ABSTRACT
AIMS: Microacinar morphology has been reported as a stage-independent prognostic variable for colorectal cancer. We have undertaken a retrospective analysis to test this observation in patients with colonic as well as rectal adenocarcinoma, and to assess interobserver agreement for identifying microacinar morphology. METHODS AND RESULTS: One hundred and eighty patients with colorectal cancer, 144 colonic and 36 rectal, were assigned to either macroacinar (144) or microacinar (36) groups. There was excellent correlation between two observers examining the colorectal cancers independently (kappa statistic, kappa = 0.87). Patients with microacinar tumours had a significantly reduced median survival compared with macroacinar lesions (46 and 87 months, respectively, P = 0.022) and this morphology was significantly associated with higher Dukes' stage (P = 0.007). Microacinar morphology lacked statistical significance when examined in a multivariate analysis with other prognostic variables, both when colonic and rectal tumours were examined either separately or as a combined colorectal group, but small numbers of rectal tumours in this study make it difficult to draw useful conclusions about microacinar morphology in rectal adenocarcinoma. CONCLUSIONS: Microacinar morphology was strongly associated with poor prognosis when examined in isolation but was not found to be an independent factor when examined with other prognostic variables.
