Independent Predictors of Pathological Complete Response in Breast Cancer: A Multivariable Analysis

Neoadjuvant chemotherapy (NAC) is widely used in the treatment of breast cancer, though response significantly varies across patients (Wang et al., 2017; Abidi et al., 2023). Biomarkers such as HER2 status, hormone receptor (HR) status, and MammaPrint risk score are commonly used to predict treatment outcomes, but their independent predictive value remains unclear.

This study aimed to determine which of these biomarkers independently predict pathological complete response (pCR) when analyzed simultaneously. Publicly available data from 654 patients in the I-SPY2 trial (GSE194040) were analyzed using chi-square tests and multivariable logistic regression.

HER2-positive status (OR = 4.62, 95% CI: 3.04–7.11, p < 0.001) and ultra-high MammaPrint risk (OR = 3.10, 95% CI: 1.95–4.98, p < 0.001) were identified as significant independent predictors of pCR, while HR status was not significant after adjustment (OR = 0.67, 95% CI: 0.43–1.03, p = 0.068). The model demonstrated moderate predictive performance (AUC = 0.70).

These findings suggest that evaluating biomarkers in isolation may overestimate their predictive importance, and that multivariable analysis provides a more accurate approach for interpreting treatment response, supporting better informed clinical decision-making.