distal_2

R p roxim al

R central

R d is ta M

R distal_2

F ig u re 4.4

R adial displacem ent o f M linkthm and M linkthiC k through loading and unloading, subjected to tree expansion, achieving an initial outer diam eter o f 3 mm.

The radial percentage recoil for the proxim al end w as found to be 25% higher than the central and distal end for bo th stents. H ow ever, the radial

percentage recoil w as found to be consistently higher throughout the M linkthin stent structure com pared w ith the Mlinkthick, see F igure 4.5.

I mlinkthin m lin kth ick

o a 2 a> Q.

0
R proximal F ig u re 4.5 R central R d is ta M R distal_2

C om parison o f percentage radial recoil o f Mlinkthm and

Mlinkthick, subjected to free expansion, achieving an initial outer

diam eter o f 3 mm.

64

D uring loading the stent structure m ay shorten giving rise to stent foreshortening. N egative foreshortening values w ere found for free expansion o f the stents w hich indicated that the length o f the stent structure elongated during loading. Shortening o f the stent structure can also occur during unloading o f the stents, w hich gives a m easure o f the longitudinal recoil o f the stents. No

significant longitudinal recoil w as observed during the free expansion o f either the Mlinkthin or M linkthiC k stent structures, see Table 4.1. D ogboning has been observed in stents during expansion, w hereby the ends o f the stents radially expand to a larger extent than the central portion o f the stents. D ogboning w as not observed in the M linkthin or M linkthiC k during free

expansion, in fact, a high degree o f radial retraction w as observed at the proxim al ends o f both stents, see Table 4.1. On exam ination o f the von M ises stress contours w ithin the stent structures, it can be seen that the highest stresses are concentrated in the area o f the arcs. The arcs appear to act as plastic hinges allow ing the stent structure to expand, see Figure 4.6. The m axim um von M ises stresses in bo th stents are

concentrated in the arcs and w here the arcs connect to the longitudinal struts o f the stent. The m axim um von M ises stress m agnitude for Mlinkthin w as found to be 604.4 M Pa and for M linkthick w as found to be 596.9 M Pa. The analysis also

shows that the longitudinal struts o f the stents experience low stresses during expansion, see Figure 4.6. The resulting contours o f Total Equivalent Plastic strain throughout the expanded structure o f M linkthjn and Mlinkthick, w ere taken at the end o f loading. H igh plastic strains are evident on the arcs o f the stent structure and the m axim um Total Equivalent Plastic strain was found to be 0.0924 for Mlinkthin and 0.0997 for Mlinkthick, see Figure 4.7.

65

MPa
500 400 300

200

M link,thin

100
0

Mlink.thick F ig u re 4.6 The resulting von M ises structures expansion. stress contours throughout the to free

o f M linkthin and

M linkthiC k5 subjected

0.05 0.04 0.03

0.02 0.01
0

Mlinkthin

M linkthick F ig u re 4.7 The resulting Total Equivalent Plastic strain contours

throughout the structures o f M linkthin and Mlinkthick, subjected to free expansion. 66

4.3

Case Study 2: Expansion of Mlinkth in and Mlinkthick inside a stenotic coronary artery achieving the same initial expanded lumen diameter

Four sim ulations w ere carried out, expanding the proxim al and distal halves o f Mlinkthin and M linkthiC k inside a stenotic coronary artery, represented as an idealised cylindrical vessel. Mlinkthin and Mlinkthick w ere expanded to achieve the same initial expanded lum en diam eter (internal diam eter o f vessel), 0 / = 3 mm, see Figure 4.8. The pressures required to expand Mlinkthin and Mlinkthick w ere 8.2 M Pa and 10.5 M Pa, respectively. The sam e pressure w as applied for both the

proxim al and distal halves o f each structure. The radial displacem ent throughout both structures o f Mlinkthin and Mlinkthick w as found to be highly non-uniform . T he low est values o f radial

displacem ent w ere observed at the proxim al and distal ends o f the stent structures, see Figure 4.9. R adial displacem ents w ere m easured at k ey locations on both the proxim al and distal h a lf o f each stent, see Figure 4.10, in order to com pute the structural characteristics o f the stents through the loading and unloading process o f the stents expansion, see Table 4.2. It can be seen that both stents, Mlinkthjn and Mlinkthick have the same pattern o f radial displacem ent distribution. The radial displacem ent w as found to be in increasing order, as follows: the least radial displacem ent in the proxim al end, follow ed b y the distal end and then the central region o f the distal half. The highest radial displacem ent was found in the central region o f the proxim al h a lf o f the stent structure, see Figure 4.11. Foreshortening w as found to be low in

m agnitude and sim ilar for both Mlinkthin and Mlinkthick, but in contrast to free expansion, the stents did foreshorten to some degree, see Table 4.2. U pon unloading, longitudinal recoil w as found to be low in both stent structures, how ever, the fact that the values are negative indicates that the stents undergo elongation on unloading, unlike in free expansion. R adial retraction o f

67

the stents was observed at the ends o f the stents w hen com pared w ith the radial diam eter o f the central portion o f the stent structures. This negative dogboning was found to be m ore prom inent on the proxim al h a lf o f the stents structures.

Proxim al H alf Stenotic vessel Stent M lm k thill

D istal H alf

Proxim al H alf
Mlinkthick

D istal H alf

F ig u re 4.8

D eform ed geom etry o f the proxim al and distal halves o f Mlinkthin and Mlinkthick, scaffolding a stenotic vessel w ith thickness o f 0.8 mm, achieving the same initial vessel lum en diam eter o f 0 i = 3 mm.

68

mm

Proxim al Mlinkthin mm

D istal

Proxim al

, Mlmkthick

D istal

Figure 4.9

Radial displacem ent distribution throughout the stents structure o f Mlinkthin and M linkthiC k, under the influence o f a stenotic vessel, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm.

Proxim al

Central p roxh f

Central p ro x h f

D istal

Figure 4.10

The proxim al, central and distal points o f interest.

69

Mlink||,i

M link,|liCk

Stent before expansion

R o rig Long

(mm) 0.77 3.6

(mm) 0.86 3.6

Stent after loading

R lo a tl p ro x im al R lo a d centT al p ro x h f R lo a d c en tra l d is h f R lo a d dislal L l o a d p ro x h f L l o a d d is h f

1.39 1.6 1.5 1.45 3.39 3.41

1.42 1.62 1.56 1.48 3.44 3.43

Stent after unloading

^ -u n lo a d p ro x im al R u n lo a d c en tra l p ro x h f R u n lo ad c en tra l d is h f R u n lo a d distal L u n lo a d p ro x h f L u n lo a d d is h f

1.12 1.19 1.16 1.13 3.5 3.51

1.19 1.29 1.27 1.21 3.51 3.49

C alculated param eters Longitudinal proxhrrecoil Longitudinal

d is h f

(% ) -3.2 -2.9 6.8 5.3 -15.1 -3.4

(%) -2.0 -1.7 4.4 4.7 -14.1 -5.4

recoil

Foreshortening pr0xhr Foreshortening D ogbonm g D ogboning

d is h f

p r(Jx h r d is h f

T ab le 4.2

G eom etric data o f M in ik in and Mlinkthk* through loading and unloading, under the influence o f stenotic vessel o f 0.8mm thickness, achieving the same initial vessel lum en diam eter o f 0 i = 3 mm.

70

Mlinkthin
S loading unloading

Mlinkthick
loading unloading

R proximal

R central proxhf

R central dlshf

R distal

R proximal

R central proxhf

R central dlshf

R distal

F ig u re 4.11

R adial displacem ent o f Mlinkthm and Mlinkthick through loading and unloading, under the influence o f stenotic vessel, achieving the sam e initial vessel lum en diam eter o f 0 / = 3 mm.

R adial recoil throughout the Mlinkthick structure w as found to be less than the Mlinkthin, see Figure 4.12. The percentage o f radial recoil w as observed to follow the sam e pattern as the radial displacem ent w ith the least recoil found in the proxim al end, followed by the distal end and the central region o f the distal half, and finally, the greatest recoil w as observed in the central region o f the proxim al half.
II mlinkthin mlinkthick

R proximal

R central proxhf

R central dishf

R distal

F ig u re 4.12

C om parison o f percentage radial recoil o f M linkthm and Mlinkthick, subjected to free expansion, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm.

71

The resulting von M ises stress contours show sim ilar stress distribution patterns as observed during free stent expansion w hereby the highest stresses are along the arcs o f the stent structure, see Figure 4.13. The m axim um von M ises stress m agnitude for M linkthin was found to be 603.1 M Pa and 675.1, and for M linkthick was found to be 575.1 M Pa and 658.7, for the proxim al and distal halves respectively.

M Pa

580 522 464

406 348 290 Mlinkthin 232 174 116 58 0 Proxim al

Mlinkthick

Proxim al D istal

F ig u re 4.13

The resulting von M ises

Stress contours throughout the

structure o f Mlinkthin and Mlinkthick, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm.

72

Subsequently the resulting Total Equivalent Plastic strain throughout Mlinkthin and M linkthiC k structures show ed the same pattern o f distribution, see Figure 4.14, w hereby high plastic strains are observed on the arcs o f the stent structure. The m axim um Total Equivalent Plastic strains w ere found to be 0.094 and 0.142 for the proxim al and distal h a lf o f M linkthin, respectively, and 0.086 and 0.144 for the proxim al and distal h a lf o f Mlinkthick, respectively.

0.08 0.07 0.06

0.05 Proxim al 0.04

M linkthin

D istal

0.03

Proxim al Mlinkthick

D istal

Figure 4.14

The

resulting

Total

Equivalent

Plastic

Strain

contours

throughout the structure o f Mlinkthin and Mlinkthick, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm.

73

4.3.1 Stress Analysis of the Stented Vessels

It is o f interest to note that both stresses induced in the artery on loading and unloading m ay contribute to restenosis. The stresses on loading w ill act as a

stim ulus for acute damage w hilst the stresses on unloading w ill rem ain in the artery long-term and m ay be related to long-term damage.

I.

Stresses after stent loading; Stim ulus for acute dam age

The Mlinkthin and Mlinkthick w ere expanded to achieve the sam e initial lumen diam eter o f 0 / = 3 mm. The stresses generated w ith the vessel w all on loading

w ere exam ined to determ ine the degree o f acute dam age as it reached the initial lum en diam eter o f 0 / = 3 mm. The stresses in the vessels w ere exam ined in the circum ferential direction, radial direction and the longitudinal direction. It w as observed that

tensile stresses w ere induced w ithin the stented vessel in the circum ferential direction, see Figure 4.15, and com pressive stresses in the radial direction, see Figure 4.16, w hile in the longitudinal direction, both tensile and com pressive stresses w ere induced w ithin the stented vessel, see Figure 4.17. It is clear that the vessel stented w ith Mlinkthick has considerably m ore tissue stressed at high levels than the vessel stented w ith MlinkthickF or initial loading, the volum e o f tissue stressed to various levels was com puted for specific stress values in the circum ferential, radial and longitudinal directions w ithin the stented stenotic vessels, see Figure 4.18 and Figure 4.19. The highest m agnitude o f stress induced w ithin the stented stenotic vessels was found to be the tensile circum ferential stress, w here som e tissue w as stressed above 6 M Pa. The low est stresses observed w ere the longitudinal com pressive

stresses, w here less than 5% volum e o f tissue w as stressed at above 40 kPa. The volum es o f tissue stressed by the com pressive stresses in the radial and longitudinal direction w ere found to be sim ilar in the vessels stented by the

Mlinkthin and Mlinkthick- D ifferent volum es o f tissue w ere subjected to elevated

tensile stresses in the circum ferential and longitudinal directions in the vessels stented w ith the two different stents, see Figure 4.18 and Figure 4.19

74

MPa

Proxim al H alf
M linkthin

D istal H alf

Proxim al H alf Mlinkthick

D istal H alf

Figure 4.15

Tensile circum ferential stresses induced in the arterial w all stented by M linkthin and M linkthiC k, at loading, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm.

75

MPa

O -1

-2
-3 -4 Proximal H alf Mlinkthin D M H alf

Proximal H alf Mlinkthick

Distal H alf

F ig u re 4.16

Com pressive radial stresses induced in the arterial wall stented by Mlink,hin and Mlinkthick, at loading, achieving the same initial vessel lumen diam eter o f 0 / = 3 mm.

76

M Pa
4

3
2
1 0 -1

Proxim al H alf
M linkthin

D istal H a lf

-2
-3 -4

Proxim al H alf M link,thick

D istal H a lf

F ig u re 4.17

Tensile and com pressive longitudinal stresses induced in the arterial w all stented by Mlinkthin and M linkthick, at loading, achieving the sam e initial vessel lum en diam eter o f 0 / = 3 mm.

77

B th in_proxhf B thin dishf B th ic k p ro x h f thick dishf

20%

0 ) E 16% _3 O >
in

2 12 %

M o

a > 8% O ) m a >
g 4%

0-

A a > 6 M Pa Figure 4.18

B a < -1 M Pa

The percentage stress volum es b y the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel by M linkthin and M linkthick at loading, achieving the same initial vessel lum en diam eter o f 0 / = 3 mm.
3thin_proxhf Bthin dishf E thick proxhf Bthick dishf

16%

0) E

1 12%
> 0 ) (0 8% > * " o d > s> ra ** c < u 4% k_ a >
a.
V) (A

C a > 1 M Pa Figure 4.19

D a < -40 kPa

The percentage stress volum es b y the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel b y Mlinkthin and M linkthiC k, at loading, achieving the same initial vessel lum en diam eter o f 0 / = 3 m m .

