Background/Objectives: Postoperative circulating tumour DNA (ctDNA) can identify patients at risk of breast cancer recurrence. We estimated the accuracy of serial surveillance in hormone receptor-positive, HER2-negative early breast cancer and examined its implications for intervention trials. Methods: We considered randomised trials and prospective or retrospective cohorts of surgically treated stage I–III disease with linked postoperative ctDNA and recurrence results for the target subtype. The quantitative synthesis required a serial postoperative testing programme and a recoverable two-by-two table. We searched PubMed, checked citations, and reviewed primary publications, supplements and public datasets for evidence available through 6 November 2026. Sensitivity and cutoff-dependent specificity were pooled using separate binomial-normal random-effects models. Results: Nine studies contributed 1,294 patients and 137 documented recurrences, including the TRAK-ER surveillance cohort. Pooled sensitivity for documented recurrence was 86.1% (95% CI 79.7–93.6), and specificity relative to recurrence status at each study cutoff was 93.0% (95% CI 87.8–96.5). Across seven studies with subtype-specific data, the median of study median lead times was 13.0 months (range 7.5–21.0). Accuracy estimates reflected recurrence status at the study-specific reporting cutoffs. Conclusions: Serial ctDNA detected most documented recurrences in the included cohorts. Estimates reflect the assays, sampling schedules and follow-up of selected populations. A positive ctDNA result with negative contemporaneous imaging may identify patients for assessment of eligibility for an interception trial. The time available to start treatment and the clinical benefit of doing so require prospective evaluation.