Clinicopathologic and epidemiological characteristics of prognostic factors in post-surgical survival of colorectal cancer patients in Jiangsu Province, China

• Post-surgical survival differences for CRC was primarily due to clinicopathologic features. • RCC patients had worse overall survival compared to LCC and ReC in several subgroups. • Epidemiological factors and treatment regimens alone could not explain survival variations. • Pairwise interactions between some subgroups were significant. Poor survival among colorectal cancer (CRC) patients has been widely associated with clinico-epidemiological features and treatment regimen. In Jiangsu (China), however, it is not known which one of the prognostic factors explains the survival disparities among patients with CRC. This prospective study using 1078 patients (stages I-IV) that underwent surgery at Jiangsu Hospital, explored the relevant factors affecting the prognoses of right-side colon cancer (RCC), left-side colon cancer (LCC) and rectal cancer (ReC) patients. Of these cases, 234 (21.7%), 241 (22.4%) and 603 (55.9%) were found to have RCC, LCC and ReC respectively. Compared to LCC, RCC exhibited a greater proportion of older patients, poorly differentiated carcinomas, higher T-stage and higher TNM-stage. The overall survival (OS) for RCC was 60 vs .78 or 77 months for LCC or ReC respectively (P = 0.030). There were no significant differences in OS between LCC and ReC across the subgroups (P = 0.633). In multivariate analysis, RCC patients had age (>60 vs . ≤60 years, HR = 1.529, P = 0.019), N-stage (N1 vs . N0, HR = 4.056, P = 0.012) and M-stage (M1 vs . M0, HR = 3.442, P < 0.0001) as independent prognostic factors, whereas smoking status was found to be a predictor of mortality (smoker vs . nonsmoker, HR = 2.343, P = 0.017) for LCC. In addition, age (>60 vs . ≤60 years, HR = 2.199, P < 0.0001), alcohol consumption (drinker vs . nondrinker, HR = 0.510, P = 0.034), tumor grade (Poor vs . well/moderate, HR = 2.759, P = 0.031) and T-stage (T3-4 vs . T1-2, HR = 1.742, P < 0.0001) were found to be predictors of mortality for ReC. There were significant pairwise interactions across subgroups. Furthermore, significant differences were observed for LCC vs . RCC (OS, HR = 0.783, P = 0.039), but no statistically significant differences for ReC vs . RCC (P = 0.149) and LCC vs . ReC (P = 0.355). Nevertheless, significant differences remained between ReC vs . RCC for male (HR = 0.591, P = 0.009), drinker (HR = 0.396, P = 0.005), rural resident (HR = 0.437,P = 0.022), tumor grade (well/moderate, HR = 0.475, P = 0.022), T-stage (T1-2, HR = 0.362, P = 0.001), N-stage (N0, HR = 0.604, P = 0.011), M-stage(M0, HR = 0.401, P = 0.006) and TNM-stage (I-II, HR = 0.567, P = 0.005). Statistically significant differences were observed for LCC vs . RCC for gender (female, HR = 0.495, P = 0.003) and T-stage (T1-2, HR = 0.417, P = 0.010) as well as for LCC vs . ReC in patients with smoking habits (HR = 1.951, P = 0.002) and M-stage (M0, HR = 2.291, P = 0.003). These findings suggest that the variations in CRC post-surgical survival in China may be primarily explained with the clinicopathologic features and epidemiological characteristic of the patients. Patients with RCC had significantly worse OS compared to both LCC and ReC in several subgroups.