Ultrasound-based radiomics and habitat analysis for noninvasive assessment of Ki-67 overexpression in breast cancer.

BACKGROUND

Accurate preoperative assessment of Ki-67 proliferation index remains a clinical challenge in breast cancer management. Conventional ultrasound radiomics often fails to fully capture intratumoral heterogeneity, suffers from overfitting, and includes redundant features that limit generalizability.

METHODS

In this retrospective study, we analyzed preoperative ultrasound images and immunohistochemical results from 288 women with pathologically confirmed breast cancer. We extracted both conventional radiomic features and intratumoral habitat features, computed risk scores, and integrated them with clinicopathological variables (e.g., progesterone receptor status, lymph node involvement) to construct a nomogram. Model performance was assessed by area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA).

RESULTS

The Clinics_Habitat_Radiomics model achieved AUCs of 0.877 (95% CI: 0.826-0.929) in the training cohort and in the validation cohort, the model achieved an AUC of 0.830, with a sensitivity of 60.3% and specificity of 91.7%, significantly outperforming other models. Calibration curves indicated close agreement between predicted probabilities and observed outcomes (Hosmer-Lemeshow test: p = 0.14 [training], p = 0.19 [validation]). DCA demonstrated superior net clinical benefit across a range of threshold probabilities compared with single-modality approaches.

CONCLUSIONS

The integration of habitat analysis with ultrasound-based radiomics enables the development of a nomogram that synergistically incorporates multimodal imaging features and clinicopathological parameters, offering a non-invasive predictive tool for Ki-67 expression in breast cancer. This model not only enhances the precision of tumor biology assessment but also provides actionable insights for optimizing therapeutic regimens, monitoring treatment responses, and stratifying prognostic risks, thereby bridging the gap between radiomic diagnostics and personalized oncology care.