Effects of tumor location and mismatch repair on clinicopathological features and survival for non-metastatic colon cancer: A retrospective, single center, cohort study

Abstract

Objective: To analyze the differences in clinicopathological features of colon cancers and survival between patients with right- versus left-sided colon cancers. Methods: This was a retrospective cohort study. Information on patients with colon cancer from January 2016 to August 2020 was collected from the prospective registry database at Peking Union Medical College Hospital . Primary tumors located in the cecum, ascending colon, and proximal two-thirds of the transverse colon were defined as right-sided colon cancers (RCCs), whereas primary tumors located in the distal third of the transverse colon, descending colon, or sigmoid colon were defined as left-sided colon cancers (LCCs). Clinicopathological features were compared using the χ2 test or Mann-Whitney U test. Survival was estimated by Kaplan-Meier curves and the log-rank test. Factors that differed significantly between the two groups were identified by multivariate survival analyses performed with the Cox proportional hazards function. One propensity score matching was performed to eliminate the effects of confounding factors. Results: The study cohort comprised 856 patients, with TNM Stage I disease, 391 (45.7%) with Stage II, and 336 (39.3%) with Stage III, including 442 (51.6%) with LCC and 414 (48.4%) with RCC and 129 (15.1%). Defective mismatch repair (dMMR) was identified in 139 patients (16.2%). Compared with RCC, the proportion of men (274/442 [62.0%] vs. 224/414 [54.1%], χ2=5.462, P=0.019), body mass index (24.2 [21.9, 26.6] kg/m2 vs. 23.2 [21.3, 25.5] kg/m2, U=78,789.0, P<0.001), and well/moderately differentiated cancer (412/442 [93.2%] vs. 344/414 [83.1%], χ2=22.266, P<0.001) were higher in the LCC than the RCC group. In contrast, the proportion of dMMR (40/442 [9.0%] vs. 99/414 [23.9%], χ2=34.721, P<0.001) and combined vascular invasion (106/442[24.0%] vs. 125/414[30.2%], χ2=4.186, P=0.041) were lower in the LCC than RCC group. The median follow-up time for all patients was 48 (range 33, 59) months. The log-rank test revealed no significant differences in disease-free survival (DFS) (P=0.668) or overall survival (OS) (P=0.828) between patients with LCC versus RCC. Cox proportional hazards model showed that dMMR was significantly associated with a longer DFS (HR=0.419, 95%CI: 0.204‒0.862, P=0.018), whereas a higher proportion of T3-4 (HR=2.178, 95%CI: 1.089‒4.359, P=0.028), N+ (HR=2.126, 95%CI: 1.443‒3.133, P<0.001), and perineural invasion (HR=1.835, 95%CI: 1.115‒3.020, P=0.017) were associated with poor DFS. Tumor location was not associated with DFS or OS (all P>0.05). Subsequent analysis showed that RCC patients with dMMR had longer DFS than did RCC patients with pMMR (HR=0.338, 95%CI: 0.146‒0.786, P=0.012). However, the difference in OS between the two groups was not statistically significant (HR=0.340, 95%CI:0.103‒1.119, P=0.076). After propensity score matching for independent risk factors for DFS, the log-rank test revealed no significant differences in DFS (P=0.343) or OS (P=0.658) between patients with LCC versus RCC, whereas patient with dMMR had better DFS (P=0.047) and OS (P=0.040) than did patients with pMMR. Conclusions: Tumor location is associated with differences in clinicopathological features; however, this has no impact on survival. dMMR status is significantly associated with longer survival: this association may be stronger in RCC patients.