Analysis of the efficacy and safety of regorafenib <i>versus</i> chemotherapy retreatment in 3<sup>rd</sup> line treatment of metastatic colorectal cancer

Introduction. Therapeutic options for 3rd-line treatment of metastatic colorectal cancer (mCRC) include chemotherapy retreatment with previously used regimens and regorafenib. Due to the lack of phase III prospective studies and conflicting retrospective results, we performed a comparative analysis of the efficacy and safety of chemotherapy ± targeted therapy retreatment versus regorafenib in a large patient cohort. Materials and methods. The study included mCRC patients from 4 oncology centers in the Russian Federation treated between 2010 and 2021 who received 3rd-line therapy either with chemotherapy ± targeted agents (epidermal growth factor receptor or vascular endothelial growth factor antibodies) (CTT group) or with regorafenib. The primary endpoint was 9-month overall survival (OS); secondary endpoints were progression-free survival (PFS), objective response rate (ORR) and toxicity. Univariable and multivariable analyses were conducted to identify prognostic factors. Results. A total of 334 patients were included. Median follow-up was 30.3 months (95 % confidence interval (CI) 26.0–34.5) in the CTT group and 23.8 months in the regorafenib group. The groups differed in several baseline characteristics: primary tumor resection was performed more frequently (76.1 % versus 67.6 %; p = 0.088) and ECOG 0 status was more common (25.2 % versus 14.2 %; p = 0.006) in the CTT group. Patients treated with regorafenib more frequently had elevated carcinoembryonic antigen ≥50 ng / mL (45.3 % versus 65.3 %; p 0.001) and indolent disease course ≥20 months (64.3 % versus 49.1 %; p = 0.005), while a «treatment-free interval» ≥6 months after oxaliplatin or irinotecan was observed less frequently (68.1 % versus 43.3 %; p 0.001). Chemotherapy retreatment demonstrated a statistically significant advantage in median OS (18.1 months versus 10.2 months; hazard ratio (HR) 0.51; 95 % CI 0.39–0.67; p 0.0001) and PFS (6.0 month versus 2.8 months; HR 0.54; 95 % CI 0.43–0.67; p 0.0001). ORR was 6.1 % in the CTT group compared to 0.6 % in the regorafenib group. In multivariable analysis, unfavorable prognostic factors were worse ECOG status (ECOG 1: HR 1.77; p = 0.002; ECOG 2: HR 2.56; p = 0.001), right-sided primary tumor location (HR 1.98; p 0.001), presence of ≥2 liver metastases (HR 1.52; p = 0.021), absence of radical resection of the primary tumor (HR 1.91, 95 % CI 1.32–2.78; p = 0.001), shorter interval between 1st-line and 3rd-line therapy (a longer course 20 months correlated with better prognosis: HR 1.38, 95 % CI 1.04–1.85; p = 0.028), and carcinoembryonic antigen ≥50 ng / mL (HR 1.38, 95 % CI 1.04–1.82; p = 0.026). Based on these factors, a prognostic model was created that identifies favorable (median OS 19.4 months), intermediate (median OS 13.0 months), and poor (median OS 6.6 months) risk groups. The C-index was 0.698 (standard error 0.02). Conclusion. Chemotherapy retreatment showed superiority over regorafenib in 3rd-line treatment of mCRC. Its higher efficacy was confirmed across all prognostic subgroups. Predictive analysis did not identify any subgroup with advantage from regorafenib therapy.