This record contains raw data related to to protocol ONC/OSS-02/2019 "Analysis of combined assessment of Radiomics pattern in [18]F-FDG-PET/CT, Immunogenic Cell Death induction and Microbiota in the prediction of response to NeoAdjuvant Chemotherapy in breast cancer patients: a proof-of-concept study (RIMNAC study)." AbstractIn breast cancer (BC), pathological complete response (pCR) prediction after neoadjuvant chemotherapy (NAC) remains challenging. Here, we hypothesize that combining radiomics-based [18]F-FDG PET/CT features, intratumoral microbiota profiling, and clinical data could refine NAC response prediction. To this aim, 110 BC patients were prospectively enrolled. Intratumoral microbiota was analysed from a baseline BC biopsy. Other data collected included clinicopathological parameters, germline pathogenic variants (GPVs), and pre-treatment [18]F-FDG PET/CT radiomics. Least Absolute Shrinkage Selection Operator regression was used for feature selection, followed by the development of a machine-learning model using logistic regression. Corynebacterium presence was inversely associated with pCR likelihood (62.5% in residual disease vs. 36.7% in pCR; p=0.01). The analysis further identified GPVs, and three radiomic features as pCR key predictors. The integrated model achieved a cross-validated AUC of 0.79±0.09. The proposed approach offers a tool for pre-treatment patient stratification, enabling personalized NAC strategies to enhance efficacy and minimize unnecessary toxicity.