457 Analyses of fats, oils, and lipoproteins. Champaign, Illinois: American 26. Torel J, Cillard J, Cillard P. Antioxidant activity of flavonoids and Oil Chemists’ Society, 1991: 524-54. reactivity with peroxy radicals. Phytochemistry 1986; 25: 383-85. 21. Singleton SL, Rossi JA. Colorimetry of total phenolics with 27. Kanner J, Mendel H, Budowski P. Prooxidant and antioxidant effects of phosphomolybdic-phosphotungstic acid reagents. Am J Enol Vitic ascorbic acid and metal salts in a &bgr;-carotene-linoleate model system. 1965; 16: 144-58. J Food Sci 1977; 42: 60-64. 22. Frankel EN, German JB, Davis PA. Headspace gas chromatography to 28. Moroney M-A, Alcaraz MJ, Forder RA, et al. Selectivity of neutrophil determine human LDL oxidation. Lipids 1992; 27: 1047-51. 23. Kanner J, Harel S. Initiation of membranal lipid peroxidation by 5-lipoxygenase and cyclo-oxygenase inhibition by an anti- activated metmyoglobin and methemoglobin. Arch Biochem Biophys inflammatory flavonoid glycoside and related aglycone flavonoids. 1985; 237: 314-21. J Pharm Pharmacol 1988; 40: 787-92. 24. Afanaslev IB, Dorozhko AI, Brodskii AV, et al. Chelating and free radical 29. Das NP. Studies on flavonoid metabolism: absorption and metabolism of scavenging mechanisms of inhibitory action of rutin and quercetin in (+ )-catechin in man. Biochem Pharmacol 1971; 20: 3435-45. lipid peroxidation. Biochem Pharmacol 1989; 38: 1763-69. 30. Gugler R, Leschik M, Dengler HJ. Deposition of quercetin in man after 25. Husain SR, Cillard J, Cillard P. Hydroxyl radical scavenging activity of single oral and intravenous doses. Eur J Clin Pharmacol 1975; 9: flavonoids. Phytochemistry 1987; 26: 2489-91. 229-34. Mesorectal excision for rectal cancer Concern about world wide local recurrence rates Introduction for rectal cancer of 20-45%, together with anxiety at To determine the standards of outcome that are the recent proliferation of adjuvant therapies, led us acceptable for surgical treatment of rectal cancer, we to review the efficacy of total mesorectal excision examined our data on total mesorectal excision (TME), (TME) with which no adjuvant therapy had been which was first reported in 1982.1 The temptation to combined. dismiss our low local and overall recurrence rates as a Precise, sharp dissection is undertaken around statistical aberration-eg, the result of case mix and the integral mesentery of the hind gut, which analytical techniques-was scrutinised objectively. Further envelopes the entire mid rectum. This procedure objective corroboration was provided by an oncologist adds to operative time and complications but has (R. D. H. R.) who has overseen our follow-up clinic for 14 years. been claimed to eliminate virtually all locally recurrent disease after "curative" surgery. Patients and methods Independent analysis (J. K. M.) of prospective Our independent validator and assessor (J. K. M.) reviewed all follow-up data extended over a 13-year interval cases of rectal cancer seen in our unit and updated the anterior resection results published in 1986.1 He then examined a high-risk (1978-91; mean 7·5 years). The actuarial local recurrence rate after curative anterior resection at 5 group of patients who were most likely to recur according to the definitions of Krook et al from the North Central Cancer Treatment years is 4% (95% Cl 0-7·5%) and the overall Group (NCCTG).4 These investigators excluded Dukes’ A (Astler recurrence rate is 18% (10-25%). 10-year figures Coller A + Bl) lesions, those more than 12 cm from the anal verge, are 4% (0-11%) and 19% (7-32%). In view of the and those considered "not curative" by the surgeon. Figures 1 and 2 show the distribution of cases according to the operative high-risk classification used for the North Central management and Dukes’ stage, respectively. J. K. M. found 2 Cancer Treatment Group (NCCTG), which has led patients who had received systemic 5-fluorouracil for six months to a trend to chemoradiotherapy, a similar group of because of personal preference (one death, one survival), and he high-risk Basingstoke cases was constructed for excluded these to leave a high-risk group of 135 cases. Data are stored on a computer database and they are updated regularly by a comparison purposes. This group included 135 full-time research worker. Patients are examined and undergo consecutive Dukes’ B (B2) and Dukes C cancer carcinoembryonic