The Large Bowel Cancer Project is a collaborative prospective study of 4228 patients with a histologically proven adenocarcinoma, of whom 2336 (55 per cent) survived a 'curative' resection. Follow-up information is available on 2220 patients (95 per cent). Subsequently, 309 (14 per cent) have developed a local recurrence confirmed by: biopsy (127; 41 per cent), clinical examination (77; 25 per cent), X-ray (15; 5 per cent), a raised CEA (2; 1 per cent), or some other method - e.g. CT scan or a confident unbiopsied laparotomy finding (88; 29 per cent). Statistically significant factors (chi2 test, P less than 0.05) associated with local recurrence are: Dukes' classification: A 4 per cent; B 13 per cent; C 18 per cent Tumour differentiation: Well 11 per cent; Moderate 14 per cent; Poor 21 per cent Obstruction: Absent 13 per cent; Present 21 per cent Perforation: Absent 13 per cent; Present 28 per cent Tumour mobility: Freely mobile 11 per cent; Others 21 per cent Operation performed (rectal and rectosigmoid tumours): Abdomino-perineal 12 per cent; Anterior resection 18 per cent; Surgeon (Consultant only): Range less than 5 per cent to greater than 20 per cent. Stratification of the above variables altered only the statistical significance pertaining to tumour differentiation (P less than 0.1, d.f. = 2). In particular, the differences between Consultant surgeons remained.

SUMMARY A review of histopathology reports on 2046 patients in the large bowel cancer project showed considerable observer variation in histological grading, Dukes staging, and lymph node harvest. These parameters have a well-established relationship to prognosis, but, if they are to be applied for both clinical and research purposes, they must be assessed consistently. A minimal level of information which should be recorded from a resection specimen is suggested, with a description of the methods by which this information can be obtained.The large bowel cancer project was initiated in 1976, and currently specimens are sent to the 22 histopathology departments from the 84 participating surgeons. Although many departments have more than one histopathologist to deal with these specimens, they have been treated as 22 observers for this analysis.The objective of this part of the study was to assess the consistency of reports on the histopathology of the resected specimens. We anticipated that there might be considerable observer variation in histological grading, which is a subjective process, but we did not expect significant differences in the staging of local tumour spread, which is an objective assessment with sharp delineation between subgroups, or in lymph node harvest. We report here the results on 2046 resected tumours. We found clinically important and statistically significant differences between histopathology departments in the reporting of these specimens. Differences between hospitals have been calculated using the x2 test for independent samples. (Fig. 1). The difference between observers was statistically significant (p < 0-001). Some lack of uniformity is inevitable in grading, but the order of difference shown can be due only to different techniques of assessment. In the present study of 2046 specimens 26 0 were well, 58% moderately, and 16% poorly differentiated;however, the proportion placed in each grade by the different observers varied widely: well-differentiated 3-93 %, moderately 8-82 %, and poorly 5-30 %

1988 patients with an adenocarcinoma of the rectum (1292; 65 per cent) or rectosigmoid (696; 35 per cent) have been studied. A resection (1700 patients) or polypectomy (124 patients) was performed in 1824 (92 per cent) and, of the former, 1376 patients either underwent abdominoperineal (AP) excision of the rectum (788 patients) or an anterior resection (598 patients). The in-hospital mortality was 63 patients (8 per cent) for AP and 44 (7 per cent) for anterior resection, and a curative resection had been performed in 504 (71 per cent) of those undergoing an AP, and 393 (71 per cent) of those undergoing an anterior resection. Follow-up information is available for 478 patients (95 per cent) who underwent an AP and 370 (94 per cent) who underwent an anterior resection. More patients have developed a local recurrence after an anterior resection (67; 18 per cent) than after AP (57; 12 per cent) (Logrank chi2 = 6.6, d.f. = 1, P less than 0.02) (stratified for sex and Dukes' stage). This difference is not accounted for by a lesser margin of distal clearance after an anterior resection; firstly because the margin of clearance was not different in those who did and those who did not develop a local recurrence (AP: whole group = 4.4 cm, local recurrence = 4.5 cm; anterior resection: whole group = 3.0 cm, local recurrence = 3.1 cm) and secondly because for each centimetre of distal clearance there was a consistently greater probability of recurrence for anterior resection (Logrank chi2 = 9.1, d.f. = 1, P less than 0.01) (stratified for sex, Dukes' stage and distal clearance margin).

Modifications of Dukes' (1932) classification of rectal tumours have led to confusion. From the data of 2518 patients who had undergone curative colorectal surgery the interrelationships between tumour penetration, grade, vascular invasion and pattern of lymph node involvement have been examined and their individual relevance to survival determined. Subdivision of Dukes' A cases into those confined to the muscularis mucosae (A) and those penetrating into, but not through, the bowel wall (B1) should be abandoned. Despite interrelationships between lymph node status, grade of tumour and vascular invasion, they all contribute prognostic information independent of each other. Apical lymph node involvement, more than four lymph nodes involved and extensive primary tumours with nodal involvement all carry a bad prognosis. Although interrelated each variable is individually relevant. However, subgroups of patients with Dukes' C tumours have an observed survival significantly better than expected. When few lymph nodes are involved or the primary tumour is confined to the bowel wall but lymph nodes are involved, the expectation of life is equivalent to Dukes' B.

Summary and conclusionsPatients with obstructing large-bowel cancer may be treated by primary tumour resection or the conventional staged tumour resection, and a prospective study comparing these two treatments was carried out. The postoperative outcome in 174 patients (of whom 90 underwent primary and 47 staged tumour resection) showed that the overall mortality was similar in both groups but that the duration of hospital stay in patients who underwent primary tumour resection was half that of those who underwent staged tumour resection. The mortality for primary tumour resection, however, was unexpectedly high for lesions proximal to the splenic flexure and unexpectedly low for lesions distal to this point. Of patients with distal tumours in whom a staged resection was planned, 35% died after a loop colostomy. The most striking result was that the ratio of postoperative death for trainee surgeons compared with fully trained surgeons was 3:1.