Vous êtes sur la page 1sur 8

Original article

Prognostic signicance of the circumferential resection margin following total mesorectal excision for rectal cancer
A. Wibe, P. R. Rendedal*, E. Svensson*, J. Norstein*, T. J. Eide, H. E. Myrvold and O. Sreide on behalf of the Norwegian Rectal Cancer Group
Department of Surgery, University Hospital Trondheim, *Cancer Registry of Norway, and Departments of Pathology and Surgery, The National Hospital, Oslo, Norway Correspondence to: Dr A. Wibe, Department of Surgery, University Hospital Trondheim, N-7006 Trondheim, Norway (e-mail: arne.wibe@medisin.ntnu.no)

Background: Knowledge of prognostic factors following resection of rectal cancer may be used in the

selection of patients for adjuvant therapy. This study examined the prognostic impact of the circumferential resection margin on local recurrence, distant metastasis and survival rates. Methods: A national population-based rectal cancer registry included all 3319 new patients from November 1993 to August 1997. Some 686 patients underwent total mesorectal excision with a known circumferential margin. This shortest radial resection margin was measured in xed specimens. None of the patients had adjuvant radiotherapy. Results: Following potentially curative resection and after a median follow-up of 29 (range 14 60) months, the overall local recurrence rate was 7 per cent (46 of 686 patients): 22 per cent among patients with a positive resection margin and 5 per cent in those with a negative margin (margin greater than 1 mm). Forty per cent of patients with a positive margin developed distant metastasis, compared with 12 per cent of those with a negative margin. With decreasing circumferential margin there was an exponential increase in the rates of local recurrence, metastasis and death. Conclusion: The circumferential margin has a signicant and major prognostic impact on the rates of local recurrence, distant metastasis and survival. Information on circumferential margin is important in the selection of patients for postoperative adjuvant therapy.
Paper accepted 20 November 2001 British Journal of Surgery 2002, 89, 327334

Introduction

Histopathological reporting of rectal cancer traditionally includes information on resection margins, usually the bowel resection margins. Tumour involvement of the proximal bowel margin is unusual, and distal intramural tumour extension more than 2 cm from the primary lesion is uncommon. The circumferential margin of resection has less frequently been reported but, as pointed out by Quirke et al.1 and others25, circumferential margin involvement appears to be a strong predictor of local recurrence and hence of survival. Others have argued that circumferential margin involvement is more an indicator of advanced disease than of incomplete excision6. However, from a clinical point of view, these papers16 consider
The Editors have satised themselves that all authors have contributed signicantly to this publication

heterogeneous patient populations that are difcult to interpret. The Norwegian Rectal Cancer Project was initiated in 1993. The objectives were to enhance the quality of surgery in Norway by the introduction of total mesorectal excision (TME) as the standard resection technique, to reduce local recurrence rates by operation alone, to improve survival after curative surgery, and to establish a quality control instrument with feedback to participating hospitals. The present study examined the role of the circumferential margin and its prognostic importance for local recurrence, distant metastasis and survival in a national cohort of patients undergoing major rectal resection for cancer. Data from patients who undergo a standardized operation, and in whom the mesorectum is excised, may better reect the potential importance of the circumferential margin and may elucidate whether circumferential margin involvement reects advanced disease rather than
British Journal of Surgery 2002, 89, 327334
327

2002 Blackwell Science Ltd

328

Circumferential margin in rectal cancer A. Wibe, P. R. Rendedal, E. Svensson et al.

Major resection, curative intent ( n = 2121) Not TME (n = 450)

Total mesorectal excision (n = 1671) CRM not given (n = 861)

CRM given (n = 810) Bowel or tumour perforation (n = 48) Radiotherapy (n = 76)

Study group (n = 686)

Fig. 1 Flow chart of the selection process for the study group.

