Monitoring of therapeutic efficacy to CDK4/6 inhibitors and early detection of metastatic relapse in breast cancer by ultra-deep sequencing of plasma cell-free DNA.

Abstract

e15544 Background: CDK4/6 inhibition substantially improves progression-free survival for women with advanced estrogen receptor-positive breast cancer. Despite this, most patients experience acquired resistance. The analyses of cancer specific genomic aberrations in plasma cell free DNA, or commonly referred to as ‘liquid biopsy’, has generated immense interest. The role of liquid biopsy enables monitoring of therapeutic efficacy to detect relapse in a minimally invasive manner. It would drastically impact risk stratification, treatment selection and patient clinical outcome in advanced stage breast cancer. Methods: We have recruited consecutive patients with metastatic breast cancer who received CDK4/6 inhibitor treatment in combination with endocrine therapy between April 2018 and December 2019. Somatic mutations identified by the Oncomine Pan-Cancer cell free assay were used to assess the mutation profiles of circulating tumor DNA (ctDNA) by ultra-deep sequencing (~56,000 depth) of the plasma DNA. Plasma samples (N = 146) were collected before and during treatment (2 years) or until disease progression. Paired white blood cells (WBC) were also sequenced to exclude clonal hematopoiesis-derived mutations. Results: Twenty-eight patients were enrolled. Based on RECIST criteria, 1 patient showed complete response (CR) within 18 months of treatment, 5 patients showed partial response (PR) 6-18 months from start of treatment, 9 patients showed stable disease (SD) condition within 3-18 months and 13 patients developed progressive disease (PD) within 1-12 months. In 22 of 28 patients (79%), ctDNA was detected prior to start of treatment (12 responders; 10 PD) and used as a monitoring marker. Majority of mutations detected in pre-treatment samples were TP53, PIK3CA and ESR1. Among the 12 ctDNA positive responders, ctDNA levels of 9 patients were undetectable within 2 weeks to 6 months from start of treatment (median 3 months). ctDNA profiles were more reflective of clinical response compared to the tumor markers CEA and CA15-3. In 7 of the 10 PD patients with ctDNA monitoring markers, ctDNA showed higher sensitivity to predict disease progression compared to tumor markers that remained stagnant or below normal threshold. ctDNA also predicted PD earlier than radiological images with a median lead time of 3 months. Conclusions: ctDNA assessment for therapy monitoring and early relapse detection in advanced breast cancer is feasible and provides a basis to evaluate early therapeutic interventions.