antigen testing every three months for the first operations, both anterior resection and abdominal- two years, then six monthly for three years, and yearly thereafter. perineal excision, for tumours below 12 cm from the No patient was lost to follow up and all deaths have been anal verge. Results from TM E alone are substantially documented; three doubtful cases were classified as rectal cancer deaths. superior to the best reported (NCCTG) from Recurrence and survival curves were generated by the Kaplan- conventional surgery plus radiotherapy or Meier method and the log-rank statistic was used for comparisons. combination chemoradiotherapy: 5% local Confidence intervals were taken at the 95% level. Our study is recurrence at 5 years compared with 25% and ongoing but this analysis is at a median follow-up time of 7-7 years, 13·5%, respectively; and 22% overall recurrence which corresponds with the NCCTG trial. compared with 62·7% and 41·5%, respectively Definitions (Dukes’ B cases [B2], 15%; Dukes’ C cases, 32%). The operation was defined as curative when, at the end of the Meticulous TM E, which encompasses the whole procedure, the surgeon believed that all grossly detectable cancer field of tumour spread, can improve cure rates and had been removed. Local and overall recurrence are defined as reduce the variability of outcomes between disease detected or suspected within the pelvis (local) and anywhere surgeons. Far more genuine "cures" of rectal cancer are possible by surgery alone than have generally ADDRESSES: University of British Columbia (Prof J. K. MacFarlane, MD), Vancouver, Canada; Wessex Radiotherapy Centre, (R. D. H. been believed or are currently accepted. Better Ryall, FRCR) Southampton, UK; Colorectal Research Unit (R. J. surgical results are an essential background for the Heald, MChir), Basingstoke District Hospital, UK. Correspondence to Mr R. J. Heald, Colorectal Research Unit, Basingstoke District Hospital, more selective use of adjuvant therapy in the future. Aldermaston Road, Basingstoke, Hampshire RG24 9NA, UK. 458 AR, anterior resection. AP, abdominal-perineal excision. (local or systemic) during life or at necropsy, respectively. The 25 % of procedures defined as non-curative included those with distant metastases, together with a smaller group (11% locally non- invasive) in which residual disease was believed to remain on the side wall of the pelvis. Low-risk exclusions included 51 cases of Dukes’ A cancer, with one case of local recurrence (probably due to implantation and which was subsequently cured) and two deaths from cancer. The exclusion of lesions 12-15 cm from the anal verge eliminated only five cases of tumour recurrence (none local). Surgical technique Meticulous sharp dissection of the avascular plane between mesorectum and parietes is completed under direct vision. The excised specimen therefore includes the entire posterior, distal, and lateral mesorectum out to the plane of the inferior hypogastric plexuses, which are carefully preserved. Anteriorly the specimen includes intact Denonvillier’s fascia and the peritoneal reflection (fig 3 upper). The characteristic smooth bi-lobed encapsulated appearance posteriorly and distally reflects the contours of the pelvic floor and the midline anococcygeal raphe (fig 3 middle). Implantation is prevented by aqueous washout below the clamp before the anorectum is divided, and in the pelvis itself. Results The first part of our study provides current lifetables for comparison with the figures published in 1986 for curative anterior resections (fig 4).’ Actuarial local recurrence is unchanged at 4% (95% CI 0-7-5%) as is overall recurrence (18%; 10-25%). If locally non-curative anterior resections are included, these figures are altered to 6% (0-11%) and 22% (13-29%). 10-year data for curative anterior resections show recurrence rates of 4% (0-11%) and 19 % Fig 3-Operative stages. (7-32%), respectively. Inclusion of abdominal-perineal Upper antenor aspect of mesorectum dunng surgery with seminal excisions alters overall recurrence to 19% at 5 years and vesicals in front and Denonmllier’s fascia behind 22% at 10 years. Extramural large-vein invasion and Middle posterior aspect of mesorectum dunng surgery showing the tumour differentiation have now become statistically characteristic bi-lobed lipoma appearance. Lower: the pnncipal parasympathetic engent nerve root preserved on significant indicators (p<0-001). Recurrence in cases of the right side of the pelvic wall Dukes’ C carcinoma has fallen from 42% to 32% at the 5-year point, and there is a trend to fewer episodes of local recurrence in the second half of the series; only two of the six occurring after curative anterior resections between 1986 and the end of 1992. The surgical results ofTME were then analysed in detail for 135 consecutive high-risk cases (table 1). 