CRM, circumferential margin; TME, total mesorectal excision

incomplete excision. Finally, the information obtained may guide the selection of patients for adjuvant radiotherapy and/or chemotherapy.
Patients and methods

In a national audit, 3319 patients with rectal adenocarcinoma were diagnosed and treated in 55 hospitals (seven university, 14 district general and 34 local hospitals) in Norway from November 1993 to August 1997. TME was used in 79 per cent and a non-TME resection in 14 per cent of the operations, and the operative technique was not stated in 7 per cent. A major rectal resection (i.e. anterior or abdominoperineal resection or a Hartmann procedure) with a curative intent was performed in 2121 patients with a tumour at or below 16 cm from the anal verge. Of these, 686 patients underwent TME, and data on the circumferential margin were available from the pathology report (Fig. 1). None of these patients had adjuvant radiotherapy. Adjuvant chemotherapy was not part of the national treatment policy for rectal cancer (formulated by the Norwegian Gastrointestinal Cancer Group)7. However, 3 per cent of the patients had chemotherapy as they were included in a separate randomized clinical trial in the period from January 1993 to October 1996. A disease-specic database was established in the Cancer Registry of Norway, a population-based registry with compulsory reporting of cancer. Clinical and operative
British Journal of Surgery 2002, 89, 327334 www.bjs.co.uk

data were submitted on project-specic forms, and copies of pathology reports gave additional information on staging. Follow-up was complete and facilitated by a unique national 11-digit personal identication number system. Survival data were delivered by the Norwegian Cause of Death Registry. Before commencing the Norwegian Rectal Cancer Project, colorectal surgeons were taught the main principles of TME. To standardize rectal cancer surgery on a national level, the surgeons attended several workshops, including live video demonstrations that pinpointed important steps of the procedure, and they had the opportunity to assist a surgeon experienced in the technique. The TME procedure was dened as sharp dissection under clear vision, with excision of the rectum and mesorectum contained within the mesorectal fascia. This does not necessarily mean an excision of the entire mesorectum down to the pelvic oor or anal canal, but the standard policy for division of the rectal tube and the mesorectum was 5 cm below the tumour. During the rst years of the national audit, pathologists were introduced to standardized handling and reporting of the operative specimen and to the use of tumour node metastasis (TNM) criteria. The pathology laboratories were encouraged to follow the quality control principles formulated by Quirke et al.1,8 in their assessment of the specimens, emphasizing the role of the circumferential margin and the mesorectal lymph nodes. The fresh specimens were opened along the antimesenteric border, avoiding the tumour if possible, and xed in formalin. Specimens were then sectioned in the transverse plane in order to identify the shortest distance (in millimetres) from the outermost part of the tumour to the lateral resection margin (Fig. 2). If macroscopic tumour deposits were noted in the mesorectum outside the main tumour, the measurement was made to the outer border of the deposit. In this study a circumferential margin greater than 1 mm was dened as an uninvolved margin1. Local recurrence was dened as recurrent disease in the pelvis, including recurrence at the site of the bowel anastomosis and of the perineal wound9. Metastases denote recurrent disease outside the pelvis, including metastases in para-aortic and/or inguinal lymph nodes. Local recurrence rates are given as the sum of local recurrence occurring in isolation and concomitant with distant metastasis; the rate of metastasis incorporates distant metastases with or without local recurrence. Recurrent disease was diagnosed clinically or by imaging. Histopathological or cytological verication was encouraged but not required. The follow-up regimen adhered to the principles formulated by the Norwegian Gastrointestinal Cancer Group7. This regimen included clinical examination,
2002 Blackwell Science Ltd

A. Wibe, P. R. Rendedal, E. Svensson et al. Circumferential margin in rectal cancer 329

Fig. 2 Illustration of the circumferential resection margin

included in the two other models. In these analyses the best prognostic group acted as reference, giving a hazard ratio greater than 1 for those at greater risk of local recurrence, metastasis or death. The Cox analyses did not incorporate T1 tumours, as no patient with such a tumour had a local recurrence or died, and only one patient developed metastases (if T1 tumours were included in the analyses, the HRs for T24 tumours would not have provided any useful information). The Cox analyses were performed both with missing data as separate categories and without missing data, although similar results were produced. No interactions were identied. The role of the circumferential margin as a predictor of local recurrence was examined further in a univariate regression model, illustrated by logarithmic curve estimation weighted by cases (i.e. circumferential margin groups with more patients counted more than those with fewer patients). How well the model ts the data is given by r2. Sensitivity, specicity and predictive values are all time dependent, and thus they have to be judged in comparison with the median follow-up. Statistical analyses were performed using the Statistical Package for the Social Sciences program (SPSS, Chicago, Illinois, USA).
Results

