In HER2-negative metastatic breast cancer (MBC), early assessment of treatment efficacy is crucial to optimize therapeutic strategies. Quantification of circulating tumor DNA (ctDNA) allows evaluation of tumor burden and tumor cell death. However, the prognostic value of early ctDNA kinetics under treatment remains to be established in this setting. The aim of this study was to assess the association between the decline in mutant allele fraction (MAF) of ctDNA and progression-free survival (PFS) in patients receiving second-line chemotherapy.

This study was based on the pre-screening phase of the MONDRIAN trial (NCT04720729), which included patients with HER2-negative MBC initiating second-line chemotherapy. ctDNA was quantified using droplet digital PCR (ddPCR) on blood samples collected at baseline (day 0) and day 15. Changes in MAF (%) over time were modeled using a linear mixed-effects model to characterize individual decay trajectories. Low MAF values were left-censored by interval to account for limits of detection and quantification. The association between the individual ctDNA decay slope and PFS was analyzed using a Cox proportional hazards model adjusted for prognostic variables, with a 30-day landmark analysis to avoid immortal time bias.

Among the 104 patients included in MONDRIAN at the time of data extraction (July 2024), 89 had complete follow-up data available for survival analysis. Modeling of MAF decline using linear mixed-effects models revealed substantial heterogeneity in individual ctDNA trajectories. The 30-day landmark analysis showed that a faster decline in MAF was associated with improved PFS in univariable Cox analysis (HR = 0.71 [0.52-0.97], p = 0.031). After adjustment for prognostic factors (ECOG performance status, histological subtype, triple-negative status), a higher rate of MAF decline remained significantly associated with better PFS (HR = 0.69 [0.50-0.94], p = 0.020). Sensitivity analyses using different landmark durations or dichotomizing MAF decline according to the median (rapid vs. slow decline) yielded similar trends.

Early changes in ctDNA decrease speed, assessed through MAF and analyzed using mixed-effects models, appear to be a promising biomarker for predicting PFS in patients with HER2-negative metastatic breast cancer receiving second-line chemotherapy. This readily accessible biomarker could help identify patients at higher risk of early progression, enabling adaptation of imaging follow-up schedules and/or consideration of earlier therapeutic modifications