Radiomics has been spotlighted as imaging biomarker for estimation of intratumoral heterogeneity (ITH) which is regarded as the main reason for resistance to tumor treatment. Although a number of studies has shown clinical evidences that separate measurement of metabolic ITH by texture features (TFs) on 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) has prognostic ability in various tumors, there has been no consensus regarding the best parameter representing ITH. Besides, it is yet uncertain that TFs are useful for estimation of histopathologic markers, prediction of response to neoadjuvant chemotherapy (NAC), or prognostic ability in breast cancer. To depart from the traditional approach, we evaluated the clinical usefulness of integrated metabolic radiomics using unsupervised clustering with 109 TFs measured from pretreatment 18F-FDG PET/CT scans of 73 patients with locally advanced breast cancer (LABC) underwent NAC before surgery. Our study shows that metabolic radiomics patterns of LABC are associated with Ki67 expression, achievement of pathologic complete response after NAC, and risk of recurrence. Integrated metabolic radiomics has potential for clinically relevant pretreatment biomarker with predictive and prognostic ability for personalized management in LABC.