(CRM). Redrawn with permission from the American College of Surgeons from J Am Coll Surg 1997; 184: 8492. Original illustration by Kari C. Toverud

proctoscopy, ultrasonography of the liver, chest radiography and determination of serum carcinoembryonic antigen level. The individual follow-up programme was left to the discretion of the participating hospital. Commonly, follow-up consultations were undertaken every 3 months for 2 years and then every 6 months until 5 years after operation. Median follow-up was 29 (range 1460) months. Statistical analysis Patient and treatment characteristics were described using frequency tables. Missing data were included as `not stated'. Cumulative hazard rates (HRs) of local recurrence, metastasis and crude survival were estimated by means of univariate KaplanMeier plots. The importance and statistical signicance of potential prognostic factors were analysed with Cox proportional hazard regression models, and signicant variables were included in the nal regression analyses. If one variable was signicant in one of the three Cox analyses, that variable was
2002 Blackwell Science Ltd

Patient, treatment and tumour characteristics are given in Table 1. The frequencies of local recurrence, metastasis and death according to circumferential margin are summarized in Table 2. The overall local recurrence rate was 7 per cent. Of the 65 patients with circumferential margin involvement (margin 01 mm), the local recurrence rate was 22 per cent (14 of 65) compared with 5 per cent (32 of 621) in the group without margin involvement (margin greater than 1 mm) (P < 0001). There were no signicant differences in patient characteristics between those with information on the circumferential margin and those without (n = 861; Fig. 1). The rate of local recurrence was slightly lower (6 per cent) in the group of patients without information on the resection margin than in the study group. This difference was not statistically signicant. Cumulative local recurrence rates related to length of follow-up and circumferential margin categories are given in Fig. 3. The impact of the circumferential margin is further outlined in Fig. 4, which gives estimated local recurrence rates according to circumferential margin (r2 = 094). The cumulative HRs of distant metastasis and overall survival related to circumferential margin are given in Figs 5
www.bjs.co.uk British Journal of Surgery 2002, 89, 327334

330

Circumferential margin in rectal cancer A. Wibe, P. R. Rendedal, E. Svensson et al.

Table 1

Characteristics of 686 patients undergoing total mesorectal excision for rectal cancer
No. of patients Median (range) age (years) Age group (years) < 50 5059 6069 7079 > 80 Sex F M Tumour level (cm) 1216 611 <6 Missing Circumferential margin (mm) 01 25 610 1119 > 20 Tumour status (T) 1 2 3 4 Node status (N) 0 1 2 Dukes' stage A B C Type of resection Anterior resection Abdominoperineal resection Hartmann procedure Differentiation Low Moderate High Missing 69 (4890) 36 103 177 254 116 (5) (15) (26) (37) (17)

Table 2

Frequency of local recurrence, metastasis and mortality according to circumferential margin


No. of patients Local recurrence Metastasis 26 24 29 7 14 (40) (14) (17) (6) (9) Mortality 24 32 27 12 10 (37) (19) (16) (9) (6)

Circumferential margin (mm) 01 65 14 (22) 25 170 13 (8) 610 168 12 (7) 1119 127 5 (4) > 20 156 2 (1) Total 686 46 (7)

100 (15)

105 (15)

294 (43) 392 (57) 179 310 162 35 65 170 168 127 156 32 166 452 36 (26) (45) (24) (5) (9) (25) (24) (19) (23) (5) (24) (66) (5)

Values in parentheses are percentages

100 01 mm 25 mm 610 mm > 10 mm

80 Local recurrence (%)

60

40

20

426 (62) 177 (26) 83 (12) 165 (24) 261 (38) 260 (38) 488 (71) 176 (26) 22 (3) 70 544 43 29 (10) (79) (6) (4)

0 No. at risk 01 mm 25 mm 610 mm > 10 mm

10

20

30

40

50

Time after operation (months) 65 170 168 283 51 152 157 271 41 107 110 204 29 64 73 136 14 34 35 70 7 7 20 23

Fig. 3 Local recurrence rate (cumulative hazard) according to

circumferential resection margin (KaplanMeier analysis)

30

Values in parentheses are percentages

Local recurrence rate (%)

20

and 6. An increasing margin was accompanied by a decrease in the proportion of patients who later developed distant metastases and by improved survival. The prognostic impact of the circumferential margin was examined by multivariate analyses adjusted for known predictors of outcome (Table 3). Both circumferential margin and tumour status were found to be signicant prognostic factors of local recurrence. The risk of developing local recurrence increased by a factor (HR) of 12 (95 per
British Journal of Surgery 2002, 89, 327334 www.bjs.co.uk

