A regression analysis of prognostic factors after resection of Dukes' B and C carcinoma of the rectum and rectosigmoid. Does post-operative radiotherapy change the prognosis?

The prognostic value of several clinical and histopathological characteristics has been evaluated in patients with Dukes' B and C carcinoma of the rectum and the rectosigmoid. Data on 260 Dukes' B and 208 Dukes' C tumours entered into a prospective, randomized clinical trial of post-operative radiotherapy (50 Gy given with 2 Gy/fraction in an overall time of 7 weeks) were analyzed by means of the Cox proportional hazards model. The Dukes' stages B and C were analyzed in two separate multivariate analyses. In patients with Dukes' B tumours, a poor prognosis was associated with age above 60, perineural and venous invasion, tumour located less than 10 cm from the anal verge and elevated pre-operative carcinoembryonic antigen (CEA) (greater than 3.2 ng ml-1). In patients with Dukes' C tumours, perineural and venous invasion, tumour located less than 10 cm from the anal verge, and elevated pre-operative CEA were associated with a poor prognosis. In addition, a large tumour diameter had a strong, negative influence on the prognosis. Males seemed to have a poorer prognosis than females among the Dukes' C patients. Resection of neighbouring organs was also associated with a poor prognosis in this stage. Post-operative radiotherapy as administered in the present series had no significant influence on prognosis. Based on the derived prognostic models patients with a hazard of death above the median in each stage were selected. A separate analysis of the survival in these high risk patients showed no survival benefit from radiotherapy. The proportional hazards model may be a useful tool in selecting patients for more aggressive adjuvant treatment.