Real-world comparison of triplet- versus doublet-chemotherapy with bevacizumab in patients with newly diagnosed metastatic colorectal cancer (mCRC).

Abstract

3574 Background: Meta-analysis of five randomized intensification trials in mCRC demonstrated an overall survival (OS) benefit of triplet chemotherapy (5FU/leucovorin, oxaliplatin, irinotecan) plus bevacizumab (Triplet/Bev) compared to doublet chemotherapy plus bevacizumab (Doublet/Bev) (HR 0.81). Guidelines recommend first-line Triplet/Bev for patients with excellent performance status and normal organ function who can withstand increased toxicity, but such patients are less represented in the real world. In this retrospective cohort study, we examine the real-world effectiveness of Triplet/Bev compared to Doublet/Bev in the U.S. Methods: We used the nationwide Flatiron Health electronic health record-derived database, comprising de-identified patient-level structured and unstructured data curated via technology-enabled abstraction from ~ 280 cancer clinics. Included were mCRC patients treated with first-line Triplet/Bev or Doublet/Bev between 10/23/14 and 10/31/22. OS by treatment was assessed using the Kaplan Meier method and adjustment for confounding variables was performed using Cox proportional hazards modeling with stabilized inverse probability of treatment weighting (IPTW). Pre-specified confounding variables included age, gender, race/ethnicity, year of metastatic diagnosis, health insurance, practice setting, KRAS/NRAS/BRAF mutation status, MMR/MSI status, synchronous/metachronous metastases, primary tumor sidedness, CEA, renal dysfunction, hepatic dysfunction, and ECOG performance status. Missing data were imputed using multiple imputation with chained equations. Results: Among 10,016 patients, 391 (3.9%) received Triplet/Bev and 9,625 (96.1%) received Doublet/Bev. Patients who received Triplet/Bev were younger (median age 52 vs 63 years) and were more likely to be treated at academic practices (27% vs 8%), to have synchronous metastatic disease (83% vs 65%), and to have KRAS/NRAS/BRAF wild-type disease (30% vs 17%). In the univariate analysis, median OS was 30.0 months (95% CI 24.7 – 34.7) in the Triplet/Bev cohort versus 23.4 months (95% CI 22.8 – 24.1) in the Doublet/Bev cohort (HR 0.73; 95% CI 0.62 – 0.85; p < 0.001). In Cox proportional hazards modeling with IPTW, there was no significant difference in hazard of death between Triplet/Bev and Doublet/Bev (HR 0.92; 95% CI 0.75 – 1.12; p = 0.4). Conclusions: In this large, national, real-world population of patients with mCRC, first-line treatment with Triplet/Bev was not associated with improved survival compared to Doublet/Bev after accounting for differences in baseline patient characteristics. The benefit to Triplet/Bev observed in randomized trials does not translate well to a more heterogeneous, real-world population. Sub-group analyses including by age, tumor sidedness, KRAS/NRAS/BRAF status, and baseline CEA level will be performed.