10

10

20

30

Circumferential resection margin (mm)

Fig. 4 Relationship between circumferential resection margin and local recurrence rate

2002 Blackwell Science Ltd

A. Wibe, P. R. Rendedal, E. Svensson et al. Circumferential margin in rectal cancer 331

100 01 mm 25 mm 610 mm > 10 mm

100

80 Rate of metastasis (%)

80 Cumulative survival (%)

60

60

40

40 01 mm 25 mm 610 mm > 10 mm 10 20 30 40 50

20

20

0 No. at risk 01 mm 25 mm 610 mm > 10 mm

10

20

30

40

50 No. at risk 01 mm 25 mm 610 mm > 10 mm

Time after operation (months) 65 170 168 283 46 148 156 265 32 97 108 196 22 58 70 133 11 34 32 68 7 9 17 23 65 170 168 283 57 155 158 272

Time after operation (months) 42 112 115 205 32 67 75 136 16 37 36 72 8 9 20 24

Fig. 5 Rate of metastasis (cumulative hazard) according to

circumferential resection margin (KaplanMeier analysis)

Fig. 6 Crude survival rate according to circumferential resection

margin (KaplanMeier analysis)

cent condence interval (c.i.) 26 to 558) when the circumferential margin was reduced from 20 to 01 mm. The sensitivity and specicity of circumferential margin involvement in predicting local recurrence were 30 and 92 per cent respectively, and the positive and negative predictive values were 22 and 95 per cent (Table 4). Circumferential margin involvement increased the risk of developing distant metastasis signicantly by a factor (HR) of 47 (95 per cent c.i. 23 to 95) (Table 3). Also the T and N statuses were signicant predictors of distant metastases: N2 status increased the risk of metastasis by a factor (HR) of 28 (95 per cent c.i. 16 to 50). (The low HR for the T4 category of tumour may be explained by few T4 tumours.) As a predictor of metastasis, circumferential margin involvement had a sensitivity and a specicity of 26 and 93 per cent respectively, and positive and negative predictive values of 40 and 88 per cent (Table 4), which means that 40 per cent of the patients with a positive margin developed distant metastases compared with 12 per cent of those with a negative margin. Decreasing circumferential margin increased mortality by a factor (HR) of 37 (95 per cent c.i. 17 to 84) (Table 3). Other independent prognostic factors were tumour status, age of the patient, resection type (increased risk associated with the Hartmann procedure as a result of the selection of high-risk patients to this operation) and tumour differentiation. In prediction of mortality the circumferential margin had a sensitivity and a specicity of 23 and 93 per cent, and
2002 Blackwell Science Ltd

the positive and negative predictive values were 37 and 87 per cent respectively (Table 4).
Discussion

In previous publications that have addressed the issue of circumferential margin involvement, the patient populations under study were not homogeneous as some of the patients received adjuvant radiotherapy, and some studies also included palliative procedures1,2,4,5. Additionally, patients in whom there was intraoperative perforation of the tumour or bowel wall, a factor that would inuence the prognosis negatively, were included in one study3 and not described in others. Most previous reports were single hospital series, and thus may not have been representative of patients in general. In the present study every patient underwent a standardized operation (TME). Other treatment-related factors of possible prognostic importance, such as preoperative radiotherapy and peroperative bowel or tumour perforation, were excluded (Fig. 1). The pathological examination of the specimens was performed in a routine setting, but efforts were made to standardize handling of the specimens. Thus, the present study included a patient cohort, treated for cure in a standardized clinical manner that is particularly suited to address the prognostic importance of the circumferential margin. The frequency of circumferential margin involvement in the present study was 9 per cent (65 of 686), markedly lower
www.bjs.co.uk British Journal of Surgery 2002, 89, 327334

332

Circumferential margin in rectal cancer A. Wibe, P. R. Rendedal, E. Svensson et al.

