Predicting breast cancer pathological complete response with clinical and imaging data

INTRODUCTION: Pathological complete response (pCR) is a key prognostic indicator in breast cancer (BC) patients receiving neoadjuvant chemotherapy (NAC). Unimodal prediction models are limited, underscoring the need for multimodal machine learning approaches. METHODS: This retrospective study included 211 BC patients. A radiomics score (Radscore) was developed from multiparametric MRI, and integrated with clinical predictors using machine learning to build three models: clinical, radiomics, and a combined clinical-radiomics model. Their performance was systematically compared. RESULTS: The combined clinical-radiomics model significantly outperformed unimodal models, with an AUC of 0.937 in the training cohort and 0.853 in the validation cohort. Decision curve and calibration analyses suggested potential clinical utility and accuracy. CONCLUSION: The clinical-radiomics model accurately predicts pCR to NAC in BC, surpassing unimodal methods. Its dynamic nomogram aids in personalizing treatment decisions.