126 anterior resections and 9 abdominal-perineal excisions (6-1%) have been done in this group, although this figure is possibly low because more abdominal-perineal excisions were undertaken by other surgeons. There were 79 male and 56 female patients and the average age at operation was 67 years 459 leakage, all patients having had a contrast enema; and units of blood transfused. Significant differences were found with the first four criteria. Cases of Dukes’ stage B (85 % 5-year survival) were significantly better for overall recurrence than cases of Dukes’ stage C (68%) (p 0-05-0-025) with no differences = in local recurrence (3 B, 2 C). Although numbers of abdominal-perineal excisions are small, 5/9 recurrences do yield significant worse results (p= <0 01). There were no differences or trends for any of the last six variables. During 13 years of follow-up, 11 patients have developed non- colorectal malignancies (table II). Discussion The management of the patient with rectal cancer has challenged oncologists for decades. Ten years ago, Silen Fig 4-Life-table to recurrence for curative anterior observed that over half of all recurrences are located in the resections during 1978-91. pelvis and constitute a failure of surgery in a disease that is (range 36-90). 3 patients died within 30 days of surgery: a unique in its surgeon variability.s Disappointing results 2% peri-operative mortality rate. from more traditional surgical practices have led to clinical trials of various regimens of radiotherapy and During 13-7 years, 38 patients have died of causes other than rectal carcinoma. Their average age at surgery was 74-9 chemotherapy.". The current consensus is that irradiation reduces local recurrence but survival benefit remains years (range 41-5-90-7) (table II). Overall recurrence-free cancer-specific survival was 78% unproven.8,11Pahlman, in a Swedish trial, showed the (95% CI 68-88%) according to life-table analysis at 5 years. superiority of pre-operative over postoperative treatment, 24 patients had a recurrence, 84% of whom were apparent and he has suggested that pre-operative radiotherapy would within 3-5 years, 92% within 5 years, and 100% at 6 years. approximately halve the risk of local recurrence in any given Their average age at surgery was 69 4 years and survival group at risk.12 Krook and colleagues have claimed that the combination averaged 2-7 years. Local recurrence rate was 5% on life-table analysis at 5 years (95% CI 0-11%). 7 patients of chemotherapy and radiotherapy improves "significantly and substantially" the surgical results for rectal cancer with a developed a local recurrence, 5 after anterior resection (5/128=3-8%), and 2 after abdominal-perineal excision "poor prognosis".4 However, should prognosis really be so (2/9=22%); 4 of these cases were local alone, while 3 had poor for Dukes’ B (B2) and Dukes’ C carcinoma? And are metastases. current surgical techniques the best for the challenge posed Life tables for local and overall recurrence were by rectal cancer? We believe that the answer is no to both constructed for subgroups according to the following questions. criteria: choice of operation (anterior resection or Whilst the NCCTG and National Surgical Adjuvant abdominal-perineal excision); Dukes’ stage; histological Project for Breast and Bowel Cancers trials were proceeding, a prospective study of total mesorectal excision was under grading; extramural vascular invasion (identified by section of the mesorectum outside the cancer); tumour height above way in Basingstoke.1 Despite similar criteria and consecutive anal verge (providing anterior resection is chosen) (4-8 cm patients, our high-risk cases may not be precisely and 8-12 cm); anastomotic height above anal verge (2-4 cm comparable, but the use of such groups does have two and 4-6 cm); distal muscle tube margin (on the specimen advantages. First, exclusion of Dukes’ A (or A+B1) and without stretching and after fixation); sex; anastomotic non-curative cases limits the effect of selection or referral factors. Second, the study groups comprise the cases most TABLE I-LOCAL AN D TOTAL RECURRENCERATES FOR TME AND likely to be influenced by treatment so that differences NCCTG DATA become apparent more readily. 