Table 3

cancer

Risk factors for local recurrence, metastasis and mortality in 654 patients undergoing total mesorectal excision for rectal
Local recurrence Hazard ratio P* 001 10 28 (05, 154) 56 (12, 255) 55 (12, 251) 12 (26, 558) 005 10 21 (07, 61) 49 (13, 183) 049 10 15 (08, 31) 14 (06, 34) 076 10 21 32 26 22 (02, (04, (03, (03, 176) 245) 204) 188) 017 10 07 (03, 16) 33 (07, 156) 029 10 19 (02, 142) 34 (04, 287) 10 14 (04, 46) 23 (06, 80) 10 08 (05, 14) 13 (03, 55) 020 10 12 (05, 29) 31 (12, 81) 10 12 10 11 07 (04, (04, (04, (02, 31) 28) 29) 21) 065 10 08 (05, 13) 42 (20, 86) 0002 10 21 (13, 34) 28 (16, 50) 075 10 08 10 26 37 (02, (03, (08, (11, 30) 35) 85) 123) < 0001 10 20 (10, 40) 09 (03, 29) < 0001 10 10 (06, 16) 12 (07, 22) < 0001 10 07 20 15 47 (03, (10, (07, (23, 18) 38) 30) 95) 004 10 19 (10, 35) 34 (14, 78) 071 Metastasis Hazard ratio P* < 0001 10 18 29 31 37 (08, (14, (15, (17, 43) 61) 65) 84) 002 Mortality Hazard ratio P* 001

Circumferential margin (mm) > 20 1119 610 25 01 Tumour status (T) 2 3 4 Node status (N) 0 1 2 Age group (years) < 50 5059 6069 7079 > 80 Resection type Anterior resection APR Hartmann procedure Differentiation High Medium Low

Patients with T1 tumours were not included. Values in parentheses are 95 per cent condence intervals. APR, abdominoperineal resection. *Multivariate analysis and Cox regression

Table 4

Predictive values of circumferential margin involvement (01 mm) for local recurrence, metastasis and mortality in 686 patients undergoing total mesorectal excision for rectal cancer
Sensitivity (%) Local recurrence Metastasis Mortality 30 26 23 Specicity (%) 92 93 93 PPV (%) 22 40 37 NPV (%) 95 88 87

PPV, positive predictive value; NPV, negative predictive value

than that of others1,2,4,6. Similarly, the local recurrence rate in the whole group of patients was low (7 per cent). Consequently, the sensitivity (30 per cent) and positive predictive value (22 per cent) of circumferential margin involvement for the development of local recurrence were much lower than the values reported by Quirke et al.1 and others2. More important, however, is the high specicity (92 per cent) and negative predictive value (95 per cent) of an
British Journal of Surgery 2002, 89, 327334 www.bjs.co.uk

uninvolved circumferential margin (greater than 1 mm). Such a level of prediction can guide clinical practice, i.e. whether to use adjuvant postoperative radiotherapy or not. Only 5 per cent of patients in the present series with an uninvolved circumferential margin developed local recurrence. Thus, if it may be assumed that radiotherapy can halve the local recurrence rate, radiotherapy would have overtreated 97 per cent of the patients. The local recurrence rate was inversely related to the circumferential margin (Table 2). Fig. 3 demonstrates that the risk increment with time is of relatively little clinical importance as long as the margin is greater than 1 mm. It is with a margin of 1 mm or less that risk increases substantially, and an exponential increase in risk is demonstrated in Fig. 4. This is of practical importance, as a complete excision of the mesorectum will optimize the probability of obtaining a circumferential margin greater than 1 mm even in patients with a locally advanced tumour. The circumferential margin and tumour status inuenced local recurrence signicantly, while node status did
2002 Blackwell Science Ltd

