Medical Engineering & Physics 35 (2013) 350356

Contents lists available at SciVerse ScienceDirect

Medical Engineering & Physics
j our nal homepage: www. el sevi er . com/ l ocat e/ medengphy
A sensor for needle puncture force measurement during interventional
radiological procedures
J. Zhai
a,
, K. Karuppasamy
b
, R. Zvavanjanja
b
, M. Fisher
c
, A.C. Fisher
d
, D. Gould
b
, T. How
a
a
Institute of Ageing and Chronic Disease, University of Liverpool, Duncan Building, Liverpool L69 3GA, UK
b
Department of Radiology, Royal Liverpool University Hospital, Prescot Street, Liverpool L7 8XP, UK
c
Department of Cardiology, Royal Liverpool University Hospital, Prescot Street, Liverpool L7 8XP, UK
d
Department of Medical Physics and Clinical Engineering, Royal Liverpool University Hospital, Duncan Building, Liverpool L69 3GA, UK
a r t i c l e i n f o
Article history:
Received 10 January 2012
Received in revised form25 May 2012
Accepted 29 May 2012
Keywords:
Needle insertion
Arterial puncture
Interventional radiology
Liver biopsy
Virtual reality training
a b s t r a c t
Computer-based simulation for interventional radiology training has attracted increasing attention in
recent years because of its potential to train remotely from patients and to provide objective assessment
of prociency. Yet developing a high delity simulator with realistic tactile feedback requires accurate
knowledge of forces exerted on medical devices during interventional radiology procedures. This paper
presents the development and validation of a force sensor for the measurement of axial forces generated
during needle, and combined cannula/trocar, puncture procedures in patients. In order to assess the
performance of this sensor, in vitro measurements were obtained using needle penetration of porcine
liver, kidney and muscle. The results were compared with forces measured by means of a tensile tester.
Calibration results showed that the force sensor has high sensitivity and linearity. Comparison of the
force proles obtained from the sensor and the tensile tester shows that good agreement was achieved
in the in vitro studies for all the tissues tested.
Preliminary clinical force measurements during arterial puncture and liver biopsy procedures have
been performed in patients. An example of force recording for each procedure type is presented.
Crown Copyright 2012 Published by Elsevier Ltd on behalf of IPEM. All rights reserved.
1. Introduction
Interventional radiology (IR) medical techniques involve the use
of needles, guidewires and catheters to diagnose and treat a range
of pathologies in arteries and organs, using imaging technologies
such as ultrasound or X-ray to guide the manipulation of these
instruments. IRrequires a considerable level of prociency to attain
technical success whilst avoiding complications. Traditionally, this
expertise is acquired by practising on animals, physical models or
patients. However, animal-based training is costly and controver-
sial, with limited relevance of animal anatomy and pathology to
humans. An alternative is to use physical models which accurately
represent human anatomy, often produced by rapid-prototyping
techniques. However, these models are expensive and are easily
destroyed by multiple needle punctures. Training on real patients
occurs in the traditional apprenticeship method, yet this prolongs
procedures and can increase risks to patients.
In recent years, there has been considerable interest in using
computer-based simulation for IR training [14]. Virtual IR train-
ing allows deliberate practice of procedural skills, remote from

Corresponding author. Tel.: +44 1865283642.