78

II.

Stresses after stent unloading; Stim ulus for long-term dam age

The pressure w as rem oved from the stent once the 3 m m diam eter lum en was achieved. The resulting final arterial lum en diam eter, 0 F, w as found to be The proxim al h a lf

different for vessels stented w ith Mlinkthin and Mlinkthick-

m odel stented w ith Mlinkthin produced a final lum en diam eter o f 0/. = 2.38 m m and the distal h alf produced 0 F = 2.32 mm. The proxim al h a lf m odel stented w ith Mlinkthick produced a final lum en diam eter o f 0 F = 2.58 m m and the distal h a lf produced 0 F = 2.54 mm. These variations in the final lum en diam eter w ere as a result o f the difference in the radial recoil o f both stent structures, w here M linkthin w as found to have higher recoil than MlinkthickThe stresses in the vessels w ere exam ined in the circum ferential direction, see Figure 4.20, radial direction, see Figure 4.21, and the longitudinal direction, see Figure 4.22, and sim ilar stress patterns w ere observed in the stented vessels as those observed on loading . A decrease in stresses in all direction was found in the stented stenotic vessels. As the pressure was com pletely rem oved, leaving the stents scaffolding open the arterial w all, the volum e o f tissue stressed to various levels was com puted for specific stress values in the circum ferential, radial and longitudinal directions w ithin the stented stenotic vessels, see F igure 4.23 and Figure 4.24. The tensile circum ferential stresses w ere found to be the highest m agnitude o f stress induced w ithin the stented stenotic vessels, w here som e tissue w as stressed above 0.5 MPa. Sim ilar to loading, the longitudinal com pressive stress w as found to be the least m agnitude o f stress induced in the stented vessels w ith less than 5% stress volum es above 40 kPa. Significant differences in volum es o f tissue stressed at high levels w ere observed in the vessels stented w ith the Mlinkthin and Mlinkthick, see Figure 4.23 and Figure 4.24.

79

Proxim al H alf Mlinkthick

D istal H alf

Figure 4.20

Tensile circum ferential stresses induced in the arterial w all stented by M linkthin and M linkthiC k, at unloading, resulting in different final lum en diameter.

80

P r o x im a l H a lf

D ista l H a lf

Mlinkihin

P r o x im a l H a lf

D ista l H a lf

Mlinkihick

Figure 4.21

C o m p r e s s iv e radial str e sse s in d u ced in th e arterial w a ll sten ted b y Mlink,i,j and M link,hjC k, at u n lo a d in g , r e su ltin g in d iffer en t fin al lu m e n d iam eter.

Proxim al H alf Mlinkthick

D istal H alf

Figure 4.22

Tensile and com pressive longitudinal stresses induced in the arterial w all stented by M linkthm and Mlinkthick, at unloading, resulting in different final lum en diameter.

82

H thin proxhf thin dishf B th ic k proxhf B thick dishf

25%

E 2 20 % o

0 )

> < /> U )

V)

4 - 15%

g, 10%

C L

(0 a > o L_ a > 5%

A a > 500 kPa Figure 4.23

B a < -100 kP a

The percentage stress volum es by the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel b y Mlinkthm and M linkthiC k, at unloading, resulting in different final lum en diameter.

B thin_proxhf B thin dishf Bthick_proxhf B th ick dishf

o 16%
E

O
12%

e
o

Q.

8% 0) S) ra *c V 4% a o >

c
a > 100 kPa Figure 4.24
c t<

D -40 kPa

The percentage stress volum es b y the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel b y Mlinkthm and M linkthiC k, at loading, resulting in different final lum en diameter.

83

4.4

Case Study 3: Expansion of Mlinkth jn and Mlinkthick inside a stenotic coronary artery achieving the same final expanded lumen diameter

Four sim ulations w ere carried out, expanding the proxim al and distal halves o f M linkthin and M linkthiC k inside a stenotic coronary artery, represented as an idealised cylindrical vessel. For this case study, Mlinkthin and Mlinkthick were

expanded and allow ed to recoil to achieve the sam e final expanded lum en diam eter o f 0 /- = 2.28 m m , see Figure 4.25. The pressures required to expand The same

Mlinkthin and Mlinkthick w ere 8.2 M Pa and 2.2 M P a respectively.

pressure was applied for both the proxim al h a lf and distal h a lf o f each structure. The radial expansion throughout bo th m odels w as found to be highly non-uniform . The low est values o f radial displacem ent w ere observed at the

proxim al and distal ends o f the stent structures, see F igure 4.26. Radial displacem ent w ere m easured at k ey locations on bo th the proxim al and distal h a lf o f each stent, see Figure 4.10, in order to com pute the structural characteristics o f the stents through the loading and unloading process o f the stents expansion, see Table 4.3. A sim ilar pattern o f radial displacem ent distribution to case study 2 was observed, and the pattern was sim ilar for both the M linkthin and M linkthiC k- The least radial displacem ent was found in the proxim al end, follow ed by the distal end and then the central region o f the distal half. The highest radial displacem ent w as found in the central region o f the proxim al h a lf o f the stent structure, see Figure 4.27. F or this case study, M linkthin was found to have greater radial = 3 m m and 0 / hick = 2.48 mm, to achieve = 2.28 m m .

displacem ent at loading, w here 0 / sim ilar final vessel lum en diam eter,

U pon loading, foreshortening w as observed to be low in m agnitude, how ever, in contrast to case study 2, M linkthm foreshortens to a larger degree than the Mlinkthick, see Table 4.3. Again, the negative dogboning was found more

84

prom inent in the proxim al region o f the stent structures, how ever, the M linkthiC k showed this to a lesser extent. Upon unloading, elongation occurred in both stents structures, w hich gave rise to the negative values for longitudinal recoil. A greater degree o f

longitudinal recoil w as observed in the M linkthi th an the M linkthiC k and the m agnitude o f the longitudinal recoil in Mlinkthick was found to be very low in this case, see Table 4.3.

Proxim al h a lf M lin k thick

D istal h alf

F ig u re 4.25

D eform ed geom etry o f proxim al and distal h a lf o f M linkthm and Mlinkthick scaffolding a stenotic vessel, thickness o f 0.8 mm, achieving the same final vessel lum en diam eter o f 0 f = 2.28 mm.

85

mm

Proxim al Mlinkthin mm

D istal

Proxim al

, Mlinkthick

D istal

Figure 4.26

R adial displacem ent distribution throughout the stents structure o f Mlinkthin and Mlinkthick, under the influence o f a stenotic vessel, achieving the same final vessel lum en diameter.

[ m lin k th in m lin kth ick

m lin kth in m lin k th ic k

I 1-6
1 1.2

I g 0.8 a . (A
D 0.4 "5 C*

T S 0

i i:m
R proximal R central proxhf R central dishf

ii
R distal

R proximal

R central proxhf

R central d ishf

R distal

a R adial displacem ent o f Mlinkthm and Mlinkthick through (a) loading and (b) unloading, under the influence o f a stenotic vessel, achieving the sam e final vessel lum en diameter.

Figure 4.27

86

M lin k ti,m

M lin k thiC k

Stent before expansion

R o iig Long

(mm) 0.77 3.6

(mm) 0.86 3.6

Stent after loading

R lo a d p ro x im al R lo a d c cm ra l p ro x h f R l o a d c cn tra l d is h f R lo a d distal L |o a d p ro x h f L lo a d d is h f

1.39 1.6 1.5 1.45 3.39 3.41

1.18 1.27 1.24 1.23 3.54 3.53

Stent after unloading

R u n lo a d p ro x im al R u n lo a d c cn tra l p ro x h f R u n lo a d c cn tra l d is h f R u n lo a d d istal L u n lo a d p ro x h f L u n lo a d d is h f

1.12 1.19 1.16 1.13 3.5 3.51

1.1 1.19 1.16 1.15 3.55 3.55

Calculated parameters

(%) -3.2 -2.9 5.8 5.3 -15.1 -3.4

(%) -0.3 -0.6 1.7 1.9 -7.6 -0.8

Longitudinal Longitudinal

p io x h f d is h f

recoil

recoil

Foreshortening Foreshortening

p r0 x h r d is h f

D ogboning pr0xhr D ogboning

d is h r

Table 4.3

G eom etric data o f Mlinkmm and M IinkthjC k through loading and unloading, under the influence o f stenotic vessel o f 0.8mm thickness, achieving the sam e final vessel lumen diameter.

87

R adial recoil throughout the M linkthiC k structure w as found to b e less than the Mlinkthm, see Figure 4.28. In fact, the radial recoil o f M linkthm w as found to be three tim es greater than the Mlinkthick- M linkthin shows the highest radial recoil at the region w here it experienced the highest radial displacem ent. The resulting V on M ises stress contours show sim ilar stress distribution patterns as observed in previous case studies, w hereby the highest stresses are along the arcs o f the stent structure, see Figure 4.29. The m axim um V on M ises stress m agnitude for Mlinkthm was found to be 603.1 M P a and 675.1, and for Mlinkthick w as found to be 493.2 M Pa and 559.5 for proxim al and distal halves respectively. The resulting Total Equivalent Plastic strain throughout Mlinkthm and Mlinkthick, structures show ed the sam e pattern o f distribution, see Figure 4.30, w hereby high plastic strains are observed on the arcs o f the stent structure. The m axim um Total Equivalent Plastic strain w as found to be 0.094 and 0.142 for the proxim al and distal h a lf o f M linkthm respectively, and 0.037 and 0.072 for the proxim al and distal h a lf o f Mlinkthick respectively.

mlinkthin mlinkthick

R proximal

R central proxhf

R central dishf

R distal

Figure 4.28

C om parison o f percentage radial recoil o f Mlinkthm and Mlinkthick, subjected to free expansion, achieving the sam e final vessel lum en diameter.

88

MPa
_ 580 522 464

406 348 290 Mlinkthin 232 174 116 58 0 Proxim al Distal

Figure 4.29

The resulting von M ises

Stress contours throughout the

structure o f M linkthin and M linkthiC k, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the same final vessel lum en diameter.

89

0.08
0.07 0.06

0.05 Proxim al 0.04 M lin k thin 0.03 D istal

Proxim al Mlinkthick

F ig u re 4.30

The

resulting

Total

Equivalent

Plastic

Strain

contours

throughout the structure o f M linkthm and Mlinkthick, under the influence o f stenotic vessel o f 0.8 m m thickness, achieving the same final vessel lum en diameter.

90

4.4.1 Stress Analysis of the Stented Vessels

I. Stresses after stent loading: Stim ulus for acute dam age

The Mlinkthin and M linkthiC k w ere expanded and recoil to achieve a final lum en diam eter o f &f = 2.28 mm. The initial lum en diam eter achieved b y Mlinkthin and Mlinkthick w ere 0 / thin = 3 m m and 0 / thick = 2.48 mm. The stresses generated w ith the vessel w all on loading w ere exam ined to determ ine the degree o f acute dam age as it reached the initial lum en diameter. C om paring the stresses induced in both vessels, it is evident that the vessel stented w ith M linkthin has significant m ore tissue stressed at high levels than the vessel stented w ith Mlinkthick, see Figure 4.31-4.33. The volum e o f tissue stressed to various levels w as com puted for specific stress values in the circum ferential, radial and longitudinal directions w ithin the stented stenotic vessels, see Figure 4.34 and Figure 4.35. Sim ilar to

case study 2, the tensile circum ferential stress show ed significantly higher m agnitudes at 800 kPa, induced in the stented vessels com pared w ith the other stresses. In contrast to the previous case study, Mlinkthin w as found to cause the greater volum es o f tissue stressed at high levels w hen com pared w ith Mlinkthick in all directions. Sim ilar to the previous case study, the low est m agnitude o f stress was found to be the com pressive longitudinal stress, w here less than 5.1% volum e o f tissue w as stressed above 40 kPa.

91

MPa
5.0 4.5 4.0 3.5 3.0 Proxim al H alf 2.5 Mlinkthin D istal H alf

2.0
1.5

1.0
0.5 0

Proxim al H a lf Mlinkthick

D istal H alf

F ig u re 4.31

Tensile circum ferential stresses induced in the arterial wall stented by M linkthm and M linkthiC k, at loading, achieving different initial vessel lum en diameter.

92

MPa
0
-0.5
-

1.0

1.5 2.0 2.5

P r o x im a l H a lf

D ista l H a lf

Mlinkthin
-

- 3 .0

3.5

-4 .0

P r o x im a l H a lf M linkthick

D ista l H a lf

F ig u re 4.32

C o m p r e s s iv e radial str e sse s in d u ce d in th e arterial w a ll sten ted b y Mlink,hin and M link,hjC k, at lo a d in g , a c h ie v in g d iffe r e n t in itial v e s s e l lu m e n d iam eter.

93

MPa

2.0
1.5

1.0
Proxim al H alf Mlinkthin

D istal H alf

Proxim al H alf Mlinkthick

D istal H alf

F ig u re 4.33

Tensile and com pressive longitudinal stresses induced in the arterial w all stented by M linkthm and Mlinkthick, at loading, achieving different initial vessel lum en diameter.

94

B Sthin jjro x h f B thin_dishf B thick_proxhf B th ic k d is h f

v 80 % E 3 O > (ft 60% (ft 0 ) (n o 40% (D O ) n 0 ) D .

c s 20%
4-1
L_

0% A a > 800 kPa F ig u re 4.34 B a < -200 kPa

The percentage stress volum es b y the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel b y M linkthjn and M linkthiC k, at loading, achieving different initial vessel lum en diameter.
Bthin_jjroxhf H th in dishf B th ic k p r oxhf B thick dishf

100.0%

ai

80.0%

O > (ft (/)

N

a >

60. 0%

liii * "v

o C T 40.,0%

0 ) 0) 20 .0 % 0_
o

4 m -k V ^

0 . 0%

C a > 200 kPa Figure 4.35

D a < -40 kPa

The percentage stress volum es b y the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel b y M linkthm and M linkthiC k, at loading, achieving different initial vessel lumen diameter.