7 of our 8 local recurrences in the curative group took place in this subgroup and only 7 recurrences of any kind occurred among excluded patients. This result suggests that the NCCTG definition of "high risk" is indeed valid and that such subgroups are ideal for comparisons between series. Inclusion of all locally non- curative anterior resections only increases local recurrence by 2% and overall by 4%. TABLE ll-DEATHS FROM OTHER CAUSES Fig 5-Comparison of lifetables for recurrence-free interval Average age at operation 74.9 years. between NCCTG series (left) and our series of TME (right). 460 In the NCCTG control arm, surgery plus radiotherapy recurrence rates from 33% to only two local recurrences produced a 5-year local recurrence rate of 25% and an among 41 cases: simply by application of this technique and overall recurrence rate of 62-7%.2 The addition of the support of his colleagues in referring cases.16 At least 6 chemotherapy to the treatment arm reduced these figures to British colorectal surgeons, applying these principles, have 13-5% and 42%, respectively. It is these differences that published or presented single-figure local recurrence rates have influenced treatment throughout the world. The after curative resections, and similar results are undoubtedly Basingstoke actuarial local recurrence figure for high-risk achieved by some surgeons in the USA and elsewhere. cases is 5% and for overall recurrence is 22% (table I and Institutional figures have almost invariably been inferior. fig 5). These results are also superior to those from the Specialisation could create a background of less variable NSABP trial where a 5-year recurrence rate of 70% was surgical results against which indications for adjuvant seen in a similar group treated by surgery alone.7 therapy can be more selectively defined by careful analysis of Such large differences imply that viable local tumour risk factors for both local and disseminated recurrence. residues remained more frequently after operation in these Widespread adoption of the principles of TME outlined in studies than in our own. The only credible explanation is this paper should lead to a reduction in the number of that the field of spread of cancer in the pelvis is susceptible to patients requiring combination therapy, together with a what may at first sight seem to be small differences in greater certainty that they will benefit from such therapy. surgical technique. TME is more difficult and time TME was developed from an anatomical understanding consuming but not more radical. The dissection avoids the based on what is surgically feasible: its importance depends autonomic nerves by defining the place within them-this on results. The data presented here suggest that the usual "holy plane" is subjected to sharp dissection under direct field of spread of rectal cancer is confined within the vision throughout.13 Complete excision of the enveloping mesorectal envelope and that rectal carcinoma is far more mesorectal tissues in a covering of thin but recognisable curable by surgery alone than has generally been believed or fascia protects the pelvic walls, genito-urinary structures, is currently accepted. Most carcinomas that recur initially presacral nerves, parasympathetic roots (fig 3 lower), and within the pelvis could probably have been cured by better neurovascular bundles, and creates a specimen with a thin surgery. Concentration of effort and resources into the mobile covering of areolar tissue. The use of the same improvement of training in surgical technique would pay procedure in all cases, irrespective of the location of the great dividends in the management of this common visceral tumour in the rectal tube, requires specific surgical skills and malignancy. strengthens the case for specialised referral patterns. Over We thank the dedication of Mrs Rosemary Sexton, our full-time research 90 % of all patients referred to our unit during this time were assistant, the support of the Royal College of Surgeons and Physicians of operated on by one surgeon (R. J. H.). Canada