Abstract P6-18-19: Real-world survival of heavily pretreated patients with refractory HR+, HER2- metastatic breast cancer receiving single-agent chemotherapy - A comparison with MONARCH 1

Background In MONARCH 1 (NCT02102490), abemaciclib demonstrated promising single-agent activity and tolerability in a population of heavily pretreated women with refractory HR+, HER2- metastatic breast cancer (MBC). 1 Confirmed objective response rate (ORR) was 19.7% (95% CI: 13.3, 27.5) and at 18 months minimum follow-up median overall survival (OS) was 22.3 months. Due to the single-arm trial design of MONARCH 1, there is a need to view these results in clinical context relative to available treatment options. This study compared the OS results of abemaciclib in MONARCH 1 vs that in a real-world single-agent chemotherapy cohort with similar patient and disease characteristics. Methods MONARCH 1 study design and key eligibility criteria were previously described. 1 The real-world cohort was based on Flatiron Health electronic health records-derived, nationally representative (USA-based) database comprising patient-level structured and unstructured data, curated via technology-enabled abstraction, for patients with MBC between January 1, 2011 through February 28, 2018. A real-world single-agent chemotherapy cohort was created based on the key eligibility criteria of MONARCH 1 and included patients diagnosed with HR+, HER2- MBC who received single-agent chemotherapy (eribulin, capecitabine, gemcitabine, or vinorelbine) following 1-2 prior chemotherapy regimens in the metastatic setting, had an ECOG PS of 0-1, and no prior CDK4 & 6 therapy. The index date was the start of the eligible single-agent chemotherapy, and patients were followed from the index date until date of death, loss to follow-up, or end of the database, whichever occurred earlier. OS results were adjusted using 2 methods (Mahalanobis distance matching and entropy balancing with bootstrapping) to account for baseline demographic and clinical differences between the real-world and trial cohorts. Results A real-world cohort (n=281) with eligibility criteria similar to the MONARCH 1 population (n=132) was identified. A subsequent matching based on Mahalanobis distance was performed to match MONARCH 1 population (n=108) with the real-world cohort (n=108). The matched cohorts demonstrated similar patient and disease characteristics. Median OS was 22.3 months in the abemaciclib arm vs 13.6 months in the matched cohort with an estimated hazard ratio (HR) of 0.54 (95% CI: 0.37, 0.77). Results of a sensitivity analysis performed using entropy balancing were consistent with an adjusted median OS of 12.7 months in the real-world cohort (n=281)with HR of 0.57 (95% CI from bootstrapping: 0.44, 0.78). Conclusion Methodological advances to adjust for potential biases, and improvements in data quality, have evolved enabling the ability to leverage a real-world cohort as an external comparator arm. This study demonstrates the ability to create a real-world chemotherapy cohort suitable to serve as a comparator for MONARCH 1. These exploratory results suggest a survival advantage and adequately place the clinical benefit of abemaciclib monotherapy in clinical context. References Dickler et al, CCR 2017 Citation Format: Rugo H, Dieras V, Cortes J, Patt D, Wildiers H, O9Shaughnessy J, Zamora E, Yardley DY, Carter GC, Sheffield KM, Li L, Andre VA, Derbyshire RE, Li XI, Frenzel M, Huang Y-J, Dickler MN, Tolaney SM. Real-world survival of heavily pretreated patients with refractory HR+, HER2- metastatic breast cancer receiving single-agent chemotherapy - A comparison with MONARCH 1 [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P6-18-19.