23 Background: Cetuximab (CET) administered weekly (q1w) at 250 mg/m², after an initial dose of 400 mg/m², is approved in combination with chemotherapy (CT) for the treatment of (K) RAS wild-type metastatic colorectal cancer (mCRC). The use of CET 500 mg/m2 administered with CT every 2 weeks (q2w) is according to US clinical practice guidelines and observed routinely. In this study, we compared q2w vs q1w regimens on overall survival (OS) in a presumed first-line (1L) treatment subcohort and present updated data on the noninferiority of q2w vs q1w in line-agnostic (1L+) treatment using US real-world data. Methods: Using IBM MarketScan, a large US insurance claims database, we classified a cohort of mCRC patients treated between 07/2010 and 12/2016 with CET+CT as q1w or q2w based on observed infusion patterns. Absence of mCRC-related treatment claims preceding CET initiation date (defined as the index date) qualified as CET treated in 1L. A validated algorithm was used to determine patient death. Confounding was accounted for using high-dimensional propensity scoring (hdPS) with inverse probability of treatment weights. OS was compared using Cox proportional hazards regression. Imbalanced confounders after hdPS were added to the Cox model. In 1L+, noninferiority of the q2w regimen was tested with a margin hazard ratio (HR) of 1.25. However, noninferiority could not be tested in 1L due to the small sample size; a test for difference was used instead. Results: Of 2,730 CET-exposed mCRC patients (updated), 1,779 (65.2%) and 951 (34.8%) were classified in q1w and q2w groups, respectively, among which 557 (31.3%) and 316 (33.2%) received CET in 1L. The HR (95% CI) for OS of q2w vs q1w in 1L was 1.10 (0.92-1.31; crude), and 1.05 (0.86-1.29; adjusted; p for difference: 0.625). In 1L+, crude and adjusted HRs were 1.05 (0.94-1.18) and 0.94 (0.85-1.03), respectively, rejecting the inferiority hypothesis at p < 0.001. Conclusions: Only a third of patients received CET in 1L in this study. OS was statistically noninferior in q2w vs q1w in 1L+, and adjusted results in 1L suggest no differences between both treatment schedules. However, more data would be needed to formally test the noninferiority hypothesis in 1L.
Patient demographics and management landscape of metastatic colorectal cancer in the third-line setting: Real-world data in an Australian population