95

II.

Stresses after stent unloading; Stim ulus for long-term dam age

The pressure w as rem oved from the stents and th e stents w ere allow ed to recoil to a final lum en diam eter o f 0/.- = 2.28 mm. The stresses generated w ithin the vessel w all on unloading w ere exam ined to determ ine the degree o f long-term damage. The stresses in the vessels w ere exam ined in the circum ferential direction, see Figure 4.36, radial direction, see Figure 4.37, and the longitudinal direction, see Figure 4.38. A decrease in stresses in all direction w as found in the stented stenotic vessels at unloading. The pattern o f stresses induced in the

stented vessels rem ained the same as loading, how ever, the difference in stresses in the tw o stented vessels on loading, w as observed to reduce, as bo th stented vessels b y M linkthin and M linkthick achieved the sam e final lum en diameter. The volum e o f tissue stressed to various levels was com puted for specific stress values in the circum ferential, radial and longitudinal directions w ithin the stented stenotic vessels, see Figure 4.39 and Figure 4.40. The tensile circum ferential stress rem ained the highest m agnitude o f stress induced w ithin the stented stenotic vessels, w here some tissue w as stressed above 0.5 M Pa. Sim ilar to loading, the longitudinal com pressive stress w as found to be the least m agnitude o f stress induced in the stented vessels w ith less than 4.1% stress volum e at above 40 kPa.

96

MPa
0.50 0.45

0.40 0.35 0.30

Proxim al H alf Mlinkthin

D istal H alf

0.25

0.20
0.15

0.10
0.05

0 Proxim al H alf Mlinkthick D istal H alf

Figure 4.36

Tensile circum ferential stresses induced in the arterial wall stented by Mlinkthin and Mlinkthick, at unloading, achieving the same final vessel lum en diameter.

97

Proximal Half M link,hin

Distal Haif

Proximal Half

_. .. , Mlinkthick

Distal Half

Figure 4.37

Compressive radial stresses induced in the anerial wall stented by Mlinkthi and Mlinklhick, at unloading, achieving the same final vessel lumen diameter.

98

M Pa

1
0.20
0.15

0.10
0.05 0 -0.05

I
Proxim al H alf
M lin k thin

D istal H a lf

Proxim al H alf Mlinkthick

D istal H alf

F ig u re 4.38

Tensile and com pressive longitudinal stresses induced in the arterial wall stented by Mlinkthm and Mlinkthick, at unloading, achieving the same final vessel lum en diameter.

99

H thin_proxhf thin_dishf thick_proxhf thick_dishf

tu 12%

E
3

tfl
|2
V)

8%

*-

0 0) O )
4%

0) o k. 0) Q .

0% A a> 300kPa F ig u re 4.39 B a < -1 0 0 k P a

The percentage stress volum es b y the tensile circum ferential stress (A) and radial com pressive stress (B), w ithin the stented vessel b y M linkth,n and M linkthiC k, at unloading, achieving the same final vessel lum en diameter.
H th in p ro x h f H thin dishf B Bth ic k p ro x h f thick dishf

18% (U
E 15%

tn 0 L_ 2
o
03

W 12%

9%
6%
3%

S
p

CL

0% C o > 40 kPa D a < -40 kP a

F ig u re 4.40

The percentage stress volum es by the tensile longitudinal stress (C) and longitudinal com pressive stress (D), w ithin the stented vessel by M linkthjn and M linkthiC k, at unloading, achieving the same final vessel lum en diameter. 100

4.5 Case Study 4: Expansion of Mlinkth ilI and Mlinkthick inside a localised stenotic coronary artery achieving the same initial expanded lumen diameter
Tw o sim ulations were carried out, expanding the distal halves o f Mlinkthin and Mlinkthick inside a localised stenotic coronary artery vessel, represented as a localised crescent-shaped axisym m etric stenosis. The vessel incorporated the The pressure

properties o f plaque and healthy porcine coronary arterial tissue.

used to expand Mlinkthin was 41 M Pa and 45.7 M Pa for Mlinkthick- B oth Mlinkthin and Mlinkthick expansions achieved an initial lum en diam eter, 0 /, o f 3.18 m m , see Figure 4.41.

A rtery

Plaque

Stent
Mlinkthick

F ig u re 4.41

D eform ed geom etry o f distal halves o f Mlinkthin and Mlinkthick, scaffolding a stenotic vessel, thickness o f 0.8 m m , achieving the same initial vessel lum en diam eter o f 0 / = 3.18 mm.

1 01

The radial expansion throughout the structures o f both M linkthin and Mlinkthick was found to be highly non-uniform , see Figure 4.42. D ue to the

geom etry o f the local stenotic vessel (i.e. plaque geom etry), the distal end o f the stent structures w ere observed to experience the highest radial displacem ent.
mm

0.92 0.87 0.82 0.76 0.71

0.66
0.60

mm

Mlinkthin

0.86
0.81 0.75 0.70 0.65 0.59 0.54

Mlinkthick

F ig u re 4.42

Radial displacem ent distribution throughout the stents structure o f Mlinkthin and Mlinkthick, under the influence o f a localised stenotic vessel, thickness o f 0.8 mm, achieving the same initial vessel lum en diameter.

102

R adial displacem ent w ere m easured at k ey locations on both the proxim al and distal h a lf o f each stent, see Figure 4.10, in order to com pute the structural characteristics o f the stents through the loading and unloading process o f the stents expansion, see Table 4.4.

Mlinkthm

Mlinkthick

Stent before expansion R-orig Long

(mm) 0.77 3.6

(mm ) 0.86 3.6

Stent after loading R-load central Rload distal Lload 1.45 1.68 3.28 1.46 1.68 3.25

Stent after unloading Runload central R-unload distal Ljnload 1.12 1.27 3.51 1.18 1.3 3.43

Calculated param eters Longitudinal recoil Foreshortening D ogboning

(%) -7.0 8.7 13.7

(%) -5.5 9.7 13.1

T a b le 4.4

Geom etric data o f M linkthin and Mlinkthick through loading and unloading, under the influence o f localised stenotic vessel o f 0.8m m thickness, achieving the sam e initial vessel lum en diam eter o f 0 / o f 3.18 m m and a final lum en diam eter, 0 F o f 2.44 m m for M linkthm and 0 F o f 2.52 m m for Mlinkthick-

103

B oth stents, M linkthin and M linkthiC k have the sam e pattern o f radial displacem ent distribution. H ow ever, in contrast to case study 2 and 3, the central region w as observed to experience low er radial displacem ent than the distal end, see Figure 4.43. For this analysis, bo th M linkthin and M lin k thiC k w ere expanded to achieve an initial lum en diam eter, 0 / o f 3.18 mm w hile at the end o f loading, the final lum en diam eter for Mlinkthin was found to be 2.44 m m and 0 f o f 2.52 mm for Mlinkthick. Foreshortening w as observed at loading w hich w as relatively high in m agnitude in com parison to the other case studies. U nlike in the free expansion o f stents, and the expansion o f the stents in idealised stenotic vessels, dogboning w as observed in M linkthm and Mlinkthick during expansion, w hereby M linkthiC k show ed a slightly low er degree o f dogboning. U pon unloading, the radial recoil o f Mlinkthin w as found to be greater than the recoil in the Mlinkthick and the highest radial recoil w as observed at the region w here the highest radial displacem ent was observed, see Figure 4.44. Longitudinal recoil w as found to be negative, indicating elongation occured. The resulting von M ises stresses contours w ere exam ined and found to be sim ilar for both stents, see Figure 4.45, w hereby the arcs w ere subjected the high level o f von M ises stress. The m axim um V on M ises stress m agnitude for Mlinkthin w as found to be 690.7 M Pa and for M linkthiC k w as found to be 493.7 M Pa. The resulting contours o f Total Equivalent Plastic strain throughout the expanded structures o f Mlinkthin and Mlinkthick, w ere taken at the end o f loading. H igh plastic strains are evident on the arcs o f the stent structures and the m axim um Total Equivalent Plastic strain w as found to be 0.156 for M linkthjn and 0.171 for Mlinkthick, see Figure 4.46.

104

 mlinkthin mlinkthicK

mlinkthin mlinkthick

0.9

R c e n tra l

R d ista l

R c e n tra l

R d is ta l

F ig u re 4.43

C om parison o f radial displacem ent o f Mlinkthin and Mlinkthick inside a localised stenotic vessel throughout the structure at (a) loading and (b) unloading, achieving an initial lum en diameter, 0 / o f 3.18 m m and a final lum en diam eter, 0 f o f 2.44 m m for

Mlinkthin and 0 // o f 2.52 m m for Mlinkthick-

mlinkthin mlinkthick

25
Percentage Radial Recoil (%)

20
15

10
5

0
R central R distal

F ig u re 4.44

Com parison o f percentage recoil o f Mlinkthin and Mlinkthick inside a localised stenotic vessel, achieving an initial lum en diameter, 0 / o f 3.18 mm.

105

M Pa

550 495 440

385 330 275 Mlinkthin

220
165

110
55 0

Mlinkthick

F ig u re 4.45

The resulting von M ises stress contours throughout the structure o f Mlinkthin and M linkthiC k, under the influence o f a localised stenotic vessel, achieving an initial lum en diam eter, 0 / o f 3.18 m m and a final lum en diam eter, 0 p o f 2.44 m m for Mlinkthin and 0 f o f 2.52 m m for Mlinkthick-

106

0.08
0.07 0.06 0.05
M linkthin

0.04 0.03

0.02 0.01
0

Mlinkthick

F ig u re 4.46

The

resulting

Total

Equivalent

Plastic

strain

contours

throughout the structure o f M linkthin and Mlinkthick, under the influence o f a localised stenotic vessel, achieving an initial lum en diameter, 0 ; o f 3.18 m m and a final lum en diam eter, 0/r o f 2.44 mm for Mlinkthin and 0 ^ o f 2.52 m m for Mlinkthick-

107

4.5.1 Stress Analysis of the Stented Vessels

I. Stresses after stent loading; Stim ulus for acute dam age

The M linkthin and Mlinkthick w ere expanded to achieve the sam e initial lum en diam eter o f 0 / = 3.18 mm. The stresses generated w ithin the vessel w all on

loading w ere exam ined to determ ine the degree o f acute dam age as it reached the initial lum en diameter. The stresses in the vessels w ere exam ined in the circum ferential, radial and longitudinal directions. C om paring the stresses induced in bo th vessel, it was found that there w as a considerable m ore tissue stressed at high levels found in stented vessels by Mlinkthin and M linkthiC k, see Figure 4.47-4.49. The sam e pattern w as found in case study 2, w here in bo th case studies, the stents w ere expanded to achieve the same initial expanded lum en diameter. The volum e o f tissue stressed to various levels w as com puted for specific stress values in the circum ferential, radial and longitudinal directions w ithin the stented stenotic vessels, see Figure 4.50. In contrast to other case

studies, the m agnitude o f stresses induced w ithin the stented vessels w as found to be significantly higher. The highest m agnitude o f stress induced w ithin the

stented stenotic vessels w as produced b y the tensile stresses in the circum ferential and longitudinal direction, w hereby som e tissue was stressed above 60 M P a and 20 M Pa respectively. The least m agnitude o f stress w as found to be the

com pressive longitudinal stress, w here less than 5% volum e o f tissue at a high m agnitude o f 0.5 M Pa. M eanw hile, there w as some tissue stressed experiencing com pressive radial stress at above 10 M Pa.

108

Mlinkthick

F ig u re 4.47

Tensile circum ferential stresses induced in the arterial wall stented by M linkti,in and M linkti,jck, at loading, achieving the same initial lum en diam eter, 0 / o f 3.18 mm.

109

MPa
0 -5

-10
-15

-20
-25 -30

M link.hi,,

-35 -40 -45 -50

Mlink'hirk

F ig u r e 4 .4 8

C o m p r e s s iv e radial str e sse s in d u ce d in the arterial w a ll ste n ted b y Mlinkti,,,, and M lin k t|C k, at lo a d in g , a c h ie v in g the sa m e in itial lu m e n d ia m eter, 0 / o f 3 . 1 8 m m .

110

MPa 50
40 30 20 10 0 M lin k tbin

-10 -20
-30 -40 -50

M linkthick

F ig u re 4.49

Tensile and com pressive longitudinal stresses induced in the arterial w all stented by Mlinkthm and M linkthiC k, at loading, achieving the same initial lum en diam eter, 0 / o f 3.18 mm.

Ill

10.0%

Percentage o f Stress Volume

8.0% 6.0% 4.0% 2.0%

0.0%

a > 6 0 M Pa

a < -1 0 M P a

o>20M Pa

a < -0 .5 M Pa

F ig u re 4.50

The percentage volum e stress b y the tensile circum ferential stress (A), radial com pressive stress (B), tensile longitudinal stress (C) and com pressive longitudinal stress (D), at loading, w ithin the stented vessel b y M linkthm and Mlinkthick, achieving an initial lum en diameter, 0 / = 3.18 m m at loading.

112

II.

