Key points are not available for this paper at this time.
The predictive accuracy of the HER2DX pCR score for pCR following anti-HER2-based neoadjuvant CT, as compared to MRI response, is unknown. With the increasing reliance on MRI response in clinical trials for guiding surgical and systemic therapy decisions, understanding the discrimination ability of HER2DX pCR score is crucial. This study assesses the effectiveness of HER2DX pCR score in predicting pCR beyond MRI response. HER2DX was assessed on pre-treatment baseline FFPE tumor biopsies from 147 patients (pts) with stage I-III BC treated with trastuzumab-based CT at Hospital Clinic Barcelona. The primary aim was the association of HER2DX pCR score as a continuous and as a group variable with pCR status, factoring in MRI binary response (>90% response vs. others). The secondary aim was the predictive capacity of HER2DX pCR score within MRI-responders and non-responders. Logistic regression models were used for statistical analysis. MRI data pre- and post-therapy were available for 131 (89.7%) pts. The overall pCR rate in this subset was 51.9% (95% CI 43.0-60.7%). In univariable analysis, MRI response was associated with pCR (63.o% vs 34.0%, odds ratio OR=3.30, p<0.001), as well as HER2DX pCR continuous score (OR=1.57, p<0.001) and HER2DX pCR groups (high vs low: OR=16.51, p<0.001). Area under the curve was 0.64 for MRI and 0.78 for HER2DX pCR-groups. In a multivariable model, HER2DX pCR showed a significant association with pCR (high vs low: OR=13.87, p<0.001), while MRI response did not (OR=2.06, p=0.100). HER2DX pCR-score also was associated with pCR in both MRI-response and MRI-non-response groups (all p<0.001). Among MRI-responders, pCR rates were 72.4%, 74.2%, and 33.3% in the high, medium, and low HER2DX pCR score categories, respectively. Among MRI-non-responders, the pCR rates were 80.9%, 61.5%, and 3.6%, respectively. The HER2DX pCR score outperforms MRI response as a predictor of pCR in HER2+ BC undergoing anti-HER2-based neoadjuvant chemotherapy. It reliably predicts pCR rates regardless of MRI response status, emphasizing its effectiveness in guiding neoadjuvant treatment optimization.
Pascual et al. (Wed,) studied this question.