Circulating tumor DNA (ctDNA)-based minimal residual disease (MRD) measured by Guardant Reveal in patients (pts) with HER2-positive (HER2+) metastatic breast cancer (mBC) with long-term disease control on first-line trastuzumab-pertuzumab.

Abstract

1052 Background: The CLEOPATRA and PERUSE trials established the combination of a taxane with the antiHER2 monoclonal antibodies trastuzumab and pertuzumab (HP) as the gold standard first-line treatment for HER2+ mBC. In both studies, the progression events reached a plateau after 4 years and up to 30% of pts remained long-term progression-free, hypothesizing HP maintenance can be safely discontinued. We therefore evaluated whether epigenomic-based ctDNA MRD analysis can potentially identify pts with a higher chance of permanent remission. PRE-PHENIX is a multi-center observational study that explores the prevalence of MRD measured by Guardant Reveal in HER2+ mBC pts on long-term first-line HP maintenance. Methods: A total of 40 pts with HER2+ mBC on first-line treatment with HP maintenance for a minimum of 4 years were included. Confirmation of no progressive disease by CT or PET-scan in the last 3 months previous to study entry was mandatory. Plasma samples were analysed using Guardant Reveal powered by the Guardant Infinity platform, a tissue-free epigenomic assay interrogating differentially methylated regions of DNA optimized to detect breast cancer DNA from normal cell-free DNA. Two ctDNA tests were performed on each patient within a 6 – 12-week interval. Additionally, 11 pts with confirmed disease progression on antiHER2 therapy for mBC were included as case controls. The primary objective was to establish the prevalence of positive MRD in both populations and the agreement between the two tests for the Long-Term responders. Results: Median age was 63.2 years (range 30.8 – 84.4). The median duration of first-line HP treatment was 6.9 years (range 4.2 - 11.1). At diagnosis, 26 pts (65%) presented with “de novo” mBC and 20 (50%) had visceral disease. The last radiological evaluation categorized 6 pts (15%) as having stable disease (SD), 2 pts (5%) with partial response (PR), and 32 pts (80%) with complete response (CR). Among the 11 pts with confirmed progression, 2 presented exclusive Central Nervous System (CNS) disease. Guardant Reveal identified MRD in 4 long-term responders (10%), 3 out of 6 pts (50%) with SD, 1 of 2 pts (50%) with PR, and no MRD among the 32 pts with CR. A perfect agreement was observed between the two tests (Kappa-index of 1). Ten out of the 11 pts (91%) with disease progression had MRD, including the two with exclusive CNS involvement. Conclusions: Our study demonstrates clinically significant performance of a tissue-free MRD test, Guardant Reveal, as a potential non-invasive monitoring tool to guide de-escalation strategies in pts HER2+ mBC pts with long-term remissions on HP treatment. A prospective study (PHENIX) to guide HP interruption by ctDNA monitoring is planned. This study was funded by a Fundación Contigo full grant (Spain) and Guardant Health.