Cadence of circulating tumor DNA (ctDNA) testing for molecular surveillance in early-stage breast cancer (eBC).

Abstract

560 Background: Post-surgical detection of molecular residual disease (MRD) via plasma ctDNA testing is strongly associated with recurrence of eBC. Our prior real-world data suggests that adjuvant MRD testing impacts clinical care for most patients (pts) with positive ctDNA (ctDNA+) results. Ongoing trials are studying whether adjuvant ctDNA testing improves eBC outcomes. While serial testing is known to improve MRD detection rates, the optimal cadence of testing is unknown. Here, we investigated the role of timing, cadence, and quantitative results of real-world ctDNA testing on detection of clinical recurrence of eBC. Methods: We identified eBC pts with available recurrence-free survival data who had adjuvant plasma MRD testing via a clinically validated, personalized, tumor-informed mPCR-NGS ctDNA test (Signatera, Natera, Inc.). All tests were ordered in the United States in the real-world clinical setting between 2019-2024; clinical records were reviewed. The cumulative incidence of clinical recurrence after each ctDNA test was used to calculate negative predictive value (NPV) for recurrence within 3, 6, 12, 18, 24, and 30 months (mo) post-test. ctDNA levels at different timepoints prior to recurrence were analyzed separately by ER and HER2 status. Results: For 819 pts with stage I-III eBC (ER+/HER2-: 249, HER2+: 68, triple-negative [TNBC]: 502), there were 4689 total plasma samples obtained in the adjuvant setting (median 5.7 time points per patient). Median time of first adjuvant ctDNA testing was 7 mo after surgery (range: 0.1-214.9). Median follow-up was 18.1 mo (range: 0.7-239.4). For pts with ER+/HER2- tumors, median time to first test was 14.6 mo (range: 0.3-214.9) versus 16.1 mo (range: 0.3-151.8) in HER2+ and 7.1 mo (range: 0.1-178.8) in TNBC. Among pts with multiple adjuvant ctDNA tests, median interval between tests was 2.8 mo (consistent across subtypes). For ER+ and TNBC tumors NPV (95%CI) gradually decreased over time from 99.5% (98.8-99.8) and 99.7% (99.2-99.9) at 3 mo to 97.7% (96.7-98.4) and 97.8% (96.6-98.5) at 30 mo, respectively. Among ctDNA+ TNBC pts, quantitative ctDNA levels were higher among those who recurred < 3 mo after a +ctDNA test than those with recurrence in 3-6 mo; median ctDNA levels were lowest in pts who recurred > 6 mos after a ctDNA+ test (median [range], 2.7 [0.03-1089.6] vs 2.0 [0.1-15.3] vs 0.3 [0.3-1.7] MTM/mL, p = 0.0041 and p = 0.0071, respectively). This trend was less pronounced in ER+HER2- disease between pts who relapsed < 6 mo, 6-9 mo, and > 9 mo after a ctDNA+ test (median [range], 11.6 [0.26-299.4] vs 4.5 [0.2-27.0] vs 3.8 [2.3-124.5] MTM/mL, p = 0.5, p = 0.8). Conclusions: The data from this real-world analysis of tumor-informed ctDNA testing in pts with eBC during surveillance demonstrate a high NPV for both ER+ and TNBC disease. These data guide future prospective ctDNA-guided studies aimed at therapeutic interception to improve clinical outcomes.