Stresses after stent unloading; Stim ulus for long-term dam age

The pressure w as rem oved from the stent once the 3.18 m m diam eter lum en was achieved. The resulting fmal arterial lum en diam eter, 0 f , w as found to be

different for vessels stented w ith M linkthin and Mlinkthick- The localised stenotic vessel stented w ith Mlinkthin produced a fm al lum en diam eter o f 0/. = 2.44 m m w hile Mlinkthick produced a final lum en diam eter o f & f = 2.52 mm. These

variations in the fm al lum en diam eter w ere the as a resu lt o f the difference in the radial recoil o f both stent structures, w here Mlinkthin w as found to have higher recoil than MlinkthickThe stresses in the vessels w ere exam ined in the circum ferential direction, see Figure 4.51, radial direction, see F igure 4.52, and the longitudinal direction, see Figure 4.53, and sim ilar stress patterns w ere observed in the stented vessels as those observed on loading. A decrease in stresses in all direction was found in the stented stenotic vessels. In contrast to loading, the difference

betw een the stented vessels by M linkthin and Mlinkthick w as found m ore prom inent. As the stents structures w ere left scaffolding open the arterial w all, the volum e o f tissue stressed to various levels w as com puted for specific stress values in the circum ferential, radial and longitudinal directions w ithin the stented stenotic vessels, see Figure 4.54. Sim ilar to case study 2, w hereby the stents w ere expanded to the sam e initial expanded lum en diam eter, w hich resulted in different fm al lum en diam eter, the stented vessel b y Mlinkthick show ed a significant higher volum e o f tissue stressed at high levels at unloading. Sim ilar to loading, the tensile stresses w ere found higher than the com pressive stress found in the stented vessels. Som e tissue w as stressed above 1 M Pa b y the tensile circum ferential stress and 0.5 M P a b y the tensile longitudinal stress found in the stented vessels b y Mlinkthin and MlinkthickM eanw hile the

stress volum e w as com puted above 0.1 M Pa for the com pressive stress in the radial and longitudinal direction. There w as m ore tissue stressed at this level by the com pressive radial stress, see Figure 4.54.

113

MPa
4 .0 3 .5 3 .0

2 .5

2.0
1.5

1.0
L
0 .5

M lin k th ic k

Figure 4.51

T e n s ile c ir c u m fere n tia l stre ss in d u c e d in th e lo c a lis e d ste n o tic v e s s e l w a ll sten te d b y M lin k thm and M lin k thiC k, at u n lo a d in g , a c h ie v in g a fin a l lu m e n d ia m eter, and

0Fo f

2 .4 4 m m fo r Mlinkthin

0F o f 2 .5 2

m m fo r Mlinkthick-

114

M Pa

0
-0 .3
-

0.6

Mlink<thin

-0 .9

1.2

-1.5

M linkthick

Figure 4.52

C o m p r e s s iv e radial

stre ss in d u ced

in the

lo c a lis e d

ste n o tic

v e s s e l w a ll ste n te d b y M link,hjn and M linkn,iC k, at u n lo a d in g , a c h ie v in g a fin a l lu m e n d ia m eter, 0 /r o f 2 .4 4 m m for M linkn,,,, and 0 / .- o f 2 .5 2 m m fo r Mlinkihick-

115

M lin k th ic k

Figure 4.53

Tensile and com pressive longitudinal stress induced in the localised stenotic vessel w all stented by M linkthin and Mlinkthick, at unloading, achieving a final lum en diam eter, 0 F o f 2.44 mm for Mlinkthin and 0 F o f 2.52 mm for Mlinkthick-

116

 m lin k th in m lin k th ic k

15.0% 0) I 12.0% > ID 5 * -< (0 ig D ) ffl p c

Q.

9.0%

< u

6.0%

a >

3.0%

0.0% A o > 1 M Pa B a < -0 .1 M P a C c > 0.5 M P a D a < -0 .1 M Pa

Figure 4.54

The percentage volum e stress b y the tensile circum ferential stress (A), radial com pressive stress (B), tensile longitudinal stress (C) and com pressive longitudinal stress (D), at unloading, w ithin the stented vessel b y Mlinkthm and M linkthiC k, achieving a final lum en diam eter, 0 f o f 2.44 m m for Mlinkthin and 0 /- o f 2.52 mm for M linkthick-

4.6

Summary

This chapter has evaluated the results follow ing the four case studies to investigate the influence o f stents expansion on stents characteristics. The

stresses induced w ithin the stented vessels w ere thoroughly exam ined, identifying the difference in the m agnitude at loading and unloading, w hich provided a greater insight into the m echanical stim uli for restenosis.

117

Chapter 5

DISCUSSION
5.1 Introduction

The IS A R STETREO Trial by K astrati et al. [26] exam ined the clinical restenosis outcom e o f the ACS R X M ultilink and the ACS M ultilink R X Duet. These

Multi Link stents are sim ilar in design but have w ith different strut thicknesses. The expansion characteristics o f the thinner strut stent (M linkthm) and the thicker strut stent (M linkthiC k) w ere exam ined during this study using FE analyses. The stents w ere com pared under free expansion and also expanded in different vessels geom etries. The vessel geom etry and the m aterial properties o f the arterial tissue and plaque resulted in significantly different expanded stent structures. The

stresses induced w ithin the stented vessel walls b y the tw o stents were also thoroughly exam ined since these stresses are believed to potentially act as the m echanical stim uli for in-stent restenosis. This chapter therefore discusses the FE analyses carried out as follows

1. C om parison o f pressure deploym ent o f M linkt]lin and M linkthjC k for each case study. 2. C om parison o f the expanded M linkthm and M linkthiC k stents 3. C om parison o f the radial recoil o f Mlinkthm and M linkthick after deployment.

118

4. C om parison o f the stresses induced w ithin the stented vessels through loading and unloading o f pressure, determ ining the volum e o f tissue stressed at high levels in each stented vessel and hence the level o f dam age caused b y M linkthm and M linkthick-

Finally, issues o f stent design are discussed in term s o f the results obtained from this study.

5.2 Limitations of this Study

The accuracy and the reliability o f the results from a finite elem ent study are highly dependent upon the inputs, such as the geom etry, the boundary conditions applied to represent the problem adequately, and the m aterials properties assigned to each component. The deploym ent o f a stent involves the interaction betw een a balloon, stent and atherosclerotic coronary artery. The m odelling o f this process using finite elem ent analysis is a com plicated and challenging study. To

com pletely describe the problem involves m odelling the interaction betw een the balloon and stent, and the stent and artherosclerotic artery. This analysis did not include the influence o f the balloon in expanding the stents and this is therefore a lim itation o f this work. The inclusion o f a balloon in the expansion o f the stent m ay result in a m ore uniform stents structure. H ow ever the ballo o n s m aterial is believed to contribute largely to the uniform ity o f stents expansion. Future w ork could include determ ining the influence o f different balloon types on the expansion characteristics o f stents. Coronary stents are placed in coronary arteries w here the blood is pulsatile in nature and hence the stented vessel is subjected to a dynam ic pressure environm ent after im plantation. The blood flow inside the coronary vessel also creates a shear stress on the w all o f the arteries that m ay alter upon im plantation o f the stents. A lthough stented vessels m ay exhibit a biological response to

changes in shear stresses, fluid flow dynam ics w ere beyond the scope o f this study.

119

One o f the m ajor lim itations o f the finite elem ent m odels in this study is the geom etrical representation o f the stents and th e artery. The geom etric

m easurem ents o f the stents, the ACS RX M ultiLink and the A CS M ultiLink RX Duet, w ere extrapolated from a handbook o f coronary stents [89]. The stents

could have been m ore accurately described using stent dim ensions obtained from the m anufacturer o f these stents (Guidant), how ever, this inform ation could not be obtained from the m anufacturer. The idealised representation o f the stenotic

vessel geom etry also represents another lim itation to this study. D ifferent vessel geom etries w ere adopted in the analyses, although all o f the geom etries were idealised representations o f real vessels. This study focuses on exam ining the role o f strut thickness in generating stresses in the stented vessel, w hereby the absolute m agnitude o f these stresses w as not the prim ary concern, but rather the difference in the stresses generated b y the two different stents. Therefore, the finite element m odels used for this com parative study w ere adequate in order to identify the different levels o f stresses induced w ithin stented stenotic vessels b y the thinner and thicker struts stents. The study can also clearly indicate the influence o f

changes in vessel geom etry on the m agnitude o f the stresses that w ould be induced in a stented vessel. The stentss structures w ere discretised by m eans o f eight-noded isoparam etric, three-dim ensional brick elem ents w ith one elem ent through the thickness, see Figure 3.3. lim itation to this study. The m esh o f the stent structure represents another A higher order elem ent and finer m esh density m ay

contribute to the accuracy o f the m agnitude o f expansion and stresses found in the stents structure. C ontact analysis for M arc/M entat requires a finer m esh on the most deform able body o f the contact bodies, i.e. the stenosed artery in this case. Therefore this m esh density o f the stent was chosen to m inim ise the num ber o f elements and w as found reasonable. A finer m esh o f the stent structure w ould lead to a finer m esh density needed for the artery. The use o f a higher order

elem ent and a finer m esh density o f the stent structure w ould consequently result in m uch greater heavy com putational tim e and resources. The m aterial m odels represent another lim itation to this study. The

arterial w all w as represented by an isotropic M ooney-R ivlin m odel, instead o f an anisotropic viscoelastic m odel that includes the hysteresis exhibited b y arterial

120

tissue and the anisotropy observed in the com position o f the arterial wall. R esidual stress w as also absent in the constitutive m odel defined for the arterial w all properties. The M ooney-R ivlin m odel has been found to adequately

represent arterial tissue w hen determ ined b y fitting to published experim ental data from uniaxial and equibiaxial tension test data as it includes the stress-stiffening behaviour o f arterial tissue [41]. A lthough the plaque tissue w as described using a constitutive m odel defined by fitting to published data on the stress-strain behaviour o f hum an atherosclerotic plaque obtained from uniaxial tension test data alone, to the authors know ledge no other uniaxial tensile test data on hum an plaques has been published and no data on the biaxial tensile behaviour o f hum an atherosclerotic plaques could be found in the literature. The m aterial data present a direct link to the m agnitude o f stresses exam ined for these analyses. D ifferent m aterial data are likely to result in a different m agnitude o f stresses. However, this study has successfully identified the m ajor stress contributor found in the stented vessel. The lim itations outlined m ay be accepted for the purpose o f com parative analyses such that the m odels lim it the com plexity o f m odelling the process o f stent expansion w ithin an artery vessel. H ow ever, the m odels still provide

valuable inform ation on the influence o f stent design on stent expansion and vessel stresses. For this purpose all param eters, for exam ple, the m aterials used for these analyses and the geom etrical representations o f the stenotic vessels and the stents structures w ere kept identical, w ith the only variation being the strut thicknesses o f the stents for each case study.

5.3

Comparison of Pressure Deployment

The lum en pressure applied to the stents w as ram ped up for a small tim e step size. The size o f time step depended upon the am ount o f pressure applied for each analysis and w as therefore found to be highly dependant on the stent structure. The pressure w as unloaded by h a lf the tim e it took to load it. H igh tim e step sizes w ere found to cause too m uch pressure to be applied in a single increm ent resulting in the sim ulation not converging.

121

5.3.1 Comparison of thinner strut stent vs. thicker strut stent

The analyses found that the thicker strut stent needed a higher am ount o f pressure com pared w ith the thinner strut stent, in order to expand bo th stents to the same initial expanded outer diameter, as seen for case study 1, 2 and 4, see Figure 5.1. U nder free expansion, there w as a significant difference in the pressure required to expand both stents to achieve the same outer diam eter o f 3 mm. It w as found that the pressure required to expand Mlinkthick w as approxim ately three tim es higher than the pressure required to expand M linkthm to the sam e diam eter. U nder the influence o f a stenosed straight vessel, Mlinkthm required approxim ately 78% o f the pressure required to expand the Mlinkthick and 90% under the influence o f a localised stenosed. H owever, in case study 3, w hereby Mlinkthm and Mlinkthick

w ere expanded and allow ed to recoil to achieve the sam e final outer diameter, under the influence o f a stenosed vessel, the pressure required to expand Mlinkthick w as found to be significantly lower, w hereby Mlinkthick required approxim ately 27% o f the pressure required to expand the Mlinkthm-

5.3.2 Comparison of each stent design under free expansion and under the influence of various stenosed vessels

The pressure required to expand the sam e stent design under free expansion and under the influence o f a stenotic straight vessel w as found to be significantly different, see Figure 5.1. For example, to expand Mlinkthm under free expansion, the pressure required w as significantly low er at only 5.3% o f the pressure required to deploy M linkthm under the influence o f a stenotic straight vessel to achieve the same outer diameter. In case study 4, w hereby M linkthm and Mlinkthick w ere

expanded inside a localised stenotic vessel w ith plaque and arterial properties, the pressures required to expand both Mlinkthm and Mlinkthick w ere significantly higher than for the straight stenosed vessel. This high pressure has also been

show n to be associated w ith the m agnitude o f stresses found in the stented vessels. Clearly, the pressure required for stent deploym ent is highly dependant on the

122

stent design, the m aterial properties o f the atherosclerotic coronary arteries, and the geom etry o f the vessel wall.

mlinkthin mlinkthick

1 Figure 5.1

C om parison o f pressure deploym ent o f Mlinkthin and M linkthjC k; for 1) free expansion (Case study 1); 2) w ithin the straight stenotic vessel achieving the sam e initial lum en diam eter (Case study 2); 3) w ithin the straight stenotic vessel achieving the sam e final lum en diam eter (Case study 3); and 4) w ithin the localised stenotic vessel achieving the sam e initial lum en diam eter (Case study 4).

