Abstract 250P
Background
Overall survival (OS) is the gold standard endpoint to assess treatment efficacy in cancer clinical trials. In metastatic breast cancer (mBC), progression-free survival (PFS) is recognized as alternative endpoint. Evidence remains scarce regarding the degree of association between PFS and OS. Our study aimed to describe the individual-level association between PFS and OS according to first-line (1L) treatment in women with mBC managed in real-world setting for each BC subtype.
Methods
We extracted data from the ESME mBC database which gathers deidentified data from consecutive patients managed in 18 French Comprehensive cancer centers (FCCCs). Adult women diagnosed with a mBC between 2008 and 2017 were included. Endpoints were described using the Kaplan-Meier method and individual-level associations were estimated using the Spearman’s correlation coefficient. Analyses were conducted by tumor subtypes and for elderly populations (>65 to >75 years).
Results
20,033 women were eligible including 5136 patients older than 70 years. For the entire cohort, median age and median follow-up duration were 60.0 years and 62.3 months (95% CI 58.4- 63.6) respectively. Median PFS ranged from 6.0 months (95% CI 5.8-6.2) for HR-/HER2- subtype to 13.3 months (36%CI 12.7-14.3) for HR+/HER2+ subtype. Correlation coefficients were highly variable across subtypes and 1L treatment (from 0.33 for HR+/HER2+ subtype to 0.81 for HR-/HER2- subtype, for L1 chemotherapy). Among the 1804 patients with HR-/HER2- mBC, coefficients were less variable (from 0.73 to 0.81). For HR+/ HER2+ mBC patients treated with chemotherapy or endocrine therapy, the individual-level associations were moderate to strong (coefficients ranging from 0.33 to 0.43).
Conclusions
We reported the individual-level associations between PFS and OS using the largest cohort of women diagnosed with mBC in France. Value of RWD was investigated in the search of candidate endpoint for surrogacy to OS. Those results would be also useful to scientific community when designing one-single arm study combined to RWD to build synthetic reference for underrepresented patients such as elderly patients.
Clinical trial identification
NCT03275311.
Editorial acknowledgement
Legal entity responsible for the study
UNICANCER.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.