Factors Predictive of Receiving Chemoradiation Therapy in Stage II and III Rectosigmoid Cancer and the Impact on Overall Survival

Abstract

The optimal treatment paradigm for rectosigmoid cancers is unknown, and it is debated whether rectosigmoid cancers should be treated as a subset of either colon or rectal cancer. We aimed to investigate the patterns of care and overall survival (OS) benefit for rectosigmoid cancer treated with chemotherapy, radiotherapy, resection, or with a multimodal approach, utilizing a large national registry of patients. Adult patients with stage II or III rectosigmoid cancer diagnosed from 2004 to 2013 were included. Logistic and Cox regression modeling was used to identify factors predictive of receipt of chemoradiation therapy (CRT) and OS. Survival analysis was performed with Kaplan Meier and log rank analysis. Propensity score matching and sensitivity analysis was performed to address selection bias and presence of potential confounding variables. A total of 7,540 patients were identified. 4,226 patients (56%) received surgery with or without chemotherapy (S ± CT), while 3,314 patients (44%) received CRT. Of those patients who received CRT, 2,174 patients received neoadjuvant CRT, 1,128 patients received adjuvant CRT, and 12 patients received both neoadjuvant and adjuvant CRT. Utilization of S ± CT increased over the study period (p = 0.04). Factors predictive of receiving CRT on multivariate analysis included: clinical stage T3 (OR: 3.28, 95% CI 2.12 to 5.09), T4 (OR: 4.67, 95% CI 2.92 to 7.45), positive nodal status of N1 (OR 1.83, 95% CI 1.60 to 2.09) and N2 (OR 1.27, 95% CI 1.04 to 1.55). Compared to S ± CT, CRT was associated with a decreased risk of death on multivariate analysis (HR: 0.68, 95% CI 0.62 to 0.75). Median survival in years for S ± CT and for CRT was 37.49 months and 41.9 months, respectively (p < 0.01). The benefit of CRT over S ± CT persisted after propensity score matching. On subset analysis, neoadjuvant CRT was associated with a decreased risk of death on univariate analysis (HR: 0.66, 95% CI 0.59 to 0.73) but not on multivariate analysis (HR: 0.92, 95% CI 0.81 to 1.09). Adjuvant CRT was associated with a decreased risk of death on multivariate analysis (HR: 0.75, 95% CI 0.64 to 0.88). The use of CRT for stage II and III rectosigmoid cancer is decreasing. Overall, when compared to S ± CT, use of CRT is associated with an OS benefit. However, we found no advantage of neoadjuvant CRT (standard rectal cancer treatment) over S ± CT when controlling for other factors like clinical stage. Adjuvant CRT is beneficial for appropriately selected patients.