E-mail address: zhaijianhua308@gmail.com(J. Zhai).
patients, whilst reducing costs of theatre time for training and
avoiding the inevitable risk that this method of learning poses to
patients. Not only can this type of simulation offer exibility for
trainees with different skill levels to practise in realistic imaging
data sets, but also it brings the added opportunity of precise, objec-
tive performance assessments.
Needle access to tissue and organ systems is the essential rst
step in many IR procedures, such as gaining vascular or visceral
catheter access (Seldinger technique), or tissue biopsy for histo-
logical diagnosis. Prior to needle puncture, an initial tiny incision is
made, throughwhichaneedleis insertedintosubcutaneous tissues.
The location of this skin puncture and guidance of the subsequent
needle trajectory to its target, are determined by the operator using
either palpation(of bone landmarks andpulse) or ultrasoundimag-
ing. Observationof the ultrasoundimage conveys needle locationin
real time, allowing the needle tip to be carefully directed towards a
suspect soft tissuemass or tumour, abscess, or towards thelumenof
a bile duct, renal collecting systemor blood vessel. In the last case,
where an artery is successfully punctured, pulsatile blood issues
from the needle hub, conrming correct needle placement. In the
case of an abscess, successful puncture is indicated by the egress
of pus from the needle. A guidewire can then be introduced into
the punctured vessel or viscus via the bore of the needle, feeling
carefully for tactile evidence of impending complications. Once the
1350-4533/$ see front matter. Crown Copyright 2012 Published by Elsevier Ltd on behalf of IPEM. All rights reserved.
http://dx.doi.org/10.1016/j.medengphy.2012.05.012
J. Zhai et al. / Medical Engineering & Physics 35 (2013) 350356 351
Fig. 1. Thin aluminiumbeamwith integral frame (a) top viewand (b) cross-sectional viewwith strain gauges. All dimensions are in mm.
wire is correctly located in the targeted structure (kidney, bile duct,
vessel, etc.), the needle is removedleaving the wire inplace: a range
of diagnostic andtherapeutic catheters canthenbe introducedover
the guidewire.
Inthe real worldtask, tactile cues are important to safe procedu-
ral outcomes. Hence, in an IRtraining simulator, the delity of force
feedback during each step of the Seldinger technique, or during the
insertion of a cannula/trocar assembly prior to biopsy procedures,
should correctly reect an operators feel during an actual clinical
procedure. To identify this task delity, and the forces generated,
requires measurements during procedures in patients, a key ele-
ment of which must be safety, with minimal interference with the
clinical procedure being performed.
The investigation of interventional needle puncture forces has
been investigated previously by custom-made or commercially
available force measuring products [512]. Washio and Chinzei [5]
described a coaxial force sensor which can measure both the cut
force and the friction force during a needle insertion of different
tissues. Podder et al. [6] used a six-axis force sensor (Nano17

, ATI
Industrial Automation, America) toperformtheforcemeasurement
during a prostate brachytherapy. A MRI compatible 3DOF optical
force/torque sensor was also built for the same purpose [10]. It has
a force measurement range from0 to 20N with resolution of 0.1N.
Peirs et al. presented the design of a 3-axial force sensor which
measures of 5mm in diameter to provide force feedback during
minimally invasive robotic surgery [13]. It uses a exible titanium
structure as the sensing element and three optical bres to detect
the deformation when subjected forces from different directions.
To our knowledge, none of the previous sensors allows the free
passage of the jet of blood, an essential step in the Seldinger tech-
nique, to conrm that the needle tip is within the vessel lumen.
For this reason, we previously used a capacitive force sensor
(Contacts

, Pressure Prole Systems, Inc., Los Angeles, America)

which was mounted on the operators thumb using a nger glove
for in vivo force measurement during an arterial puncture proce-
dure [12]. However, the commercial or home-made force sensors
used previously are bulky: some are designed for use with robotic
systems and are unsuitable for use during clinical procedures. The
Contacts

