Iseult Browne, Javier Pascual, Rosalind Cutts, Belinda Kingston, Sarah Hrebien, Lucy Kilburn, Alex Pearson, Laura Moretti, Andrew Wardley, Iain R. Macpherson, Richard D. Baird, Rebecca Roylance, Miguel Gil‐Gil, Ángel Guerrero‐Zotano, Isaac García-Murillas, Judith M. Bliss, Alistair Ring, Miguel Martín, Nicholas C. Turner
PURPOSE: ctDNA dynamic levels may identify patients who will respond to therapy. We assessed ctDNA baseline levels and on-treatment dynamics in patients with advanced breast cancer on the plasmaMATCH trial with mutation-targeted therapies (cohorts A-D) and triple-negative breast cancer on olaparib and ceralasertib combination (cohort E). EXPERIMENTAL DESIGN: Blood samples were collected at baseline [cycle 1 day 1 (C1D1)] and before treatment on cycle 2 day 1 (C2D1). Samples were sequenced using error-corrected targeted panels (Guardant360/GuardantOMNI). Circulating DNA ratio was calculated as the ratio of C2D1/C1D1 circulating DNA ratio, and baseline ctDNA levels were associated with progression-free survival (PFS) and confirmed objective response rates (ORR). RESULTS: A total of 167 patients had assessable C1D1-C2D1 ctDNA results. There was a strong association between baseline ctDNA levels and response in cohort E; low baseline levels were associated with longer PFS (HR, 0.33; P = 0.001) and higher ORR (40% vs. 9.7%; P = 0.02). In cohorts A to D, there was a weaker association with PFS (HR, 0.60; P = 0.03) and ORR (15.2% vs. 5.7%; P = 0.17). Associations of baseline ctDNA level and ORR were validated in the independent PEARL study. For on-treatment dynamics, suppression of ctDNA below median was predictive in cohorts A to D (HR, 0.47; P = 0.001) but not in cohort E (HR, 1.02; P = 0.94). Undetectable ctDNA levels at C2D1 were associated with good outcomes in both cohorts: in cohort E with improved PFS (HR, 0.25; P = 0.01) and improved ORR (86% vs. 11%; P = 0.01). Six of seven patients with undetectable on-treatment ctDNA were BRCA1/BRCA2/PALB2 wild type. CONCLUSIONS: Baseline low ctDNA levels predict response to targeted therapy, potentially suggesting shared mechanisms between high ctDNA release and resistance to therapy. Both baseline ctDNA levels and on-treatment dynamics are a promising surrogate endpoint for drug development, with clearance of ctDNA being a robust cross-therapy surrogate for outcomes.