Tracking structural variants in early breast cancer using a high-sensitivity ctDNA assay predicts relapse in the post-neoadjuvant setting: the multicenter ALIENOR trial

BACKGROUND: Identification of breast cancer patients at high risk of relapse after neoadjuvant chemotherapy (NAC) is of high unmet clinical need. The multicenter ALIENOR study aimed to identify patients with molecular residual disease (MRD) using an ultrasensitive ctDNA assay during follow-up after NAC. PATIENTS AND METHODS: Patients whose tumor did not achieve pathologically documented complete response after NAC were enrolled into a 5-year liquid biopsy surveillance program. The primary objective was to assess relapse-free interval (RFI) in patients with or without circulating tumor DNA (ctDNA) detection during adjuvant follow-up. Whole genome sequencing was performed on primary tumor samples, and personalized multiplex digital PCR assays were used to track structural variants (SVs) in ctDNA samples. RESULTS: Eighty-three patients underwent successful SV-personalized fingerprint design (473 plasma samples). Median follow-up time from the date of surgery was 60.6 months (range 7.2-66.3 months), and 27 clinical recurrences (32.5%) were observed. ctDNA was detected in 34 patients. In 44.1% (15 of 34) of these patients, ctDNA was in the ultrasensitive range (<100 parts per million) at the earliest positive sample. Patients with ctDNA detection had a shorter RFI [hazard ratio (HR) 29.2, 95% confidence interval (CI) 6.9-124.0, P < 0.0001]. In multivariate analysis, ctDNA was the most significant prognostic factor for clinical recurrence (HR 69.7, 95% CI 13.3-365.1, P < 0.0001). The median lead times from ctDNA detection to clinical relapse were 13.9 months (range 0.3-60.6) in the entire cohort, 20.9 months (range 5.5-60.6) in estrogen receptor-positive/human epidermal growth receptor 2 (HER2)-negative subtype, 14.2 months (range 3.4-34.4) in HER2-positive subtype, and 4.3 months (range 0.3-19.0) in triple-negative subtype. Of 27 patients who experienced locoregional or distant relapse, 25 had detectable ctDNA (sensitivity 92.6%). Among 34 ctDNA-positive patients, 9 had not relapsed at last follow-up: 6 were positive only at the initial time point before adjuvant therapy, 1 had multiple positive samples without relapse at last follow-up, and 2 were positive at the last visit. The 5-year RFI of patients with a negative postsurgical plasma sample was 72.6% (95% CI 60.1% to 81.6%). CONCLUSION: This study supports the clinical validity of an ultrasensitive ctDNA assay for longitudinal surveillance of high-risk early breast cancer. ctDNA detection in the ultrasensitive range underscores the importance of high-sensitivity assays for MRD monitoring.