Adenocarcinoma of the rectum treated by abdominoperineal excision: multivariate analysis of prognostic factors.

The aims of this study were to describe the results of 100 consecutive abdominoperineal excisions, and to try to define homogeneous subgroups of survivors using the Cox regression analysis model. There was one postoperative death. No patient was lost to follow-up. Overall 5-year survival rate was 45%. Multivariate analysis selected 3 prognostic factors: lymph node involvement (p less than 0.001), local tumour extent (P = 0.08), and extension to adjacent organs (p = 0.03). Four subgroups of patients were defined. Best survival was observed in patients (n = 43) with tumour invading into the subserosa (W1) or less and without lymph node metastasis (N0), with a 5-year survival of 82% (group 1). Patients in Group 2 (n = 18) had a tumour invading into the muscularis propria or less with lymph node metastasis (W0 N+), or into the serosa or perirectal fat without lymph node invasion (W2 N0), with a 5-year survival of 51%. Patients in Group 3 (n = 17) had a tumour invading into the subserosa with lymph node invasion (W1 N+) or extension to adjacent organ without lymph node invasion (W3 N0) with a 5-year survival of 24%. Patients in group 4 (n = 22) had extension into the serosa or perirectal fat or more with lymph node invasion (W2 N+) with a 5-year survival of 2%. This scoring system is simple, practical and easy to apply. If it can be confirmed by other studies, its routine adoption could be justified.