A multivariate analysis of clinical and pathological variables in prognosis after resection of large bowel cancer.

Data on 709 patients who had a resection for colorectal carcinoma at Concord Hospital between 1971 and 1980 were studied to determine the independent effects on survival of several patient characteristics and pathological variables using the Cox regression model. Clinicopathological stage had the strongest association. Other variables ranked according to their relative importance independent of stage were: histological grade, level of direct spread, the presence of venous invasion, age and sex of the patient and the presence of obstruction.