5.4 Comparison of the Mlinkth in and Mlinkthick stents after deployment

5.4.1 von Mises stress and plastic strain

The von M ises stresses and plastic strain distributions w ithin the deployed stents show ed sim ilar patterns for all o f the sim ulations. It is interesting to note the

com parison o f m axim um von M ises stress in each stent design under different influence o f expansion, see Figure 5.2.

123

 mlinkthin thin_dishf
700
CL

m linkth ick

thin_p roxhf

thick_proxhf thick_dishf

re

w & 600 </) < /> at

(A

(A

500 ra 400 1 Figure 5.2 Com parison o f m ax von M ises stress found in the structures o f Mlinkthin and Mlinkthick, upon deploym ent for 1) free expansion (Case study 1); 2) in a straight stenotic vessel achieving the same initial lum en diam eter (Case study 2); 3) in a srtaight stenotic vessel achieving the same final lum en diam eter (Case study 3); and 4) in a localised stenotic vessel achieving the same initial lum en diam eter (Case study 4).

M urphy et al. [88] reported that the stress-strain behaviour o f 316L stainless steel stent struts is size dependent. That study suggested that certain

areas o f thinner stent struts could experience local failure, w hich w ould not be detected if bulk stress-strain curves w ere used to describe the m aterial behaviour. They proposed that any analysis o f a stent w ith strut thickness below 0.05 mm should use a size based stress-strain relationship. U nder free expansion, the specific stress-strain behaviour o f 316L stainless steel was incorporated into the analyses, according to the stents strut size. The resulting m axim um von M ises stresses w ere found only m arginally

higher for Mlinkthin at 604.4 M Pa com pared w ith 597 M Pa for MlinkthickH owever, the failure stress for the thinner strut stent was found to 693.2 M Pa

124

com pared w ith 730.4 M Pa for thicker strut stent from the data b y M urphy et al. [88]. The Mlinkthin was therefore at a greater risk o f failure than the M linkthjC kAs discussed earlier, the pressure required to expand the stent structures under the influence o f stenosed vessels w as significantly higher than that required to expand the stents under free expansion. In order to run the sim ulations the

failure stress o f the thinner strut stent had to be brought up to sam e failure point as thicker strut stent. For these case studies, w here the stent w as expanded inside a stenosed vessel, the m ax von M ises stress for the Mlinkthin was found to exceed the failure stress found by M urphy et al. [88], see Figure 5.2. Clearly therefore, the thin strut stent is highly susceptible to failure and the use o f a m aterial w ith higher failure strength m ay be necessary to prevent this risk o f failure in thin strut stents; exam ples include cobalt chrom ium which has been successfully used for thin strut stents in recent years [91].

5.4.2 Foreshortening

Foreshortening is a common problem in m etallic stents. stent in the axial direction during deploym ent.

It is the contraction o f a

In general foreshortening is not

desirable, as accurate placing and optim um longitudinal vessel coverage is com prom ised w ith increasing foreshortening in stents. Foreshortening w as found to vary considerably for all o f the case studies evaluated in this thesis, see Table

U nder free expansion, foreshortening w as absent in both stent designs. In fact, the stents w ere found to elongate b y 2% o f their original lengths. H owever, w hen the stents w ere subjected to high pressure they exhibited foreshortening, see Table 5.1. This table also shows that the larger the radial

expansion the greater the degree o f foreshortening, w hen expanding from the crim ped low profile stent to an expanded larger profile. This characteristic behaviour w as confirm ed by case study 3, w here the M linkthiC k exhibited lower foreshortening w hilst achieving a low er initial lum en diam eter com pared w ith Mlinkthin.

125

Foreshortening (%)

Thin_proxhf

th in d ish f

thick_proxhf

thick_dishf

Case study 1 Case study 2 Case study 3 Case study 4 6 6

-2 5 5 9 4 2

-2 5 2 10

T ab le 5.1

Foreshortening o f Mlinkthm and M linkthiC k after deploym ent, for all o f the case studies evaluated in this thesis.

U nder the influence o f the straight and localised stenotic vessels, both stent designs exhibited foreshortening. In case study 2 and 4, the foreshortening experienced b y M linkth,n and M linkthiC k w as found to be sim ilar for bo th stents, w here both stents w ere expanded to achieve the sam e initial lum en diam eter. This illustrates that the different strut thicknesses, for the sam e stent design, do not have a m ajor influence on the degree o f foreshortening exhibited b y the stents. H owever, in case study 4, foreshortening w as found approxim ately tw ice that observed in case study 2. This is due to the variation in the vessel geom etry,

w hereby the stents conform to the geom etry resulting in greater foreshortening around the localised stenosis.

5.4.3 Dogboning

D ogboning m ay be observed in stents during loading, w hereby the ends expand to a higher radial displacem ent than the m iddle o f the stent, causing the stent to flare at the ends. This behaviour is undesirable as the flare is likely to cause

126

dam age through the arterial wall. studies, see Table 5.2.

D ogboning was evaluated in all o f the case

D ogboning w as not observed in either the Mlinkthm or M linktbiC k during free expansion, in fact, a high degree o f radial retraction w as observed at the proxim al ends o f both stents, see Table 5.2. This behaviour is due to the fact that

there is less m etal surface area at the proxim al end o f the stents structure. Pressure is uniform ly applied on the internal surface o f the stent, and therefore the net force applied in the proxim al end o f the stent is low er than in the other areas o f the stent structure. This consequently led to low er expansion seen at the

proxim al end. The radial displacem ent distributions throughout the stents w ere found to be highly non-uniform . It w as clear that the proxim al end resulted in the least radial displacem ent under free expansion. The proxim al end o f bo th stents underw ent approxim ately 74% o f the m axim um radial displacem ent. negative dogboning w as found to occur to a sim ilar degree in both stents. Sim ilar to free expansion, bo th stent designs show ed negative This

dogboning w hen expanded inside a straight stenotic vessel. H ow ever, the degree o f retraction was lower in the stenotic straight vessel than during free expansion. The proxim al end underw ent approxim ately 88% o f the m axim um radial displacement.

D ogboning (%) Thin_proxhf thin_dishf thick_proxhf thick d ish f

Case study 1 Case study 2 Case study 3 Case study 4

-34.9 -15.1 -15.1

-16 -3.4 -3.4 13.7

-34.5 -14.1 -7.6

-8 -5.4 -0.8 13.1

T ab le 5.2

D ogboning o f M linkthm and M linkthiC k after deploym ent, for all o f the case studies evaluated in this thesis.

127

In case study 4, how ever, under the influence o f a localised stenotic vessel geom etry, a sim ilar degree o f dogboning w as exhibited by M linkthin and M linkthick, see Table 5.2. This is due to the geom etry o f the localised stenosed vessel geom etry, w hereby the stents conform ed to the plaque geom etry, resulting in the ends experiencing m ore radial displacem ent. It is clear therefore that the degree o f dogboning o f a stent is dependant on the strut thickness and design o f the stent but is m ore highly depentant on the host vessel geometry.

5.5 Radial Recoil

Radial recoil is a com mon characteristic o f m etallic stents w hereby the stents radially retract during the recovery o f elastic deform ation after the rem oval o f the balloon pressure. Radial recoil is generally regarded as undesirable since it results in lum en loss after stent deployment. R adial strength is one o f the fundam ental characteristics o f a stent since it defines the ability o f stent to w ithstand the elastic recoil and collapsing force o f a vessel w all and it is therefore needed to ensure the vessel is m aintained open. Radial recoil o f M linkthm and Mlinkthick w as found to be consistently higher for Mlinkthick com pared w ith Mlinkthick for all o f the case studies, see Figure 5.3. It is interesting to note the deviation o f radial recoil w ithin each stent. For exam ple, in case study 1, the Mlinkthin w as found to have radial recoil o f 4-5% in the central and distal regions, w hilst the proxim al end was found to radially recoil by 7%. The same pattern o f radial recoil was observed w ithin the thicker strut stent. W hilst the radial recoil w as consistently higher in Mlinkthin, the vessel geom etry also had an influence on the m agnitude o f the recoil. R adial recoil was found to be significantly higher under the influence o f a stenotic vessel com pared w ith free expansion, see Figure 5.3.

128

F ig u re 5.3

Com parison o f radial recoil o f M linkthin and M linkthiC k; 1) free expansion (Case study 1); w ithin a straight stenotic vessel achieving the same initial lum en diam eter (Case study 2); w ithin a straight stenotic vessel achieving the sam e final lum en diam eter (Case study 3); and w ithin a localised stenotic vessel achieving the same initial lum en diam eter (Case study 4).

5.6

Stresses Induced Within The Stented Vessels

Three case studies w ere carried out to establish the relationship betw een the strut thicknesses and vessel stresses. The stresses induced w ithin the vascular w all by M linkthin and M linkthiC k w ere exam ined in each case study. M igliavacca et al. [86] and Petrini et al. [48] have previously reported the stresses found in a stented vessel during stent deployment. In addition, Lally et al. [41] and H olzapfel et al. [32] have further suggested that these stresses found in the vascular w all m ay be used as a m easure o f restenosis outcome. The differences in the levels o f stresses in the circum ferential, radial and longitudinal direction w ere exam ined for bo th M linkthjn and M linkthiC kIn

addition, the von M ises stress w as also exam ined. The stresses found w ithin the vascular w all at loading w ere taken as a m easure o f acute dam age and those at

129

unloading as a m easure o f long-term damage. The acute dam age is predicted to cause the initial laceration to the arterial w all w hilst the long-term dam age m ust be m inim ized, as the stent scaffolds the vascular w all, to prevent an aggressive long-term vascular response to this dam age and ultim ately restenosis. A ll o f the stress analyses carried out indicated that the tensile circum ferential stresses w ere o f highest m agnitude in com parison to radial, longitudinal and von M ises stresses. Lally et al. [41] investigated the m axim um principal stresses, w hich were noted to act in the circum ferential direction, w hilst H olzapfel et al. [32] also investigated changes in circum ferential stress w ith variations in stent design. It is m ost likely, therefore, that the stresses acting in the circum ferential direction are m ost im portant in term s o f restenosis outcome. G enerally it was noted that as the stent expanded out against the vessel wall, w ith greater expansion o f the vessel lum en higher stresses w ere induced w ithin the vessel wall. The results also show that there is a significant difference betw een acute dam age and long-term dam age, w here the m agnitude o f stresses w ere greatly reduced due to stent and vessel recoil. The reduced stresses found at unloading, or the long-term stresses, are those that w ould contribute to long-term dam age, and hence the long-term restenotic response o f the vessel. Clearly, therefore, theses stresses are o f m ost interest in this study.

5.6.1 Stress Analysis: Case Study 2 Expansion of Mlinkthin and Mlinkthick inside a stenotic artery vessel geometry with arterial material properties, achieving the same initial expanded lumen diameter, 0i thin = 0\ thick

Mlinkthin and MlinkthiC k w ere expanded inside a stenotic vessel w here they
achieved the sam e initial expanded lum en diam eter. The circum ferential, radial and longitudinal stresses in the stented vessels w ere exam ined at loading and unloading, see Figure 5.4. It is clear that the tensile circum ferential stress are the stresses o f highest m agnitude in the stented vessels, w hilst the com pressive longitudinal stresses are low and appear insensitive to loading and unloading.

130

The difference in percentage volum es o f tissue stressed to high levels is m ore prom inent at unloading than at loading, for all o f the stresses induced in the stented vessels. It was found that M linkthick caused significantly higher percentage volum e o f tissue to be stressed at high levels than the M linkthmThis result

correlates w ith the findings o f the ISA R Stereo Trial w here the clinical result show ed that the used o f a thinner strut stent w as associated w ith a significant reduction o f angiographic restenosis than the thicker strut stent [26], w hereby the thinner strut stent resulted in 15% angiographic restenosis com pared w ith 25.8% for the thicker strut stent. It is interesting to note that this significant difference in the volum e stresses for the two stents is due to the fact that as discussed earlier, Mlinkthm resulted in a higher level o f recoil and consequently produced a sm aller expanded final lum en vessel than M linkthiC k- D uring the clinical intervention, x-ray angiography w as used as a visual aid for the operators and it w as left to the operators discretion to achieve the desired lum inal gain observed through angiography [26]. The lum inal gain observed by the operators w ould be the initial expanded lum en diam eter before the pressure was rem oved. D ue to the thin

geom etry struts o f the M linkti,m, the stent has a higher degree o f recoil and consequently results in a low er expanded final lum en diam eter than the M linkthiC k low ering the stresses induced w ithin the vessel. As a result, the stresses w ithin the vessel stented w ith the thinner strut stent w ere found to be low er than the vessel stented w ith a thicker strut stent.

131

S Sthin_proxhf

thin_dishf

H th ic k p ro x h f

B th ic k d is h f

0 ) " > 20 . 0 % tn I 15.0%

c | 25.0%
E

16.0%

>
8 12 .0 %

1 re 0 ) u

S 10.0%
5.0%

o o * 5 8.0% o CL
0 2 4.0%

0 .0%

a > 6 M Pa

o > 0.5 M Pa

o < -lM P a

a < -0 .1 M Pa

< W
E

6 .0 %

o >

= 16.0%
m (u 12.0%

E ra Q a >

E 4.0%

S 8.0%
2 .0 %

3 4.0%

0 ) o

0.0%

o > lM P a

> 0.1 M Pa

c < -40 kPa

c t<

-40 kPa

F ig u re 5.4

Com parison o f tensile circum ferential stresses (A), com pressive radial stresses (B), tensile longitudinal stresses (C) and

com pressive longitudinal stresses (D), at loading (1) and unloading (2), in vessels b y M linkthm and M linkthiC k, w hereby the stents w ere expanded to achieve the sam e initial lumen diam eter o f 3 mm, and at unloading M linkthin and Mlink,hiC k achieved a final lum en diam eter o f 2.28 m m and 2.54 mm respectively.

