This record contains raw data related to article “The predictive role of radiomics in breast cancer patients im-aged by [18F]FDG PET: preliminary results from a prospective cohort” Abstract: Background: In the last decade, radiomics emerged as a source of image-derived biomarkers. However, existing data predominantly stem from retrospective analyses. We aimed to prospectively assess the predictive role of [18F]FDG PET radiomics in breast cancer (BC) patients. Methods: we prospectively enrolled stage I-III BC patients eligible for neoadjuvant chemotherapy (NAC), who underwent staging [18F]FDG PET/CT. All patients had data regarding pathological treatment response assessed in the post-NAC surgical specimen and were grouped in pathological complete responders (pCR) and pathological residual disease (non-pCR). Radiomic PET features were extracted from the volume of interest drawn on the primary breast lesion. The predictive role of clinical, histological, and radiomic data with respect to pCR was assessed. Univariate and multivariate statistics were used for inference; principal component analysis (PCA) was used for dimensionality reduction. Results: We analyzed 53 HER2+, and 40 triple-negative (TNBC) BC patients. pCR was obtained in 24/53(45%) HER2+ and 20/40(50%) TNBC patients. Age, molecular subtype, ki-67, and stage were not statistically different between classes and couldn’t predict pCR at multivariate analysis. At univariate analysis, 10 radiomic features resulted with a p < 0.1. 3/22 radiomic principal components (PC) were found to be discriminative for pCR. Using a cross-validation approach, the radiomic PC failed to discriminate pCR vs non-pCR groups but were able to predict the stage (mean accuracy = 0.79±0.08); Conclusions: These preliminary results demonstrate the potential of radiomic features extracted from PET for staging purposes in BC patients, while their possible role in predicting pCR to NAC needs to be further investigated.