Anatomic location of colorectal cancer presents a new paradigm for its prognosis in African American patients.

Abstract

Among all racial groups in the U.S., African Americans (AA) have the highest incidence of and mortality from colorectal cancer (CRC). Although socioeconomic factors, as the major contributors to racial disparity of CRC, have been widely investigated, there is a dearth of information germane to understanding its biological basis. To better elucidate the clinicopathologic features we extracted demographic, clinical, pathologic and molecular features of 500 consecutive cases of CRC diagnosed at our institution which has an AA-predominant patient population (75% of all patients). We compared data from our AA patients with those of white patients both from our institution and from SEER and the published literature for meaningful comparison. AA patients were more likely to be at an advanced disease stage (25.9% vs. 20.8%, p = 0.041), have low grade tumors (89.2% vs. 77.5%, p<0.001) in cecum (18.7% vs. 16.2%, p<0.001) and <60-years-old than white patients (31.8% vs. 26.3%, p = 0.015). The frequency of KRAS mutation was higher in AA patients than in white patients (56.8% vs. 20.7%, p<0.001). Amongst subtypes of KRAS tested in CRC, codon 12 mutation is more common in AA than white patients (85.2% vs. 68.9%, p = 0.020). Compared with other racial groups, we found AA patients to have worse disease-free survival (HR = 3.682, p = 0.035). Also, AA patients with CRC in distal (sigmoid and rectum) or proximal (cecum) colon have worse overall survival than those with CRC in middle colon (HR = 2.926, p = 0.014), a finding not observed in white patients. In both racial groups, advanced stage, perforation, and hypertension were independent prognostic factors for overall survival (p<0.05). Similarly, low body-mass index at presentation, mucinous adenocarcinoma, lymphovascular invasion, perineural invasion and KRAS mutations were independent factors significantly associated with poor disease-free survival. Collectively, our data provide new insights into the roles of clinicopathologic features, especially anatomic distribution, in predicting outcomes of CRC in AA population.