BACKGROUND: Cyclin-dependent kinase 4/6 inhibitor in combination with endocrine therapy(CDK4/6i) is the current standard of care in fi rst line (1L) treatment for hormone receptor–positive and human epidermal growth factor receptor 2-negative (HR+/HER2–) metastaticbreast cancer (mBC). Comparative eff ectiveness evidence assessing overall survival (OS)outcomes are limited, particularly in older pts. Our primary objective was to compare OS inMedicare pts treated with 1L PAL + AI versus AI alone in the current SEER dataset.
METHODS: We conducted a retrospective analysis of HR+/HER2–mBC pts using the SEER-Medicare database. Medicare-enrolled pts aged ≥65 years diagnosed with de novoHR+/HER2– mBC between February 1, 2015, and December 31, 2019 were included. Ptswere followed from the start of 1L therapy line until death, Medicare disenrollment,enrollment in a health maintenance organization plan, or the end of available study follow-up (December 31, 2020), whichever came fi rst. Stabilized inverse probability of treatmentweighting (sIPTW) was the primary method used to balance patient characteristics atbaseline, with propensity score matching (PSM) and multivariable CPH regression assensitivity analyses. Median OS and the hazard ratio (HR) were estimated using theweighted Kaplan-Meier (KM) method and the weighted Cox proportional hazard (CPH)regression model, respectively.
RESULTS: Of 779 included pts, 296 received PAL+AI and 483 received AI alone. Medianfollow-up was 23.9 months for the PAL+AI arm and 18.2 months for the AI alone arm. AftersIPTW adjustment, median OS was 37.6 months (95% CI = 34.8-42.0) for PAL+AI vs 25.5months (95% CI = 22.0-28.9) for AI alone (HR = 0.73, 95% CI = 0.59-0.91). Results from thePSM and CPH regression analyses were consistent with the primary analysis. Tablepresents key patient characteristics.
CONCLUSIONS: This comparative eff ectiveness study showed that PAL+AI was associatedwith signifi cantly prolonged OS vs AI alone in the 1L treatment of older pts with de novoHR+/HER2- mBC treated in US clinical practice using the SEER-Medicare database. Ourresults add to the body of evidence on the survival benefi t of PAL+AI in this population.