sensor used in our previous study was found sensitive

to position changes between the sensor and the needle hub. In
addition, the placing of the sensor between the operators thumb
and the needle hub signicantly reduces the tactile feeling of the
operator which is essential to performthe Seldinger technique.
This paper describes the development of a coaxial force sen-
sor for unobtrusive measurement of the insertion force during
arterial puncture and liver biopsy. Studies were carried out to mea-
sure forces on needle puncture of porcine liver, kidney and muscle
using this force sensor, and the results were compared with those
recorded simultaneously using a tensile tester. Clinical measure-
ment of forces during arterial puncture and biopsy procedures are
currently being performed. Preliminary data obtained from one
case of arterial puncture during cardiac catheterisation and one
liver biopsy procedure are presented.
2. Methods
2.1. Sensor design
The vascular access needle (Cook Europe, Cork, Ireland) used in
our institutionconsists of a 19Gsteel needle component to whichis
attached to a moulded plastic Luer assembly (holder) which incor-
porates wings, designed to facilitate holding by an operator during
puncture procedures. An important requirement for the sensor is
that it should be small and light so that any interference with the
procedure and the normal method of use of the needle can be min-
imised. The sensor design is based on the use of a thin beam with
an integral frame machined fromone piece of aluminiumbar. The
beamis 15mm in length, 1.0mmthickness and 6.0mmwidth and
has a circular hole of diameter 2.8mm in the centre (Fig. 1a). An
aluminiumtube that ts into this hole is securely bonded onto the
beam. Force applied to the middle of the beam induces strains in
Fig. 2. Needle force sensor structure (a) sensor assembly; (b) exploded viewof the sensor and (c) sensor inserted in an arterial puncture needle.
352 J. Zhai et al. / Medical Engineering & Physics 35 (2013) 350356
Fig. 3. Simulation result of strain distribution of the beam subjected to a load of
10 N.
the beamwhichcanbe detected by means of straingauges (Fig. 1b).
Fig. 2a shows the assembled force sensor which has overall dimen-
sions of 22mm12mm, thickness 6mm. Excluding the leads and
plug, the sensor has a weight of 4.2g. The exploded view (Fig. 2b)
shows the beam, lower and upper covers and a male Luer connec-
tor which is xed to the bottom cover. The bottom section of the
Luer connector mates with the hub (holder) of an arterial puncture
needle and the top section passes through the beamand the hollow
tube but without making direct contact with either. The top cover
has a clearance hole for the central tube which protrudes through
the cover by 4mm. To incorporate familiar tactile components into
the force measurement system, the plastic hub or holder is rst
carefully removed froman arterial puncture needle. This dummy
holder is then secured within the sensors central tube and these
assembled components are then sterilised by ethylene oxide. Dur-
ing an in vivo procedural force data collection, the preassembled
dummy holder and sensor are removed from their sterile pack-
aging and inserted into the Luer lock channel of the procedural,
vascular access needles hub assembly (Fig. 2c). To performneedle
puncture, the operator grasps the dummy holder, fromwhich all
axial forces generated are transmitted to the beam, and advances
the needle assembly, using it as in a normal procedure. The ensuing
strain distribution on the beamwhen subjected to a load of 10N is
simulated using nite element analysis software Pro/ENGINEER

Mechanica

and the result is shown in Fig. 3.

Four backed U-shaped semiconductor strain gauges (part num-
ber SS-060-033-500PUB with its backing base dimension of
3.3mm7.6mm) fromMicron Instruments (Simi Valley, CA, USA)
are bonded onto the beam and wired to form a full Wheatstone
bridge circuit. Thin layers of water-resistant compounds (Gagekote
#5 and Gagekote #8, Vishay Measurements Group UK Ltd., Bas-
ingstoke, UK) were applied to the strain gauges and the wiring
inside the sensor to protect themfromblood contamination during
in vivo use and from cleaning after the measurement. A custom-
built signal conditioning circuit comprising a stabilised DC voltage
source of 5V for bridge supply and instrumentation amplier
(INA128P, Texas Instruments, USA) was developed for use with this
sensor.
2.2. Sensor calibration
The sensor was calibrated using dead weights applied to the
dummy holder whilst keeping the body of the sensor in a xed
position in a small vice. To cover the full range of forces that may
be encountered in vivo, increasing weights from 50g up to 1050g
(0.5N to 10N) were applied whilst the resulting output voltages
were recorded by means of a digital voltmeter. As the sensor can
also measure the forces during needle retraction, the calibration
was also performed with the sensor turned upside down and the
dead weights were hung fromthe holder which is nowsituated at
the bottom. This calibration procedure was carried out ten times at
roomtemperature.
2.3. Sensor validation
In vitro validation studies analysed the sensors dynamic per-
formance. Force data, obtained using the fabricated sensor during
needle puncture of porcine liver, kidney and subscapularis muscle,
were compared with data recorded simultaneously using a ten-
sile tester (Nene Instruments Ltd., Wellingborough, UK) tted with
a 5N load cell. The load cell was attached to the sensor holder
and an 18G 2-parts Kimal needle (Kimal Plc, Middlesex, UK) was
connected to the male Luer connector at the bottom of the sen-
sor. Fresh, porcine liver and kidney samples obtained within 12h
of slaughter were placed directly below the tip of the needle as
shown in Fig. 4. Downward movement of the crosshead caused
the needle to penetrate through the tissues at a constant speed
whilst theloadsignals fromtheloadcell andtheneedlesensor were
recorded continuously via an analogue-to-digital converter (USB-
1608FS, Measurement Computing Corp., USA) at a sampling rate of
100samples/s and saved to a PC. The needle insertion speed was
set to 200mm/min. Needle penetration of the porcine subscapu-
laris muscle was at a right angle to muscle axes. After reaching a
depth of between 16 and 20mm, the needle tip was retracted at
the same speed until it reached the start point when the measure-
ment was terminated. Ten punctures were made for each sample
of porcine liver, kidney and muscle.
2.4. Clinical measurements
The needle force sensor is currently being used in vivo to deter-
mine the axial forces on needles during routine liver biopsy and
arterial puncture using the Seldinger technique. Informed consent
is obtained and the study has been approved by the Research Ethics
Committee. The sensor is sterilised using ethylene oxide. An exam-
ple of the forces obtained for each procedure is presented here.
Fig. 4. Experimental setup for in vitro sensor validation.
J. Zhai et al. / Medical Engineering & Physics 35 (2013) 350356 353
Force (N)
10 8 6 4 2 0 -2
O
u
t
p
u
t

