Objective: To determine whether the presence of tumour at a free serosal surface was independently associated with pelvic recurrence or survival in patients who had a resection for clinicopathological stage B or stage C rectal cancer and who had not received adjuvant therapy.
Method: Data were drawn from a comprehensive, prospective hospital registry of all resections for rectal cancer from January 1971 to December 1998 with follow up to December 2003. Statistical analysis employed the chi(2) test or Fisher's exact probability, Kaplan-Meier estimation and proportional hazards regression, with a significance level of < or =0.05 and 95% confidence intervals (CI).
Results: In 665 patients with stages B or C tumour, 35 (5.3%; CI 3.7-7.2%) had tumour at a free serosal surface. These comprised 6/332 (1.8%; CI 0.8-3.7%) patients with stage B tumour and 29/333 (8.7%; CI 6.1-12.2%) with stage C tumour. After adjustment for other relevant variables, involvement of a free serosal surface was significantly associated with pelvic recurrence [hazard ratio (HR) 2.7; CI 1.3-5.5] and diminished survival (HR 1.6; CI 1.1-2.4) but not with systemic (only) recurrence.
Conclusion: This study has confirmed that direct tumour spread to a free serosal surface independently predicts pelvic recurrence and diminished survival after resection of clinicopathological stage B and C rectal cancer. This feature should always be sought by the pathologist and reported when present, and noted by the surgeon and oncologist. Serosal involvement should be evaluated further for its utility in selecting patients for adjuvant therapy.