Serum HER2 level as a predictor of treatment efficacy and prognosis in advanced breast cancer patients treated with trastuzumab deruxtecan (T-DXd).

1039 Background: No validated liquid biopsy biomarker currently exists to predict the efficacy of trastuzumab deruxtecan (T-DXd) in advanced breast cancer (aBC). Consequently, the role of serum HER2 (sHER2) in this setting remains unclear. This study aimed to determine whether baseline levels and early on-treatment dynamics of sHER2 are associated with progression-free survival (PFS) and objective response rate (ORR) in aBC patients receiving T-DXd. Methods: We analyzed 43 aBC patients treated with T-DXd who had baseline sHER2 data available. As the established 15 ng/mL cut-off was not associated with PFS in our preliminary analysis, the optimal baseline sHER2 cut-off for predicting PFS was determined using the maximally selected log-rank statistics method. Patients were stratified by this cut-off, and a multivariable Cox model adjusted for clinicopathological factors was used to evaluate the independent prognostic value of baseline serum HER2. In a subgroup of 20 patients, serum HER2 changes after two treatment cycles were assessed for association with efficacy outcomes. Results: Using the maximally selected log-rank statistics method, 10.94 ng/mL was identified as the optimal predictive cut-off, with PFS differing significantly between groups defined by this threshold (P=0.015). Multivariable Cox analysis confirmed that a baseline sHER2 level ≥10.94 ng/mL was an independent protective factor for longer PFS (Hazard Ratio [HR] = 0.129, 95% Confidence Interval [CI]: 0.019-0.875, P=0.038), corresponding to an approximately 87.1% reduction in disease progression risk. This predictive value was independent of primary tumor HER2 status, metastatic burden, and liver metastasis. Primary tumor HER2 status itself was not a significant prognostic factor in the model (P=0.530). Regarding ORR, no statistically significant difference was found between high and low sHER2 groups using the 10.94 ng/mL cut-off (P=0.392). In the 20-patient subgroup, early on-treatment sHER2 dynamics showed no significant association with PFS (P=0.455) or ORR (P=0.256). Conclusions: In aBC patients treated with T-DXd, a baseline serum HER2 level ≥10.94 ng/mL is a strong and independent predictive biomarker for PFS, with prognostic value potentially surpassing primary tumor HER2 status. This newly identified cut-off may help to better select patients who are most likely to benefit from T-DXd therapy using a convenient liquid biopsy approach. Clinical trial information: NCT06492447 .