(
V
)
-1
0
1
2
3
4
5
6
Mean of 10 measurements
Fig. 5. Sensor calibration results. The error bars represent the standard deviation.
The tted curved is y =0.55x 0.02, r
2
>0.999.
Table 1
Comparison of the forces measured by means of the tensile tester and the needle
force sensor using BlandAltman analysis.
Tissue Bias 95% limits
Liver 0.0065 0.0149 to 0.0278
Kidney 0.0085 0.0157 to 0.0326
Muscle 0.0092 0.0153 to 0.0337
The bias and the limits of agreement for each tissue are given.
For liver biopsy, a Kimal needle with Trocar 18G15cmwas used
whilst arterial puncture was achieved using a 19G vascular access
needle (Cook Europe, Cork, Ireland).
3. Results
3.1. Sensor calibration
Fig. 5 shows a plot of the mean output voltages for the range
of loads applied. The least-squares regression curve for these data
is y =0.55x 0.02 with regression coefcient r
2
>0.999, where y is
the voltage output and x the applied force in newtons. As the gain
of the amplier was 50, the sensitivity of the sensor was therefore
11mV/N.
3.2. In vitro measurements
Fig. 6shows the plot of the force proles measuredby the needle
sensor and the load cell during a typical experiment as the needle
penetrates through the porcine liver, kidney and muscle. To com-
pare the force measurements obtainedby the twomethods for each
tissue, the BlandAltman plots were made as shown in Fig. 7. The
bias and 95% limits of agreement are given in Table 1. In addition,
a linear regression analysis on the force data was also performed
and the regression and Pearsons correlation coefcients are given
in Table 2 for each tissue type.
Table 2
Linear regression analysis of the force data obtained fromthe tensile tester and the
needle force sensor.
Tissue Slope Y-intercept Pearsons correlation
coefcient (r
2
)
Liver 1.04 4.610
3
0.991
Kidney 1.01 7.110
3
0.998
Muscle 1.01 6.310
3
0.999
Time (s)
14 12 10 8 6 4 2 0
F
o
r
c
e

(
N
)
-0.1
0.0
0.1
0.2
0.3
0.4
0.5
Load cell
Needle sensor
a
Time (s)
12 10 8 6 4 2 0
F
o
r
c
e

(
N
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Load cell
Needle sensor
b
Time (s)
12 10 8 6 4 2 0
F
o
r
c
e