132

5.6.2 Stress Analysis: Case Study 3 Expansion of Mlinkthin and MlinkthiC k inside a stenotic artery vessel geometry with arterial material properties as in Case Study 2, however, achieving the same final lumen diameter,
0 F thin = 0 F thick

Finite elem ent sim ulations w ere carried out w here M linkthm and M linkthiC k w ere expanded inside a stenotic vessel identical to that u sed in the previous analyses b u t w here the stents w ere allow ed to recoil to achieve the sam e expanded final lum en diameter. The circum ferential, radial and longitudinal stresses in the The

stented vessels w ere exam ined at loading and unloading, see Figure 5.5.

tensile circum ferential stresses w ere again found to be the stresses o f highest m agnitude in the vessels. H owever, in this case, the difference in percentage o f tissue volum e dam age at loading is m ore prom inent than at unloading. A s it w as established

earlier that stresses increase w ith luminal gain, this prom inent difference in the stresses found at loading was due to the fact that the Mlinkthin had to be expanded m ore to produce a higher initial lum en diam eter than the M linkthiC k, as Mlinkthm recoiled m ore to achieve the same final lum en diam eter. A t unloading, the

am ounts o f volum e stressed at specific m agnitudes w ere analysed. In this case, it w as clear that Mlinkthin resulted in a m arginally higher percentage volum e o f tissue stressed to high levels w hen com pared w ith M linkthiC k, see Figure 5.5. This resulting higher volum e dam age found in M linkthin m ay be explained b y the fact that the vessel load to be supported, to m aintain the vessel open at the sam e vessel lum en diameter, is supported b y a low er area in the thin strut stent com pared w ith the thick strut stent since the artery-stent contact area is low er for the M linkthin. In addition, the m agnitude o f the pressure used for stentss expansion for these analyses w as higher for Mlinkthin at 8.2 M P a com pared w ith 2.2 M Pa for

Mlinkthick- Clearly, this difference in the pressure applied to the stents causes the
difference in the stresses found in the stented vessel w all on loading.

133

B Sthin_proxhf

Hthin_dishf

B th ickp ro xh f

thick dishf

E o

a )

0
| 30.0%

= 80 .0 %

>
60 . 0 %

O > W
55

8 20 .0 %

S i
5 20 .0 %
CL

0 .0 %

ct > 0.8 M Pa

a > 0.3 M Pa

a < -0 .2 M Pa

a < -0 .1 M Pa

E 80.0% 60.0% 2 40.0%

3.0%
1

20 .0 % c
a u

O O

U )

2.0%

I 1.0%

0 .0 %

o > 200 kPa

a > 40 kPa

a < -40 kP a

a < -40 kPa

F ig u re 5.5

Com parison

of

(A)

tensile

circum ferential

stresses;

(B)

com pressive radial stresses; (C) tensile longitudinal stresses; and (D) com pressive longitudinal stresses, at loading (1) and unloading (2), in stented vessels b y M linkthin and M linkthiC k, w hereby they w ere expanded to achieve the sam e final lumen diam eter o f 2.28 m m , w hereby at loading M linkthm achieved the inital lum en diam eter o f 3 m m and 2.48 m m for M linkthiC k.

134

5.6.3 Stress Analysis: Case Study 4 Expansion of Mlinkthin and MlinkthiC k inside a localised stenotic artery, achieving the same initial expanded lumen diameter, 0 i thm= 0 i thick

Finite elem ent sim ulations were carried out w here Mlinkthin and Mlinkthick were expanded inside a localized stenotic vessel achieving the sam e initial lum en diameter. The circum ferential, radial and longitudinal stresses in the stented In these

vessels w ere exam ined at loading and unloading, see F igure 5.6.

analyses, the tensile circum ferential stresses w ere also found to be the stresses o f highest m agnitude in both stented vessels. As for case study 2, the Mlinkthick caused greater volum es o f tissue to be stressed at high levels com pared w ith Mlinkthin- It w as noted that the m agnitudes at w hich the vascular w all w as stressed w ere significantly higher than the previous analyses, particularly on loading. The type o f plaque and the geom etry o f the

vessel w all greatly influenced the results in term s o f the m agnitude and the distribution o f the resulting stresses found in the vascular wall. The pressures

used to expand the stents inside the vessels w ere approxim ately five tim es the pressures used in the previous analyses. These high pressures resulted in the

significantly higher stresses induced w ithin the vessel w all, see Figure 5.6. These analyses illustrate the relationship betw een the stresses induced in the vascular w all and the results obtained in the clinical study, the ISA R STEREO Trial [26]. A lthough the stress analyses do not identify the exact percentage o f restenosis they do show that there is a m uch greater volum e o f tissue stressed to elevated levels in the vessels stented by Mlinkthick particularly in the

circum ferential direction at unloading.

This difference in the stresses found

w ithin the stented vessels can be identified as the potential m echanical stim uli for in-stent restenosis. Theses stresses, w hich are m arkedly different for the two

stents on unloading, are deem ed to act as a long-term stim ulus for in-stent restenosis.

135

 thin dishf thick dishf

a > 6 0 M Pa

a > 1 M Pa a < - 1 0 M P a a < -0 .

a > 2 0 M Pa

a > 0 .5 M Pa

a < -0 .5 M P a

a < -0 .1 M Pa

F ig u re 5.6

C om parison (B)

of

(A) radial

tensile stresses;

circum ferential (C) tensile

stresses;

com pressive

longitudinal

stresses; and (D) com pressive longitudinal stresses at loading (1) and unloading (2), in stented vessels b y M linkthm and Mlinkthick The stents w ere expanded to achieve the same initial lum en diam eter o f 3.18 m m , and at unloading M linkthin and Mlinkthick achieved a final lum en diam eter o f 2.44 m m and 2.52 m m , respectively.

136

5.7 Preclinical Testing Of Stents And Stent Design

Intravascular stenting has em erged over the years as am ong the m ost successful treatm ent o f coronary artery disease. H owever, this prom ising intervention Stent

presents in-stent restenosis as the m ajor lim itation to stenting procedure.

design has been identified as a m ajor factor in determ ining restenosis outcome. This finite elem ent study is used as a preclinical tool to investigate the expansion o f stents com plex structure and identify the influence o f stent design on the resulting stresses found in the stented arteries, thus direct com parison w as m ade to clinical study b y K astrati et al. [26], The stents structure after deploym ent is highly dependent on the geom etry o f the stenosed artery. The com plex geom etry o f the sten ts structures resulted in a non-uniform radial displacem ent distribution. Thinner strut stent

exhibits greater radial recoil than the thicker strut stent under free expansion. This m ay serve as a disadvantage in obtaining the adequate lum inal gain. Both stents w ere then expanded inside a various stenosed arteries to exam ine the effects o f different strut thickness stents on their structures expansion and the resulting stresses induced w ithin the stented vessels. Since radial recoil p lay a m ajor role in obtaining the sufficient lum inal gain, the stresses w ithin the stented vessels were exam ined at the end o f loading, w here the initial lum inal gain is achieved and at the end o f unloading, w here the lum inal loss is observed as radial recoil takes place. These stresses in the form er are believed to act as a stim ulus for acute dam age w hile long-term stresses are o f m ost interest in this study since these stresses m ust contribute to the long-term grow th o f in-stent restenosis. The acute loading o f the vessel m ay cause dam age to the vessel how ever acute loading is also observed in balloon angioplasty. In balloon angioplasty restenosis is not due to neotim al grow th but prim arily due to vessel recoil. Therefore, it m ay be the recoil o f the vessel that low ers the stresses in balloon angioplasty such that the vessel does not respond w ith high neointim al growth. Clearly therefore, for stenting there m ay be an optim al level o f vessel expansion that achieves a sufficiently large lum en to restore blood flow but does

137

not illicit an aggressive restenotic response and subsequently restenosis in stented vessels. The high radial recoil o f the thin strut stent m ay enable this optimum

vessel diam eter to be m ore easily achieved when com pared with the Mlinkihick and therefore result in the thin strut stent having a low er propensity for restenosis. Overall therefore, there may be an optim um expansion pressure for all stent designs that is required to achieve this optim um diam eter and w hich depends upon the vessel wall geom etry as well as the arterial m aterial properties.

138

Chapter 6

CONCLUSIONS

6.1

Main Findings

The m ain objective o f this thesis was to identify the m echanical stim uli for in stent restenosis. This study focused on exam ining the role o f strut thickness on restenosis outcome. Finite elem ent analyses w ere used to exam ine the expansion o f a thinner strut stent and a thicker strut stent in various stenosed vessel geometry. Stress analyses w ere carried out on the stented vessels and a

com parison w as m ade to the clinical data. The m ain conclusions o f this study are as follows:-

1. There is a significant difference betw een the expansion characteristics o f a stent on free expansion com pared w ith the stent w hen it is expanded inside a stenotic vessel. Free expansion o f the stents in this study show ed no

foreshortening and gave rise to elongation w hereas elongation was not observed w hen the stent w as expanded inside a stenotic vessel. 2. The geom etry o f the stenotic vessel has a significant influence on a stents expansion. In this study dogboning o f the stent w as observed w hen it was expanded inside a localised stenotic vessel w hilst negative dogboning was seen under free expansion and under the influence o f an idealised straight stenotic vessel.

139

3. For the sam e stent design, a thinner strut stent has a higher level o f recoil than a thicker strut stent. This leads to low er final lum inal gain in a vessel stented by a thinner strut stent, w hen both stents are expanded to achieve the sam e initial lum inal gain. Since the analyses show that the m agnitude o f stresses increases w ith lum inal gain, the thicker strut stent resulted in a higher lum inal gain and hence higher stresses in th e vascular wall. 4. H igher stresses occurred in the vessel w all o f the thicker strut stent for the same initial lum en diameter. Since available clinical data indicates that a thicker strut stent has a greater propensity for restenosis, this indicates that w all stresses m ay act as the stim ulus for restenosis. The difference in

stress m agnitudes and volum es o f tissues stressed at high levels in the vessels stented w ith the different strut stents w as found to be m ost pronounced in the circum ferential direction b u t w as also evident in the radial and longitudinal directions. The tensile circum ferential stresses m ay therefore be the m ain stim ulus for in-stent restenosis. 5. The analyses indicate that long-term dam age, or stresses on stent unloading, m ay be the critical stresses for stim ulating the progression o f restenosis. 6. The stenting analyses, w here both stents w ere expanded such that they achieved the sam e final lum en diameter, suggest that thicker strut stent, or any stent design, m ay have the potential to be expanded to an optim al diam eter w hereby the proliferation o f neointim a is m inim ised. Therefore, the use o f preclinical testing tools, such as finite elem ent m odelling, could be used to predeterm ine the level o f expansion that a particular stent should be subjected to in order to achieve the optim um lum inal gain and m inim ise the stim uli for in-stent restenosis. 7. V essel w all stresses, induced through stent expansion, are highly

dependant on the stent design, the lum inal pressure, the vessel geom etry and the vessel m aterial properties.

140

6.2

Future Work

The m ain findings o f this thesis have determ ined th e role o f strut thickness on restenosis outcome. This know ledge could im prove the developm ent o f

intravascular stenting. The following recom m endations are proposed:-

1. Cell tests could be carried out on vascular sm ooth m uscle and endothelial cells to identify the specific level o f stress at w hich neointim a starts to proliferate in an aggressive manner. This data could then be used in finite elem ent studies to determ ine the optim um vessel diam eter w hich w ould not reach those critical stress levels for various stent designs. The balloon pressures required to achieve these vessel diam eters, in different vessel geom etries w ith different m aterial properties, could also be established using finite elem ent techniques. 2. Realistic stenosed vessel geom etries could be obtained from in-vivo im aging, such as Intravascular U ltrasound, and u sed in the finite elem ent preclinical tests o f stents to determ ine the stent m ost suited to particular host vessel geom etries. 3. Finite elem ent analyses are highly dependent upon the accuracy o f the m aterial m odels used in the analyses. F urther tests could be carried out on hum an atherosclerotic plaque tissue to add to the lim ited data currently available. This w ould enable the range o f properties o f plaque tissues to be determ ined and the influence that changes in these properties have on the outcom e o f stenting procedures to be determ ined. Further data w ould also enable m ore accurate constitutive m odels for the tissue to be im plem ented in preclinical tests o f intravascular stents. 4. U ltim ately, preclinical testing could lead to patient-specific analyses capable o f determ ining the optim um stent design and stent deploym ent protocol that reduces the likelihood o f in-stent restenosis for a particular stenotic vessel w ith specific m aterial properties.

141

REFERENCES
[1] W ard, M ., (2005), Im plem entation o f secondary prevention o f coronary heart disease in prim ary care, H eartw ise, Irish H eart Foundations m agazine, V olum e 8, Issue 1, Spring 2005.

[2]

Central Statistics Office, [online], http://w w w .cso.ie/statistics/ (A ccessed 7th M arch 2005).

[3]

Irish H eart Foundation, Cardiovascular M ortality R ates, [online], htl>)://w w w .inshhcai1.ie/iopen24/catalog/pub//C V D M ortalit / R ates04.pdf (A ccessed 2lld M arch 2005).