A. Wibe, P. R. Rendedal, E. Svensson et al. Circumferential margin in rectal cancer 333

not. It has been well documented by others1013 that nodepositive patients are at higher risk of developing local recurrence. The analyses summarized in Table 3 show, however, that the effect of node status is negated when the variable circumferential margin is introduced in the analysis. The clinical interpretation of this nding is that the circumferential dissection that characterizes TME eliminates the inherent risk of lymph node involvement as long as the margin is larger than 1 mm. This is illustrated in the group of patients with Dukes' C tumours in whom the local recurrence rate was 9 per cent if the circumferential margin was greater than 1 mm (data not shown). The circumferential margin is also an independent risk factor for the development of distant metastases and mortality (Table 3). For metastases, circumferential margin involvement is a detrimental factor which increases risk exponentially, while there is a more gradual risk relationship between circumferential margin and survival (Figs 5 and 6). Hall et al.6 have claimed that circumferential margin involvement may reect an advanced nature of the tumour rather than an inadequate surgical clearance. There were, however, contradictory and unexpected ndings in their study. For instance, the local recurrence rate was similar in the margin-negative (11 per cent) and margin-positive (15 per cent) groups (in the present study the equivalent rates were 5 and 22 per cent). The conclusion of Hall et al. was based on a cohort of 152 patients, compared with 686 patients in the present study, which may explain the different results. The present data show that advanced disease (higher tumour status, more advanced node status and poor differentiation of tumour) is a signicant risk factor, but of less importance than obtaining a tumour-free circumferential margin. Thus, surgeons should maximize the probability of obtaining a tumour-free margin by adopting the operative principles of TME. Of all the independent risk factors, the only one that can be manipulated by treatment is the circumferential margin. In conclusion, this study has provided evidence that following routine TME and pathological examination of the surgical specimens for rectal cancer the circumferential margin is an important prognostic indicator of local recurrence, distant metastasis and survival. This knowledge can be used in the selection of patients for postoperative adjuvant therapy.
Acknowledgements

The authors are indebted to all the Norwegian surgeons and the 55 hospitals that have performed the rectal cancer surgery and reported the data over the years of the study. They are also grateful to all the pathologists who examined the specimens, and wish to thank Richard J. Heald, Erik Carlsen and Johan N. Wiig for their enthusiastic support of the Norwegian Rectal Cancer Project. Steering group
The Norwegian Gastrointestinal Cancer Group (chairperson Olav Dahl, Department of Oncology, Haukeland Hospital, Bergen).

Project group members


Tormod Bjerkeset, Innherred Hospital; Erik Carlsen, Ullevaal Hospital; Tor J. Eide, The National Hospital; Olav Dahl, Haukeland Hospital; Tom Gerner, Oppland Central Hospital, Lillehammer; Frydis Langmark, The Cancer Registry of Norway; Helge E. Myrvold, Trondheim University Hospital; Jarle Norstein, The Cancer Registry of Norway; Arthur Revhaug, Troms University Hospital; Geir Rnning, Nordland Central Hospital; Arne Skarstein, Haukeland Hospital; Jon Arne Sreide, Rogaland Central Hospital, Odd Sreide, The National Hospital (chair); Asgaut Viste, Haukeland Hospital; and Johan N. Wiig, The Norwegian Radium Hospital.

Participating hospitals and surgeons


Aker Hospital (Arild Nesbakken), Akershus Central Hospital (Arne E. Frden), Aust-Agder Central Hospital (Tore Gauperaa), Buskerud Central Hospital (Ulf Jacobsen), Brum Hospital (Rocio Rosales), Diakonhjemmet Hospital (Jan Helge Solhaug), Diakonissehjemmets Hospital (Trond Haugstvedt), Gjvik Hospital (Olav Hstmlingen), Gravdal Hospital (Bjrn Bergsdal), Hammerfest Hospital (Svein Hammelbo), Harstad Hospital (Trond Ellingsen), Haugesund Hospital (Knut Svendby), Haukeland Hospital (Arne Skarstein), Hedmark Central Hospital (yvind Graadal, Indre stfold Hospital (Thor Arnestad), Inherred Hospital (Tormod Bjerkeset), Kirkenes Hospital (Lennart Wa Kongsberg Hospital (Fredrik Kylberg), hlby), Kongsvinger Hospital (Mikael Bech), Kristiansund Hospital kan Steen), Larvik Hospital (Odd Ruistuen), Oppland (Ha Central Hospital, Lillehammer (Bjrn H. Nilsen), Lrdal Hospital (Fritjof Lund-Larsen), Molde Hospital (Arve Sb), Moss Hospital (Torgeir Lvig), Mre og Romsdal Central Hospital (Gustav R. Hareide), Namdal Hospital (Hallvard Grsli), Narvik Hospital (Tor Teigan), Nordland Central Hospital (Jermund Liljedal Jr), Notodden Hospital (Hans
www.bjs.co.uk British Journal of Surgery 2002, 89, 327334

This work was supported by a grant from the Norwegian Cancer Society and the Norwegian Medical Association's Fund II for Quality Assurance.
2002 Blackwell Science Ltd

334

Circumferential margin in rectal cancer A. Wibe, P. R. Rendedal, E. Svensson et al.

