Preoperative prediction of extensive intraductal component in invasive breast cancer based on intra- and peri-tumoral heterogeneity in high-resolution ultrafast DCE-MRI.

Preoperatively predicting extensive intraductal component in invasive breast cancer through imaging is crucial for informed decision-making, guiding surgical planning to mitigate risks of incomplete resection or re-operation for positive margins in breast-conserving surgery. This study aimed to characterize intra- and peri-tumor heterogeneity using high-spatial resolution ultrafast DCE-MRI to predict the extensive intraductal component in invasive breast cancer (IBC-EIC) preoperatively. A retrospective analysis included invasive breast cancer patients who underwent preoperative high-spatial resolution ultrafast DCE-MRI, categorized based on intraductal component status (IBC-EIC vs. IBC without EIC). Propensity score matching (PSM) was employed to balance clinicopathological covariates between the groups. Personalized kinetic intra-tumor heterogeneity (ITH) and peri-tumor heterogeneity (PTH) scores were quantified using clustered voxels with similar enhancement patterns. An image combined model, incorporating MRI features, ITH, and PTH scores, was developed and assessed. Of 368 patients, 26.4% (97/368) had IBC-EIC. PSM yielded well-matched pairs of 97 patients each. After PSM, ITH and PTH scores were significantly higher in the IBC-EIC group (ITH: 0.68 ± 0.23; PTH: 0.58 ± 0.19) compared to IBC without EIC (ITH: 0.32 ± 0.25; PTH: 0.42 ± 0.18; p < 0.001). Before PSM, ITH (0.71 ± 0.20 vs. 0.49 ± 0.28, p < 0.001) and PTH (0.61 ± 0.18 vs. 0.50 ± 0.20, p < 0.001) scores remained higher in the IBC-EIC group. The Image Combined Model demonstrated good predictive performance for IBC-EIC, with an AUC of 0.91 (95% CI 0.86-0.95) after PSM and 0.85 (95% CI 0.81-0.90) before PSM. Inclusion of ITH and PTH scores significantly improved prediction capability. ITH and PTH characterization from high-spatial resolution ultrafast DCE-MRI kinetic curves enhances preoperative prediction of IBC-EIC, offering valuable insights for personalized breast cancer management.