[4]

H eart Centre, [online], http://heart.healthcentersonline.com (A ccessed 12th D ecem ber 2004).

[5]

Rankin, J.M. and Penn, I.M ., (1998), C oronary stenting: A global perspective, Journal o f the A m erican College o f Cardiology, pp.22-25.

[6]

A m erican H eart A ssociation, [online], http://w w w .am ericanheart.org/presenter.ihtm l?identifier=:1200026 (A ccessed 8th D ecem ber 2004).

[7]

M edline, [online], http://w w w .nlm .nih ov/medlm enlus/ency/nresentations/100160 I .him (A ccessed 15th N ovem ber 2004).

[8]

PTCA , [online], http://w w w .ptca.org/ (A ccessed 11th N ovem ber 2004).

[9]

H ospital In-Patient Enquiry (H IPE) System http://w w w .esri,ie (A ccessed 14th A pril 2005)

U nit,

[online],

[10]

Colom bo, A. and Tobis, J., (2001), Techniques in C oronary Stenting, (2nd ed.), M artin D unitz Lt., U nited K ingdom .

A rtery

[11]

Banka, V.S., K ochar, G.S., M aniet, A .R., D .O ., V oci, G., (1993), Progressive coronary dilation: A n angioplasty technique that creates controlled arterial injury and reduces com plications, A m erican H eart Journal, Vol. 125, pp. 61-71.

142

[12]

. tV i Guidant, http://w w w .guidant.com (A ccessed 15' D ecem ber 2004).

[13]

A dam , [online], http://w w w .adam .com / (A ccessed 20th N ovem ber 2004).

[14]

Edelm en, E.R. and Rogers, C., (1998), Pathobiologic responses to stenting, A m erican Journal o f Cardiology, Vol. 81, pp. 4E-6E.

[15]

Sigwart, V.J., Puel, J. and M irkovitch, V ., (1987), Intravascular stents to prevent occlusion and restenosis after translum inal angioplasty, N Engl J M ed, Vol. 316, pp. 701-706.

[16]

M oscucci, M ., M ansour, K.A., Kent, K .C., K untz, R .E., Senerchia, C., Baim , D.S. and Carrozza, J.P., (1994), Peripheral vascular com plication o f directional coronary atherectom y and stenting: Predictors, m anagem ent, and outcom e, A m erican Journal o f Cardiology, V ol. 74 (5), pp. 448-453.

[17]

David, H., 1998, Renal angioplasty and stenting, A sia Pacific H eart Journal, Vol. 7 (3), pp. 215-218.

[18]

Sachar, R., Yadav, J.S. and Roffi, M ., (2004), Severe B ilateral Carotid Stenosis, Journal o f the A m erican College o f C ardiology, Vol. 43 (8), pp. 1358-1362.

[19]

Levin, T., H ollow ay, S. and Feldm an, T., (2000), U se o f intracoronary stents for the prevention o f restenosis, Interventional Cardiology.

[20]

Fischm an, D .L., Leon, M .B. and Baim , D .S., (1994), A random ized com parison o f coronary-stent im plantation w ith balloon angioplasty in patients w ith coronary artery disease, N ew E ngland Journal o f M edicine, Vol. 331, pp. 496-501.

[21]

Bult, H., (2000), Restenosis: a challenge for pharm acology, Trends in Pharm acological Sciences, Vol. 21 (7), pp. 274-279.

[22]

M cClean, R. and Eigler, N .L., (2002), Stent design: im plications for restenosis, Review s in Cardiovascular M edicine, Vol. 3, pp. S16-S22.

[23]

Hausleiter, J., Kastrati, A, M ehilli, J., Schuhlen, H., Pache, J., Dotzer, F., D irschinger, J. and Schomig, A., (2002), Predictive factors for early cardiac events and angiographic restenosis after coronary stent placem ent

143

in small coronary arteries, Journal o f the Cardiology, Vol. 40, pp. 882-889.

A m erican

College o f

[24]

Escaned, J., Goicolea, J., Alfonso, F., Perez-V izcayno, M .J., H ernandez, R., Fem andez-O rtiz, A., Banuelos, C. and M acaya, C., (1999), Propensity and m echanism o f restenosis in different coronary stent designs, Journal o f the A m erican College o f Cardiology, Vol. 34 (5), pp. 1490-1497.

[25]

H offm ann, R., M intz, G.S., Haager, P.K ., B ozoglu, T., Grube, E., M ichael Gross, B eythien, C., M udra, H., vom Dahl, J. and H anrath, P., (2002), R elation o f stent design and stent surface m aterial to subsequent in-stent intim al hyperplasia in coronary arteries determ ined b y intravascular ultrasound, A m erican Journal o f Cardiology, V ol. 89, pp. 1360-1364.

[26]

Kastrati, A., M ehili, J. and D irschinger, J., (2001), Intracoronary stenting and angiographic results: Strut thickness effect on restenosis outcom e (ISAR-STEREO ) trial, Circulation, Vol. 103, pp. 2816-2821.

[27]

D um oulin, C. and Cochelin, B., (2000), M echanical behaviour m odelling o f balloon-expandable stents, Journal o f Biom echanics, Vol. 33, pp. 1461-1470.

[28]

Chua, S. N. D, M ac D onald, B. J. and H ashm i, M . S. J., (2002), Finite elem ent sim ulation o f stent expansion, Journal o f M aterials Processing Technology, Vol. 120, pp. 335-340.

[29]

M igliavacca, F., Petrini, L., Colom bo, M ., A uricchio, F. and Pietrabissa, R., (2002), M echanical behaviour o f coronary stents investigated through the finite elem ent m ethod, Journal o f B iom echanics, Vol. 35, pp. 803 811.

[30]

Petrini, L., M igliavacca, F., A uricchio, F. and D ubini, G., (2004), N um erical Investigation o f the Intravascular C oronary Stent Flexibility, Journal o f Biom echanics, Vol. 37, pp. 495-501.

[31]

Lally, C., D olan, F. and Prendergast, P.J., (2005), C ardiovascular stent design and vessel stresses: a finite elem ent analysis, Journal o f B iom echanics, Vol. 38, pp. 1574-1581.

[32]

Holzapfel, G.A., Stadler, M. and Gasser, T.C., (2005), Changes in the m echanical environm ent o f stenotic arteries during interaction w ith stents:

144

com putational assessm ent o f param etric stent designs, A M SE, Vol. 127, pp. 166-180.

[33]

H um phrey, J.D., (2002), Cardiovascular solid m echanics: cells, tissues, and organs, Springer, N ew York.

[34]

H ayashi, K., (2003), M echanical properties o f soft tissues and arterial walls, B iom echanics o f soft tissue in cardiovascular system s, W ien, Springer.

[35]

Ozolanta, I., Tetere, G., Purinya, B. and K asyanov, V ., (1998), Changes in the m echanical properties, biom echanical contents and w all structure o f the hum an coronary arteries w ith age and sex, M edical Engineering and Physics, Vol. 20, pp. 523-533.

[36]

Frederick, H .S., (1994), Biom aterials, M edical D evices and Tissue Engineering: A n Intergrated A pproach, C hapm an & Hall, U nited Kingdom.

[37]

van A ndel, C.J., Pistecky, P.V. and Borst, C., (2003), M echanical properties o f porcine and hum an arteries: im plications for coronary anastom otic connectors, Armais o f Thoracic Surgery, V ol. 76, pp. 58-65.

[38]

A hlgren, A.R., A strand, H ., Sandgren, T., V em ersson, E., Sonesson, B. and Lanne, T., (2001), D ynam ic behaviour o f com m on fem oral artery: age and gender o f m inor im portance, U ltrasound in M edicine & Biology, Vol. 27 (3), pp. 181-188.

[39]

Cheng, K .S., Tiwari, A., Baker, C.R., M orris, R., H am ilton, G. and Seifalian, A.M ., (2002), Im paired carotid and fem oral viscoelastic properties and elevated intim a-m edia thickness in peripheral vascular disease, A therosclerosis, Vol. 164 (1), pp. 113-120.

[40]

V ito, R.P. and D ixon, S.A., (2003), Blood vessel constitutive m odels 1995-2002, A nnual R eview o f B iom edical Engineering, Vol. 5, pp. 413 439.

[41]

Lally, C., Reid, A.J. and Prendergast, P.J., (2004), Elastic behaviour o f porcine coronary artery tissue under uniaxial and equibiaxial tension, A nnals o f B iom edical Engineering, Vol. 32 (10), pp. 1355-1364.

145

[42]

Holzapfel, G.A., Gasser, T.C. and Stadler, M ., (2002), A structural model for the viscoelastic behaviour o f arterial w alls: continuum form ulation and finite elem ent analysis, European Journal o f M echanm ics A /Solids, Vol. 21, pp. 441-463.

[43]

Rachev, A. and Greenwalf, S.E., (2003), R esidual strains in conduit arteries, Journal o f Biom echanics, Vol. 36 (5), pp. 661-670.

[44]

H olzapfel, G.A., Gasser, T.C. and Stadler, M ., (2000), A new constitutive fram ework for arterial w all m echanics and a com parative study o f material m odels, Journal o f Elasticity, Vol. 61, pp. 1-48.

[45]

M enzel, A., (2005), M odelling o f anisotropic grow th in biological tissues: a new approach and com putational aspects, Biom echanical M odel M echanobiology, Vol. 3, pp. 147-171.

[46]

Lally, C. and Prendergast, P.J., (2003), A n investigation into the applicability o f a m ooney-rivlin constitutive equation for m odelling vascular tissue in cardiovascular stenting procedures, In Proceedings o f the International Congress on C om putational Biom echanics, Zaragoza, Spain, pp. 542-550.

[47]

Fung, Y .C., (1993), Biom echanics: M echanical Properties o f Living Tissues , Springer, N ew York.

[48]

Petrini, L., M igliavacca, F., Dubini, G. and Auricchio, F., (2003), N um erical Analysis o f V ascular Stents Exploiting Shape-M em ory-A lloy B ehavior, 16th A IM ETA Congress o f Theoretical and A pplied M echanics.

[49]

Carm ines, D .V ., M cElhaney, J.H. and Stack, R., (1991), A piece-w ise non-linear elastic stress expression o f hum an and pig coronary arteries tested in vitro, Journal o f Biom echanics, Vol. 24, pp. 899-906.

[50]

Rogers, C., Tseng, D .Y., Squire, J.C. and Edelm an, E.R., (1999), B alloon-artery interactions during stent placem ent: a finite elem ent analysis approach to pressure, com pliance, and stent design as contributors to vascular injury, Circulation Research, Vol. 84, pp. 378-383.

[51]

Berry, J.L., M anoach, E., M ekkaoui, C., R olland, P.H., M oore, J.E. Jr, Rachev, A., (2002), H em odynam ics and w all m echanics o f a com pliance

146

m atching stent: in vitro and in vivo analysis, Journal o f V ascular Interventional Radiology, Vol. 13, pp. 97-105.

[52]

Chua, S. N. D., M acD onald, B.J. and H ashm i, M .S.J., (2004), Finite elem ent sim ulation o f slotted tube (stent) w ith the presence o f plaque and artery by balloon expansion, Journal o f M aterials Processing Technology, pp. 1772-1779.

[53]

A uricchio, F., Di Loreto, M. and Sacco, E., (2001), Finite-elem ent analysis o f a stenotic artery revascularisation through a stent insertion, Com puter M ethods in Biom echanics and B iom edical Engineering, Vol. 4, pp.249-263.

[54]

Loree, H.M ., Grodzinsky, A.J., Park, S.Y., G ibson, L.J. and Lee, R.T., (1994), Static and circum ferential tangential m odulus o f human atherosclerotic tissue, Journal o f Biom echanics, V ol. 27, pp. 195-204.

[55]

M acaya, C., Serruys, P.W ., Ruygrok, P., Suryapranata, H., M ast, G., Klugm ann, S., Urban, P., den Heijer, P., K och, K , Simon, R., M orice, M .C., Crean, P., Bonnier, H., W ijns, W., D anchin, N ., Bourdonnec, C. and M orel, M .A., (1996), Continued Benefit o f C oronary Stenting Versus Balloon A ngioplasty: O ne Y ear Clinical Follow -U p o f Benestent Trial, Journal o f the American College o f Cardiology, Vol. 27, pp. 255-261.

[56]

K iem eneij, F., Serruys, P.W ., M acaya, C., R utsch, W., H eyndrickx, G., A lbertsson, P., Fajadet, J., Legrand, V., M atem e, P., Belardi, J., Sigwart, U., Colom bo, A., Goy, J.J., Disco, C.M. and M orel, M .A ., (2001), Continued benefit o f coronary stenting versus balloon angioplasty: fiveyear clinical follow-up o f Benestent-I trial, Journal o f the A m erican College o f Cardiology, Vol. 37, pp. 1598-1603.

[57]

Rubartelli, P., N iccoli, L., Verna, E., G iachero, C., Zim arino, M., Fontanelli, A., V assanelli, C., Cam polo, L., M artuscelli, E. and Tom m asini, G., (1998), Stent im plantation versus balloon angioplasty in chronic coronary occlusions: results from the G ISSO C trial, Journal o f the Am erican College o f Cardiology, Vol. 32, pp. 90-96.

[58]

Pratsos, A., Fischm an, D.L. and Savage, M .P., (2001), Restenosis in SaphenousV ein G rafts, Current Interventional C ardiology Reports, Vol. 3, pp. 287-295.

147

[59]

Schwartz, R.S. and Henry, T.D., (2002), P athophysiology o f coronary artery restenosis, R eview o f Cardiovascular M edicine, Vol. 3 (5), pp. S49.