(
N
)
0.0
0.5
1.0
1.5
2.0
Load cell
Needle sensor
c
Fig. 6. Plot of the forces recorded by the load cell (Nene tensile tester) and needle
sensor during a single measurement for (a) porcine liver, (b) porcine kidney and (c)
porcine muscle.
3.3. In vivo measurements
The plot in Fig. 8 shows the recording of the forces measured
during a liver biopsy procedure in a patient. The force prole is
much more complex here because the needle has to pass through
different tissues before reaching the liver and the depth of pene-
tration may be up to 15cm. In this case, the needle is not inserted
continuoulsly as in the in vitro measurements, but in stages whilst
the position of the needle tip is followed by means of ultrasound
imaging to ensure safe and accurate needle penetration. The fea-
tures AE indicated in Fig. 8 can be attributed to the forces required
for the biopsy needle to penetrate different tissues. The rst peak A
corresponds to skin penetration, B and C to soft tissues and Dand E
to actual puncture of the liver. It can be seen that the forces return
354 J. Zhai et al. / Medical Engineering & Physics 35 (2013) 350356
Fig. 7. BlandAltman plots for the comparison of the results fromthe tensile tester
and the needle force sensor (a) porcine liver, (b) porcine kidney and (c) porcine
muscle.
towards the zero baseline at various points during the recording
when the force on the needle is released whilst the position of the
needle tip is veried by ultrasound imaging.
Fig. 9 shows an example of the forces recorded during femoral
artery puncture using the Seldinger technique. The rst peak,
labelled (a) on the graph, corresponds to the initial needle pene-
tration through a 25mm skin incision. Following a momentary
Time (s)
40 30 20 10 0
F
o
r
c
e

(
N
)
-2
0
2
4
6
8
10
B A
C
D
E
Fig. 8. Forces recorded during a liver biopsy procedure in a patient. The features
AE in the recording are described in the text.
pause (b) when the force drops towards zero, the needle is again
advanced (c) and then retracted (d). Two further attempts (e and
f) are made at localising the femoral artery before successful punc-
ture, conrmed by the jet of blood through the sensor and holder,
at (g). The peak force required to penetrate the arterial wall was
2.1N.
4. Discussion
During needle puncture the operator employs tactile and visual
cues to aid safe and reliable placement of the needle tip into the
target organ, tissue or vessel lumen. Measurement of the forces
experienced by the operator using a sensor that impinges on the
important tactile cues may not only affect the reliability and accu-
racy of the measurement but also the safety of the patient. The
sensor described in this paper was designed to be unobtrusive and
light so that it has minimal interference with the ongoing medi-
cal procedure. The holder of the sensor is derived directly froman
actual arterial puncture needle, which has the advantage of famil-
iarity to the operator, allowing the operator to hold the sensor in
their preferred way. On successful needle puncture into an arte-
rial lumen, a pulsatile streamof arterial blood jets through the top
holder, as per normal.
The calibration results demonstrate that the sensor has good
linearity and repeatable response with relatively low coefcient
of variations. When comparing the results obtained from in vitro
tests on porcine tissues by the needle sensor and the tensile tester
load cell, it is seen that the peak penetration force (occurring when
needle tip is about to penetrate the capsule of liver or kidney) for
liver, kidney or muscle matches well. Fig. 6a, b and c shows that the
requiredcapsulepenetrationforces measuredbythesensor andthe
tensiletester are0.42Nand0.41Nfor liver and0.67Nand0.66Nfor
Time (s)
0 20 40 60 80 100
F
o
r
c
e

