Abstract
BACKGROUND: Early prediction of treatment response is crucial for the optimal treatment of advanced breast cancer. We aimed to explore whether monitoring early changes in plasma human epidermal growth factor receptor 2 (HER2) levels using digital PCR (dPCR) could predict the treatment response in advanced breast cancer. METHODS: This was a multicenter, prospective, noninterventional clinical study of patients with advanced breast cancer. All enrolled patients underwent blood testing to measure the HER2 levels by digital PCR before treatment initiation and once every 3 weeks during the study. The primary endpoints were(a) the diagnostic value of dPCR for detecting HER2 status in the blood and(b) the relevance of potential changes in the plasma HER2 level at 3 weeks from baseline for predicting treatment response. RESULTS: Overall, 85 patients were enrolled between October 9, 2018, and January 23, 2020. dPCR had a specificity of 91.67% (95% CI: 80.61% to 97.43%) for detecting HER2 amplification, and the area under the receiver operating characteristic (ROC) curve was 0.84 (p < 0.01). A clinically relevant specificity threshold of approximately 90%, which was equivalent to a ≥15% decrease in the plasma HER2 ratio at 3 weeks from baseline, showed a positive predictive value of 97.37% (95% CI: 77.11% to 98.65%) in terms of predicting clinical benefit. Patients whose plasma HER2 ratio was reduced by ≥15% had a longer median progression-free survival (PFS) than those whose ratio was reduced by <15% (9.20 months vs. 4.50 months, p < 0.01). CONCLUSIONS: Early changes in the plasma HER2 ratio may predict the treatment response in patients with advanced breast cancer and could facilitate optimal treatment selection.