Multidimensional MRI radiomics-based model predicts recurrence risk in triple-negative breast cancer

Patients with triple-negative breast cancer (TNBC) often have a poor prognosis. Existing staging systems cannot accurately assess patients' recurrence risk, due to the tumor's significant heterogeneity. Given the lack of a reliable prognostic assessment method, this study aimed to propose a combined multimodal prognostic model. We systematically integrated the clinicopathological indicators with multi-scale radiomics features in a large-scale TNBC cohort (n = 700). Remarkably, "intratumoral habitat heterogeneity analysis" and "peritumoral microenvironment characteristics" were innovatively incorporated in our model to reflect the potential biological behavior of individual tumors. The best consistency index (C-index) illuminated the superior prognostic capabilities of the combined model. Time-dependent ROC analysis demonstrated robust AUC at the 1-, 3-, and 5-year disease-free survival (DFS). Overall, our predictive model could provide a more comprehensive description of the biological aggressiveness of TNBC, overcoming the limitations of traditional staging systems in individualized prognosis assessment and better assisting doctors in developing personalized treatment and follow-up plans.