(
N
)
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
a
b
c
d
e
f
g
Fig. 9. Force recording during an arterial puncture procedure using the Seldinger
technique. The features ag in the recording are described in the text.
J. Zhai et al. / Medical Engineering & Physics 35 (2013) 350356 355
Fig. 10. Force variations during needle insertion of porcine liver (a) two-part Kimal needle tip and (b) force prole.
kidney. Therelativedifference(absolutedifferenceintheforces/the
tensile tester force 100%) for liver is 2.4%, 1.5%for kidney and1.3%
for muscle, which can be considered as insignicant. Furthermore,
this high degree of agreement is veried from the results of the
BlandAltman analysis (Fig. 7). The correlation coefcients for all
tissues are higher than 0.996, again indicating the good agreement
between these two methods.
However, the analysis results also show some differences
between these two measurement methods. For example, in Fig. 6,
the maximum relative difference for the liver reaches 7.58% and
7.49% for muscle at some points. This could be explained by the
different sampling rates of the forces recorded by these two meth-
ods. Duringthemeasurement, thesamplingrateof thetensiletester
couldnot beadjustedindependentlybytheoperator andfor thetest
speed of 200mm/min, this was about 20Hz, whilst the sampling
rate of the data collection system of the sensor was set to 100Hz.
As a result, the tensile tester may not have recorded some critical
force data when a rapid change in puncture force occurred. This is
also veried by the fact that most of the peak forces recorded by
the sensor were slightly higher than those measured by the tensile
tester.
Because of the difference in sampling rates of the two force
recording methods, comparison of the measurements at the same
time points was achieved by cubic spline interpolation. Interpo-
lation applied to the tensile tester and needle force sensor data
allowed forces to be determined at the same time points.
The force proles presented in Fig. 6 showseveral distinct char-
acteristics which reect the tissue structures encountered and the
two-part needle tip conguration. Here we use Fig. 10 to demon-
strate the force variations during each step of the insertion of a
Kimal needle. The needle moves down towards the liver sample
and comes into contact with liver capsule at Point A. Force is built
up rapidly and the contact area at the needle tip and the liver cap-
sule starts to deform. The magnitude of this force is a function of
the stiffness of the liver tissue being deformed. At Point B, the sharp
tip of the needle inner stylet penetrates the liver capsule and the
puncture force drops slightly. Force goes up again (point C) as the
liver capsule encounters the small step produced by this needles
outer trocar just above the needle tip. This force reaches its peak at
Point D when the outer part of the needle is about to penetrate the
capsule, then drops dramatically after the penetration, increasing
again gradually at Point E until the peak at Point F, where the inser-
tion ends. The needle then commences withdrawal after Point F,
until Point G where the capsule is released fromthe needle tip. The
data in Fig. 9 is partly derived from Fig. 6a, however, the features
described can also be found in Fig. 6b and c. Apparently, the cap-
sules of the liver and kidney and the small step formed by the two
parts of the needle have dominated the pattern of force variations.
During needle insertion into tissue the force measured by the
sensor is a sumof three components: (1) tissue stiffness force, (2)
cutting force and (3) friction force [13]. The rst two components
are a function of the properties of the tissue and the characteris-
tics of the needle tip, i.e. size, sharpness and bevel angle. The third
component is the force along the length of the needle in contact
with soft tissue and comprises the static and dynamic Coulomb
friction, viscous damping and the tissue clamping around the shaft
of the needle. The friction force is therefore dependent on the
depth of penetration of the needle. The forces measured during
the liver biopsy procedure were much higher than those measured
in vitro on isolated tissues. This is to be accounted for partly by
the signicantly greater depth of penetration of the needle in vivo
(150mm compared with 20mm). It was also noted by the oper-
ator (RZ) that this particular patient had a brosed and unusually
stiff liver.
The needle force sensor was designed to record only the force
along the longitudinal axis of the needle. In calibrating the sen-
sor and during the in vitro evaluation using the tensile tester, only
axial forces were applied. However, during the in vivo measure-
ments transverse forces may be applied by the operator during the
procedure depending on how the needle is manipulated. We have
assessed the sensitivity of the sensor to transverse forces and found
that when forces were applied in the direction of the long axis of
the beam, the sensitivity was about 10% of that in the longitudi-
nal axis of the needle. However, because of the arrangement of