and the University of British Columbia, and the ongoing financial The cost of this success is not inconsequential. The support of the Wessex Cancer Trust. We also thank Prof William Silen for his meticulous dissection procedure requires additional help and advice with the manuscript. operating time (up to 25 h) and blood transfusion REFERENCES requirements increase. Anastomoses at 3-6 cm from the anal 1. Heald RJ, Ryall RDH. Recurrence and survival after total mesorectal verge have led to a 17-4% anastomotic leak rate for TME excision for rectal cancer. Lancet 1986; i: 1497-82. (11% clinical and 6-4% radiological); we now routinely 2. Heald RJ, Husband EM, Ryall RDH. The mesorectum in rectal cancer fashion a temporary protective colostomy.14 Long-term surgery: the clue to pelvic recurrence? Br J Surg 1982; 69: 613-16. anastomotic failure, significantly morbidity, and sometimes 3. Isbister WH. Personal view: Basingstoke revisited. Aust NZJ Surg 1990; 60: 243-46. a permanent stoma have marred the recovery of about 5 % of 4. Krook JE, Moertel CG, Gunderson LL, et al. Effective surgical adjuvant these patients. These drawbacks seem justified in terms of therapy for high risk rectal carcinoma. N Engl J Med 1991; 324: 709-15. the improvement in local recurrence. 5. Silen W. Colon and rectum. In: Schwartz E, ed. The year book of Of even greater importance is our fmding that improved surgery. Chicago: Year Book, 1983: 368-70. 6. Galandiuk S, Weiand HS, Moertel CG, et al. Patterns of recurrence after local excision reduces metastatic disease developing curative resection of carcinoma of the colon and rectum. Surg Gynaecol subsequent to local failure, and that the 5-year tumour-free Obstet 1992; 174: 27-32. survival of 78 % represents a real increase in patients actually 7. Fisher B, Wolmark N, Rockette H, et al. Post operative adjuvant cured of cancer. This finding is at variance with the notion chemotherapy or radiation therapy for rectal cancer results from NSABP protocol R-01. J Natl Cancer Inst 1988; 80: 21-29. that cancer is a systemic disease. Clearly, rectal cancer 8. Gastrointestinal Tumour Study Group. Prolongation of the disease free usually is not. This observation, the key to our whole interval in surgically treated rectal carcinoma. N Engl J Med 1985; 312: philosophy of this disease, is confirmed by the absence of 1465-72. recurrences beyond 59 years in this 13-year study. 9. Gerard A, Buyse M, Nordlinger B, et al. Pre-operative radiotherapy as Few papers in this century have had so profound an effect adjuvant treatment in rectal cancer: final results of randomised study of the EORTC. Ann Surg 1988; 208: 606-14. on the management of colorectal cancer as the two 10. Moertel CG, Fleming TR, MacDonald JS, et al. Levamisole and publications from the NCCTG at the Mayo Clinic and fluorouracil for adjuvant therapy of resected colon carcinoma. N Engl J Duke University.4,10 These and other studies led to the Med 1990; 322: 352-58. 11. Izar F, Fourtanier G, Pradere B, et al. Pre-operative radiotherapy as publication of a National Institutes of Health consensus adjuvant treatment in rectal cancer. World J Surg 1992; 16: 106-12. document on adjuvant treatment for colorectal cancer.1s 12. Pahlman L, Glimelius B, Graffman S. Pre versus post operative Throughout the world patients and their advisers confront radiotherapy in rectal carcinoma: an interim report from a randomized what is really a loaded question: should chemo-radiotherapy multicentre trial. Br J Surg 1985; 72: 961-66. 13. Heald RJ. The "Holy Plane" of rectal surgery. J Roy Soc Med 1988; 81: be initiated after surgery? In poorer countries, health-care 503-08. providers wrestle with the cost and personnel implications 14. Karanjia ND, Corder AP, Holdsworth PJ, Heald RJ. Risk of peritonitis of what threatens to become a revolution in the management and fatal septicaemia and the need to defunction the low anastomosis. of this common cancer. Can current priorities be correct Br J Surg 1991; 78: 196-98. when superior results can be achieved by TME without 15. NIH Consensus Conference. Adjuvant therapy for patients with colorectal cancer. JAMA 1990; 264: 1444-50. adjuvant modalities? One Swedish surgeon who visited 16. Nilsson E. Reducing the risk of local recurrence in rectal cancer. Basingstoke in 1989 has reported a reduction in local Onkologisk Forum, Trondheim November 1992.