Thomas Fjre), Odda Hospital (Gauti Arnthorsson), Orkdal Sanitetsforenings Hospital (Kristen P. Benediktsson), Rana Hospital (Ranveig K. Aspevik), The Norwegian Radium Hospital (Johan N. Wiig), The National Hospital (Odd Sreide), Ringerike Hospital (John Colin Poole), Rjukan Hospital (Kjetil Unneberg), Rogaland Central Hospital (Hartwig Krner), Sandefjord Hospital (Hans Petter Nss), s), Sogn og Fjordane Sandnessjen Hospital (John Kufa Central Hospital (Nils Sletteskog), Stensby Hospital (Johan Lykke), Stord Hospital (Olav Stray), Storkmarknes Hospital (Rune Svensen), Telemark Central Hospital (John Bjrn Nilsen), Troms University Hospital (Arthur Revhaug), Trondheim University Hospital (Helge E. Myrvold), Tynset Hospital (Bjrn Nordby), Ullevaal Hospital (Erik Carlsen), Vefsn Hospital (istein Ylvisaker), Vest-Agder Central Hospital (Leif . Heen), Vestfold Central Hospital (Geir Haarberg), Volda Hospital (Yngve Nordervald), Voss Hospital (Hans Jrgen Nielsen), stfold Central Hospital (Thorleif E. Wilhelmsen).
References
1 Quirke P, Durdy P, Dixon MF, Williams NS. Local recurrence of rectal adenocarcinoma due to inadequate surgical resection. Histopathological study of lateral tumour spread and surgical excision. Lancet 1986; ii: 9969. 2 Ng IO, Luk IS, Yuen ST, Lau PWK, Pritchett CJ, Ng M et al. Surgical lateral clearance in resected rectal carcinomas. A multivariate analysis of clinicopathologic features. Cancer 1993; 71: 19726. 3 de Haas-Kock DF, Baeten CGMI, Jager JJ, Langendijk JA, Schouten LJ, Volovics A et al. Prognostic signicance of radial margins of clearance in rectal cancer. Br J Surg 1996; 83: 781 5. 4 Adam IJ, Mohamdee MO, Martin IG, Scott N, Finan PJ, Johnston D et al. Role of circumferential margin involvement

10

11

12

13

in the local recurrence of rectal cancer. Lancet 1994; 344: 707 11. Cawthorn SJ, Parums DV, Gibbs NM, A'Hern RP, Caffarey SM, Broughton CIM et al. Extent of mesorectal spread and involvement of lateral resection margin as prognostic factors after surgery for rectal cancer. Lancet 1990; 335: 10559. Hall NR, Finan PJ, al-Jaberi T, Tsang CS, Brown SR, Dixon MF et al. Circumferential margin involvement after mesorectal excision of rectal cancer with curative intent. Predictor of survival but not local recurrence? Dis Colon Rectum 1998; 41: 97983. Norwegian Gastrointestinal Cancer Group. Colorectal and anal cancer a guideline for doctors. Tidsskr Nor Laegeforen 1999; 119(Suppl): 413. Quirke P, Dixon MF. The prediction of local recurrence in rectal adenocarcinoma by histopathological examination. Int J Colorectal Dis 1988; 3: 12731. Marsh PJ, James RD, Schoeld PF. Denition of local recurrence after surgery for rectal carcinoma. Br J Surg 1995; 82: 4658. Enker WE, Havenga K, Thaler HT, Cohen AM, Polyak T, Cranor M et al. Total mesorectal excision with pelvic autonomic nerve preservation in operative treatment of rectal carcinoma. In: Sreide O, Norstein J, eds. Rectal Cancer Surgery: Optimisation, Standardisation, Documentation. Berlin: Springer, 1997: 22043. Jatzko GR, Jagoditsch M, Lisborg PH, Denk H, Klimpnger M, Stettner HM. Long-term results of radical surgery for rectal cancer: multivariate analysis of prognostic factors inuencing survival and local recurrence. Eur J Surg Oncol 1999; 25: 284 91. Hohenberger W, Schick CH, Go hl J. Mesorectal lymph node dissection: is it benecial? Langenbecks Arch Surg 1998; 383: 4028. Zirngibl H, Husemann B, Hermanek P. Intraoperative spillage of tumor cells in surgery for rectal cancer. Dis Colon Rectum 1990; 33: 61014.

British Journal of Surgery 2002, 89, 327334

www.bjs.co.uk

2002 Blackwell Science Ltd