the strain gauges, the sensor was effectively insensitive to trans-
verse forces applied at right angle to the long axis of the beam.
Transverse forces were minimised during the in vivo measure-
ments by ensuring that needle trajectory during arterial puncture
and liver biopsy was parallel with the longitudinal axis of the
needle.
Training interventional tasks requires learning tactile, as well
as visual, perceptual cues. An operators application of force to
an instrument (needle) may include axial, lateral, perhaps rota-
tory components, to attain a specic target (e.g. for biopsy). To
commence procedures in patients without inherent knowledge of
the forces required can expose a patient to risk; considerable or
excessive force could drive a needle into an undesirable anatomical
location (aorta, gallbladder) producing a complication. Similarly,
the level of force that is sufcient to initiate needle advancement
must be learnt; hence for vascular access this must be sufcient to
traverse the anterior vessel wall, but not so much as to go through
both vessel walls. The sensor was designed specically for the pur-
pose of in vivo needle puncture force measurement. Similar sensor
may be incorporated into instruments for training and in auto-
mated robotic needle insertion devices. However, the design may
need to be altered to suit particular application.
356 J. Zhai et al. / Medical Engineering & Physics 35 (2013) 350356
5. Conclusion
The development and validation of a novel, miniature force
sensor is presented. This sensor can be used to measure needle
inserting forces unobtrusively during needle access to liver, kidney
or muscle, for example for tissue biopsy or uid drainage proce-
dures. It canalsobeusedinother medical procedures whichinvolve
the use of needles for diagnosis or treatment purposes such as for
breast biopsy, laparoscopy access and arterial puncture procedures
using the Seldinger technique. In vitro results fromthe sensor show
good agreement with those from the tensile tester. Future work
will explore the effect of needle insertion speed on the puncture
force, and the detailed forces generated during in vivo procedures
onpatients. Theinuenceof thesedatafollowingincorporationinto
simulation models for IR training can be explored through content
validation studies.
Acknowledgements
This work was funded by the Engineering and Physical Sciences
Research Council and the Health Technology Devices Programme;
it presents independent research commissioned by the National
Institute for Health Research (NIHR). The views expressed in this
publication are those of the authors and not necessarily those of
the NHS, the NIHR or the Department of Health.
Conict of interest statement
The authors declare that there are no conicts of interest with
any of the programcommittee members.
References
[1] Gould DA. Interventional radiology simulation: prepare for a virtual revolution
in training. J Vasc Interv Radiol 2007;18(4):48390.
[2] Gould DA, Reekers JA. The role of simulation in training endovascular interven-
tions. Eur J Vasc Endovasc Surg 2008;35:6336.
[3] Dawson DL, Meyer J, Lee ES, Pevec WC. Training with simulation
improves residents endovascular procedure skills. J Vasc Surg 2007;45(1):
14954.
[4] Anderson J, Chui CK, Cai Y, Wang Y, Li Z, Venbrux A, et al. Virtual reality training
ininterventional radiology: theJohns Hopkins andKent Ridgedigital laboratory
experience. Semin Interv Radiol 2002;19(2):17985.
[5] Washio T, Chinzei K. Needle force sensor, robust and sensitive detection of the
instant needle force. In: Int Conf Med Image Comput & Comput-Assis Interv,
vol. 3217. 2004. p. 11320.
[6] Podder T, Clark D, Sherman J, Fuller D, Messing E, Yu Y, et al. In vivo motion and
force measurement of surgical needle interventionduring prostate brachyther-
apy. Med Phys 2006;33(8):291522.
[7] Saito H, Togawa T. Detection of needle puncture to blood vessel using puncture
force measurement. Med Biol Eng Comput 2005;43(2):2404.
[8] Piccin O, Barb L, Bayle B, Mathelin MD, Gangi A. A force feedback teleop-
erated needle insertion device for percutaneous procedures. Int J Robot Res
2009;28(9):115468.
[9] Abolhassani N, Patel R, MoallemM. Needle insertion into soft tissue: a survey.
Med Eng Phys 2007;29:41331.
[10] Su H, Fischer GS. A 3-axis optical force/torque sensor for prostate needle
placement in magnetic resonance imaging environments. In: 2nd annual
IEEE international conference on technologies for practical robot applications.
Boston, MA, USA: IEEE; 2009. p. 59.
[11] Peirs J, Clijnen J, Reynaerts D, Brussel HV, Herijgers P, Boone
S, et al. A micro optical force sensor for force feedback during
minimally invasive robotic surgery. Sens Actuators A 2004;115:
44755.
[12] Healey AE, Evans JC, Murphy MG, Powell S, How TV, Gould DA, et al. In vivo
force duringarterial interventional radiologyneedle puncture procedures. Stud
Health Technol Inform2005;111:17884.
[13] Okamura AM, Simone C, OLeary MD. Force modelling for needle insertion into
soft tissue. IEEE Trans Biomed Eng 2004;10(51):170716.