HER2DX genomic test in HER2-positive breast cancer treated with 15 weeks of neoadjuvant paclitaxel, trastuzumab, and pertuzumab (THP): Final analysis from the BiOnHER clinical trial.

Abstract

607 Background: HER2DX is a 27-gene genomic assay providing prognostic and predictive insights in early-stage HER2-positive (HER2+) breast cancer. Although widely validated, less data exists for its performance in THP beyond the DAPHNe trial (JAMA Oncol 2023), which included 80 patients (pts). This study aimed to validate HER2DX for predicting pathological complete response (pCR) with THP and compare its performance to hormone receptor (HR) status. Methods: HER2DX (Reveal Genomics) was centrally evaluated on all tumor biopsies from the BiOnHER trial (NCT05912062), where pts with stage I-III HER2-positive breast cancer received 15 weeks of neoadjuvant THP at the Catalan Institute of Oncology. Biopsies were collected at pre-treatment baseline (D1) and day 8 (D8) after an HP loading dose but before initiating paclitaxel. HER2DX pCR group cutoffs were based on predefined thresholds for HER2DX low-, medium-, and high groups. Logistic regression and receiver-operating characteristic (ROC) curve analyses were used for statistical evaluation. The primary objective was to assess HER2DX pCR score for predicting pCR (ypT0/isN0). Secondary objectives included evaluating HER2DX performance by HR status, baseline TILs/Ki67, and additional insights from D8 data. Results: HER2DX was successfully evaluated in all 83 enrolled patients. The cohort included 65.1% T2 tumors, 62.7% cN0 status, 69.9% stage II disease, and 67.5% HR-positive tumors. The overall pCR rate was 45.8% (95% CI, 34.9–57.0%), and the ypT1miN0 rate was 54.2%. HER2DX pCR score was significantly associated with pCR (odds ratio [OR] = 5.26, P < 0.001), with an AUC of 0.835. Patients were categorized into 35.0% low, 37.5% medium, and 27.5% high HER2DX pCR score groups, with pCR rates of 13.3% (95% CI, 4.4–31.6%), 51.6% (95% CI, 33.4–69.4%), and 81.8% (95% CI, 59.0–94.0%), respectively. Among HR-negative tumors, pCR rates were 78.6% for high-pCR and 0.0% for low-pCR groups, while in HR-positive tumors, pCR rates were 87.5% and 13.8%, respectively. HR status alone was associated with pCR (OR = 0.125, P = 0.006) but lost significance in multivariable analysis including HER2DX. Baseline Ki67 (median: 35.0%) and TILs (median: 8.0%) were not associated with pCR. While D8 data offered biological insights, it did not improve predictive performance beyond baseline HER2DX. Conclusions: HER2DX is a robust predictor of pCR following neoadjuvant THP in stage I-III HER2-positive breast cancer, outperforming HR status. Baseline TILs and Ki67 were not predictive of pCR, and HER2DX D8 data did not improve predictive performance. Clinical trial information: NCT05912062 .