Real-world outcomes of trifluridine/tipiracil with or without bevacizumab in refractory metastatic colorectal cancer: a multicenter cohort study from Turkey

Abstract

Trifluridine/tipiracil (FTD/TPI; TAS-102) is an established later-line treatment option for refractory metastatic colorectal cancer (mCRC), and randomized evidence supports the addition of bevacizumab. However, real-world data from Turkey are limited. We conducted a multicenter retrospective cohort study of adults with metastatic colorectal cancer (mCRC) who received FTD/TPI plus bevacizumab (combination) or FTD/TPI monotherapy in routine clinical practice between June 2021 and May 2025, with follow-up updated until September 25, 2025. Overall survival (OS) was the primary endpoint, and progression-free survival (PFS), response outcomes in radiologically evaluable patients, and safety were secondary endpoints. Survival was analyzed using Kaplan-Meier estimates and log-rank tests, with hazard ratios estimated using Cox regression. A prespecified multivariable Cox model for OS was adjusted for ECOG performance status, number of metastatic sites, liver metastasis, age group, and treatment group. Seventy-eight patients were included (combination therapy, n = 57; monotherapy, n = 21). Median OS was 8 months (95% CI, 6.17–9.83) in the combination group and 6 months (95% CI, 5.03–6.97) in the monotherapy group (log-rank p = 0.437). Median PFS was 4 months (95% CI, 2.83–5.16) versus 3 months (95% CI, 1.92–4.07) (log-rank p = 0.409). Best radiologic response was evaluable in 65 patients. In univariable Cox regression, treatment group was not significantly associated with OS or PFS. In the multivariable Cox model for OS, treatment group remained not significantly associated with OS (adjusted HR 1.251, 95% CI 0.644–2.427; p = 0.509), whereas ECOG 1 (vs. 0) and liver metastasis (yes vs. no) were associated with worse OS. Hematologic toxicity was the dominant safety signal in both groups, and no unexpected safety signals were observed. In this multicenter Turkish real-world cohort, FTD/TPI-based therapy was feasible with manageable toxicity in heavily pretreated refractory mCRC. Comparative survival estimates between combination therapy and monotherapy were not statistically significant in unadjusted or adjusted analyses and should be interpreted as exploratory in light of baseline imbalances, subgroup-size asymmetry, and residual confounding. These findings complement, rather than challenge, the comparative efficacy estimates established in randomized trials such as SUNLIGHT.