Predictive factors for early mortality after initiation of regorafenib or trifluridine/tipiracil in refractory metastatic colorectal cancer.

Abstract

3560 Background: Regorafenib (REG) and trifluridine/tipiracil (FTD/TPI) have been recognized as standard treatments for patients (pts) with refractory metastatic colorectal cancer (mCRC). Because these drugs have limits on efficacy benefit for some pts, we are necessary to select pts who may be better not to receive REG or FTD/TPI. However, no reports are available on how to predict pts with early mortality after initiation of these drugs. Methods: We retrospectively evaluated pts with mCRC who were registered in a multicenter observational study (the REGOTAS study). The main inclusion criteria were ECOG PS of 0–2, refractory or intolerant to fluoropyrimidines, oxaliplatin, irinotecan, and anti-VEGF and anti-EGFR therapy (if KRAS wild type), and no prior use of REG or FTD/TPI. Predictive factors for early mortality (≤ 12 weeks from initiation of REG or FTD/TPI) were evaluated by multivariate analysis for survival with all variables with P values of <0.05 from the univariate analysis, using the Cox proportional hazards model. In this analysis, the pts who lived at first 15 weeks were defined as censored case. Results: A total of 523 pts (REG, 212; FTD/TPI, 311) were eligible. Predictive factors for early mortality were without primary tumor resection [adjusted hazard ratio (aHR), 1.56; p = 0.02], the low level of albumin (aHR, 2.31; p < 0.0001), the high level of CRP (aHR, 2.31; p < 0.0001), and short time from diagnosis of mCRC (aHR, 1.77; p = 0.002) by multivariate analysis. The pts harboring all these poor factors were classified into the high (H) risk of early mortality group. Two groups of pts with H and non-H were identified, with 12-week mortality rate of 39 and 14%, with 14-week mortality rate of 70 and 18%, and with median survival time (MST) of 2.9 and 7.8 months, respectively (HR of H/non-H, 4.02; p value < 0.0001). In pts treated with REG, H and non-H groups had 12-week mortality rate of 38 and 16%, 14-week mortality rate of 79 and 18%, and MST of 2.8 and 7.8 months, respectively. In pts treated with FTD/TPI, H and non-H groups had 12-week mortality rate of 41 and 13%, 14-week mortality rate of 63 and 18%, and MST of 3.2 and 7.7 months, respectively. Conclusions: Our predictive model for early mortality after initiation of REG or FTD/TPI might be useful for selecting pts who should not receive either these drugs.