[60]

Zim m erm ann, A., Pollinger, B., Rieber, J., K onig, A ., Erhard, I., K rotz, F., Sohn, H.Y., K antlehner, R., H aim erl, W ., D uhm ke, E., Leibig, M ., Theisen, K., K lauss, V. and Schiele, T.M ., (2005), Early tim e course o f neointim a form ation and vascular rem odelling follow ing percutaneous coronary intervention and vascular brachytherapy o f in-stent restenotic lesions as assessed by intravascular ultrasound analysis, Journal o f Cardiology, Vol. 94 (4), pp. 239-246.

[61]

M oreno, P.R., Palacios, I.F., Leon, M .N ., R hodes, J., Fuster, V. and Fallon, J.T., (1999), H istopathologic com parison o f hum an coronary in stent and post-balloon angioplasty restenotic tissue, A m J. Cardiol, Vol. 84 (4), pp. 462-466.

[62]

H om e, M cD., (2005), O verview o f hem ostasis and throm bosis; current status o f antithrom botic therapies, Throm bosis Research, in press.

[63]

Bauters, C. et al., (1996), M echanism and prevention o f restenosis: from experim ental m odels to clinical practice, C ardiovascular Research, Vol. 31, pp. 835-846.

[64]

Bauters, C. and Isner, J.M ., (1997), The biology o f restenosis, Progress in Cardiovascular D iseases, Vol. 40 (2), pp. 107-116.

[65]

Lowe, H.C., Schwartz, R.S., M ac N eill, B.D., Jang, I.K., H ayase, M., Rogers, C. and Oesterle, S.N., (2003), The porcine coronary m odel o f in stent restenosis: current status in the era o f drug-eluting stents, Catheterisation and Cardiovascular Interventions, Vol. 60, pp. 515-523.

[66]

Dangas, G. and Fuster, V., (1996), M anagem ent o f restenosis after coronary intervention, A m H eart J, Vol. 132, pp. 428-436.

[67]

Schwartz, R.S., Chronos, N.A. and V irm ani, R., (2004), Preclinical restenosis m odels and drug-eluting stents, still im portant, still m uch to learn, Journal o f the A m erican College o f Cardiology, Vol. 44 (7), pp. 1373-1385.

[68]

W aksm an, R., (2002), D rug-eluting stents, from bench to C ardiovascular R adiation M edicine, Vol. 3, pp. 226-241.

bed,

148

[69]

M orton, A.C., Crossm an, D. and Gunn, J., (2004), The influence o f physical stent param eters upon restenosis, Pathologie B iologie, Vol. 52, pp. 196-205.

[70]

Schachter, L.G. and Bakarat, A.I., (2001), C om putational study o f arterial flow disturbance induced by intravascular stents, The A SM E B ioengineering Conference, Vol. 50.

[71]

Carter, A.J., Lee, D.P. and Yeung, A.C., (2001), M etaling w ith new stent designs, C atheterisation and Cardiovascular Interventions, Vol. 53, pp. 426-428.

[72]

Colom bo, A., Stankovic, G. and M oses, J.W ., (2002), Selection o f coronary stents, Journal o f the A m erican C ollege o f Cardiology, Vol. 40 (6), pp. 1021-1033.

[73]

Stoeckel, D., Bonsignore, C. and Duda, S., (2002), A survey o f stent designs, M in Invas Ther & Allied Technology, Vol. 11 (4), pp. 137-147.

[74]

H autm ann, H. and Huber, R.M ., (1996), Stent flexibility: an essential feature in the treatm ent o f dynam ic airw ay collapse, European R espiratory Journal, Vol. 9, pp. 609-611.

[75]

K untz, R.E., Safian, R.D., Levine, M .J., R eis, G.J., D iver, D.J. and Baim, D.S., (1992), N ovel approach to the analysis o f restenosis after the use o f three new coronary devices, Journal o f A m erican College o f Cardiology, Vol. 19, pp. 1493-1499.

[76]

Ferrero, V., R ibichini, F., H eyndrickx, G.R., D e B ruyne, B., Piessens, M ., Carlier, S., Buchi, M ., M atullo, G., V assanelli, C. and W ijns, W ., (2004), A ngiographic and three-dim ensional intravascular ultrasound analysis o f com bined intracoronary b eta radiation and self-expanding stent im plantation in hum an coronary arteries, A m J Cardiol, Vol. 94, pp. 1237-1242.

[77]

Lansky, A.J., M intz, G.S., M ehran, R., Popm a, J.J., Pichard, A.D., Kent, K.M ., Satler, L.F. and Leon, M .B., (1998), Insights into the m echanism o f restenosis after PTC A and stenting, Indian H eart J., Vol. 50, Suppl 1, pp. 104-108.

[78]

Carrozza, J.P. Jr, Hosley, S.E., Cohen, D.J. and Baim , D .S., (1999), In Vivo A ssessm ent o f Stent E xpansion and Recoil in N orm al Porcine

149

Coronary Arteries: D ifferential O utcom e by Stent D esign, Circulation, Vol. 100, pp. 756-760.

[79]

Stankovic, G., Briguori, C., Sarais, C., Pagnotta, P., Liistro, F., M ontorfano, M ., Chieffo, A., Sgura, F., C orvaja, N., A lbiero, R., Toutoutzas, C., Bonizzoni, E., D i M ario, C. and Colom bo, A., (2001), Effect o f Strut Thickness on A ngiographic R estenosis Rate A fter Coronary Stenting, Journal o f the American C ollege o f Cardiology, Vol. 40, pp. 403-409.

[80]

Costa, M .A ., Sabate, M ., Kay, I.P., de Feyter, P.J., K ozum a, K , Serrano, P., de Valk, V ., Albertal, M ., Ligthart, J.M ., D isco, C., Foley, D.P. and Serruys, P.W ., (2000), Three-D im ensional Intravascular U ltrasonic V olum etric Q uantification o f Stent R ecoil and N eointim al Form ation o f Two N ew G eneration Tubular Stents, Journal o f the A m erical College o f Cardiology, Vol. 85, pp. 135-139.

[81]

Taylor, A .J., G orm an, P.D., K enw ood, B., H udak, C., Tashko, G. and V irm ani, R., (2001), A com parison o f four stent designs on arterial injury, cellular proliferation, neointim a form ation, and arterial dim ensions in an experim ental porcine m odel, C atheterisation and Cardiovascular Interventions, Vol. 53, pp. 420-425.

[82]

Edelm an, E.R ., Seifert, P., G roothuis, A ., M orss, A., B om stein, D. and Rogers, C., (2001), G old-C oated N IR Stents in Porcine Coronary A rteries, C irculation, Vol. 103, pp. 429-434.

[83]

Pache, J., K astrati, A., M ehilli, J., Schuhlen, H ., D otzer, F., H ausleiter, J., Fleckenstein, M ., N eum ann, F.J., Sattelberger, U ., Schm itt, C., M uller, M ., D irschinger, J. and Schomig, A., (2003), Intracoronary Stenting and A ngiographic Results: Strut Thickness Effect on Restenosis O utcom e (ISA R-STEREO -2) Trial, Journal o f the A m erican College o f Cardiology, Vol. 41.

[84]

Chua, S. N. D., M acD onald, B.J. and H ashm i, M .S.J., (2003), Finite elem ent sim ulation o f stent and balloon interaction, Journal o f M aterials Processing Technology, Vol. 143-144, pp. 591-597.

[85]

M cG arry, J.P., O Donnell, B.P., M cH ugh, P.E. and M cG arry, J.G., (2004), A nalysis o f the m echanical perform ance o f a cardiovascular stent design based on m icrom echanical m odelling, C om putational M aterials Science, Vol. 31, pp. 421-438.

150

[86]

M igliavacca, F., Petrini, L., Colom bo, M ., A uricchio, F. and Pietrabissa, R., (2003), D eploym ent o f an intravascular stent in coronary stenotic arteries: a com putational study, Sum m er B ioengineering Conference, 25 29 June.

[87]

Liang, D .K ., Yang, D.Z., Qi, M ., W ang, W .Q ., (2005), Finite elem ent analysis o f the im plantation o f a balloon-expandable stent in a stenosed artery, International Journal o f C ardiology, V ol. 104, pp. 314-318.

[89]

Serruys, P.W . and K utryk, M .J.B ., (2000), H andbook o f C oronary Stents, (3rd ed.), M artin D unitz Lt., U nited K ingdom .

[90]

W ilm s, G., v an Calenbergh, F., Stockx, L., D em aerel, P., van Loon, J. and Goffin, J., (2000), Endovascular treatm ent o f a ruptured paraclinoid aneurysm o f the carotid syphon achieved using endovascular stent and endovascular coil placem ent, A JN R A m J N euroradiol, Vol. 21, pp. 753 756.

[91]

K ereiakes, D .J., Cox, D.A., H erm iller, J.B., M idei, M .G ., Bachinsky, W .B., N ukta, E.D., Leon, M .B., Fink, S., M arin, L. and Lansky, A.J., (2003), U sefullness o f a cobalt chrom ium coronary stent alloy, The A m erican Journal o f Cardiology, V ol. 92 (4), pp. 463-466.

151

Appendix A

Simulation of a Full Model of MlinkthiC k inside a Stenotic Straight Vessel.

A.1 Introduction
Intravascular stenting procedure involves in the deploym ent o f the stent inside a stenotic coronary artery. Full three-dim ensional m odel o f the stent expanded

w ithin a stenotic vessel w ere developed to determ ine m ore realistically the expansion o f the stent and to quantify the stresses induced w ithin stented stenotic coronary arteries. The finite elem ent m odelling o f this procedure is a challenging study that involves large displacem ents and deform ation, geom etric and m aterial non-linearity. D epending on the com plications o f the m odel, finite elem ent

sim ulation can be rather expensive on tim e consum ption. The purpose o f this w ork is to study the m ovem ent o f the artery vessel relative to the stent during stent expansion inside a stenotic vessel. This is for the purpose o f providing justification in dividing the m odel into tw o halves m odels for the purpose o f saving com putational tim e and resources.

A.2 Materials and Method

Full three-dim ensional m odel o f m linkthiC k w as developed to expand inside a stenotic coronary artery w hich w as m odelled as a straight vessel. The

m odel com posed o f two bodies, the stent (m linkthiC k) and the stenotic artery. The stent has an internal radius o f 0.72 m m and outer radius o f 0.86 mm. The straight

152

vessel has an internal radius o f 1 m m and thickness o f 0.8 m m , see Figure A .I. A uniform , linearly increasing radial pressure (P) o f 10.5 M P a w as applied as a surface load to the internal surface o f the stent to deploy the stent w ithin the vessel, w hich was subsequently removed. The stent is m ade o f 316L stainless steel w ith Y oungs m odulus o f 196 GPa and Poisson ratio o f 0.3. The inelastic stress-strain behaviour was

incorporated into the analysis through a V on M ises-H ill plasticity m odel w ith isotropic hardening using data obtained from M urphy et al. [88]. The non-linear stress-strain relationship o f the arterial w all w as described using a third order M ooney-R ivlin hyperelastic constitutive equation determ ined by fitting to data from uniaxial and equibiaxial tension tests o f porcine coronary tissue by Lally et al. [41].

Figure A .l

Finite elem ent m esh o f m linkthiC k inside a stenotic vessel.

153

D ue to symmetry, only 120 degrees in circum ferential direction (onethird) w as m odelled. Cyclic sym m etry boundary conditions w ere im posed on the nodes o f the stent and the stenotic vessel in the circum ferential plane o f symmetry. B oth ends o f the stenotic vessel w ere constraint in the longitudinal

direction. Two nodes in the m iddle o f the stent w ere constraint in the longitudinal direction and one node o f the stenotic vessel w as constraint in the circum ferential direction to prevent rigid body rotations, allow ing the nodes m ovem ent relative to the constraint node. Five elements w ere assigned through the thickness o f the vessel, thirty elem ents w ere assigned in the circum ferential length and ninety-six elem ents were assigned in the longitudinal length. There w ere 14,400 elem ents in total for the stenotic vessel and 532 elements w ere assigned to the stent. The elem ent type

used w as a full integration, three-dim ensional eight-node isoparam etric arbitrary hexahedral elem ent (Type 7 in M arc M entat). A non-linear, large displacem ent analysis w as perform ed using M sc M arc M entat.

A.3 Results and Discussion

The radial displacem ent throughout the structure w as found to be highly non uniform , see Figure A.2. The stent w as expanded to achieve 3 mm diam eter. The m ovem ent o f the artery relative to stent expansion in the m iddle plane were exam ined. The m easurem ents w ere taken at full loading. It w as found that the circum ferential m ovem ent in the m iddle plane ranged from -0.02 mm to 0.03 mm, see Figure A.3 and the longitudinal m ovem ent ranged from 0.02 mm to 0.06 mm, see Figure A.4.

154

mm

Figure A.2

Radial displacem ent distribution throughout m linkthick scaffolding a stenotic vessel.

mm 0.03

0.02
0.01

Figure A .3

Circum ferential distribution throughout the m iddle plane o f the vessel.

155

mm 0.06
0.05

Figure A .4

Longitudinal distribution throughout the m iddle plane o f the vessel.

A.4 Conclusions
The result shows that the displacem ent in the m iddle plane o f the artery relative to the stent expansion was found m inim al. The com putational tim e for this m odel took 20 days to obtain 100% loading w ith a system o f the follow ing specification; Pentium 4 CPU, 1 GB o f RA M and 2.6 GHz o f speed. Therefore the full m odel was divided into tw o halves o f proxim al h a lf and distal h a lf

156