ABSTRACT
BACKGROUND: The incorporation of circulating tumor DNA (ctDNA) into the management of operable breast cancer (BC) has been hampered by the heterogeneous results from different studies. We aimed to assess the prognostic value of ctDNA in patients with operable (non metastatic) BC. MATERIALS AND METHODS: A systematic search of databases (PubMed/Medline, Embase, and CENTRAL) and conference proceedings was conducted to identify studies reporting the association of ctDNA detection with disease-free survival (DFS) and overall survival (OS) in patients with stage I-III BC. Log-hazard ratios (HRs) were pooled at each timepoint of ctDNA assessment (baseline, after neoadjuvant therapy, and follow-up). ctDNA assays were classified as primary tumor-informed and non tumor-informed. RESULTS: Of the 3174 records identified, 57 studies including 5779 patients were eligible. In univariate analyses, ctDNA detection was associated with worse DFS at baseline [HR 2.98, 95% confidence interval (CI) 1.92-4.63], after neoadjuvant therapy (HR 7.69, 95% CI 4.83-12.24), and during follow-up (HR 14.04, 95% CI 7.55-26.11). Similarly, ctDNA detection at all timepoints was associated with worse OS (at baseline: HR 2.76, 95% CI 1.60-4.77; after neoadjuvant therapy: HR 2.72, 95% CI 1.44-5.14; and during follow-up: HR 9.19, 95% CI 3.26-25.90). Similar DFS and OS results were observed in multivariate analyses. Pooled HRs were numerically higher when ctDNA was detected at the end of neoadjuvant therapy or during follow-up and for primary tumor-informed assays. ctDNA detection sensitivity and specificity for BC recurrence ranged from 0.31 to 1.0 and 0.7 to 1.0, respectively. The mean lead time from ctDNA detection to overt recurrence was 10.81 months (range 0-58.9 months). CONCLUSIONS: ctDNA detection was associated with worse DFS and OS in patients with operable BC, particularly when detected after treatment and using primary tumor-informed assays. ctDNA detection has a high specificity for anticipating BC relapse.