Effectiveness of palbociclib with aromatase inhibitors for the treatment of advanced breast cancer in an exposure retrospective cohort study: implications for clinical practice

Abstract

Abstract Background New drugs for locally advanced or metastatic breast cancer have led to clinical benefits, aside with increasing costs to healthcare systems. The current financing model for health technology assessment (HTA) privileges real-world data. As part of the ongoing HTA, this study aimed to evaluate the effectiveness of palbociclib with aromatase inhibitors (AI) and compare it with the efficacy reported in PALOMA-2. Methods A population-based retrospective exposure cohort study was conducted including all patients initiating treatment in Portugal with palbociclib under early access use and registered in the National Oncology Registry. The primary outcome was progression free survival (PFS). Secondary outcomes considered included time to palbociclib failure (TPF), overall survival (OS), time to next treatment (TTNT), and proportion of patients discontinuing treatment due to  adverse events (AEs). The Kaplan–Meier method was used and median, 1- and 2-year survival rates were computed, with two-sided 95% confidence intervals (95%CI). STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines for reporting observational studies were used. Results There were 131 patients included. Median follow-up was 28.3 months (IQR: 22.7–35.2) and median duration of treatment was 17.5 months (IQR: 7.8–29.1). Median PFS was 19.5 months (95%CI 14.2–24.2), corresponding to a 1-year PFS rate of 67.9% (95%CI 59.2–75.2) and a 2-year PFS rate of 42.0% (95%CI 33.5–50.3). Sensitivity analysis showed median PFS would increase slightly when excluding those not initiating treatment with the recommended dose, raising to 19.8 months (95%CI 14.4–28.9). By considering only patients meeting PALOMA-2 criteria, we could observe a major difference in treatment outcomes, with a mean PFS of 28.8 months (95%CI 19.4–36.0). TPF was 19.8 months (95%CI 14.2–24.9). Median OS was not reached. Median TTNT was 22.5 months (95%CI 18.0–29.8). A total of 14 patients discontinued palbociclib because of AEs (10.7%). Conclusions Data suggest palbociclib with AI to have an effectiveness of 28.8 months, when used in patients with overlapping characteristics to those used in PALOMA-2. However, when used outside of these eligibility criteria, namely in patients with less favorable prognosis (e.g., presence of visceral disease), the benefits are inferior, even though still favorable.