Radiomics Based on DCE-MRI for Predicting Response to Neoadjuvant Therapy in Breast Cancer.

Area of Science:

• Oncology
• Medical Imaging
• Radiology

Background:

• Neoadjuvant therapy (NAT) is crucial for breast cancer treatment.
• Accurate early prediction of NAT response is essential for treatment optimization.
• Current methods like diameter percentage lack sufficient predictive power.

Purpose of the Study:

• To compare radiomics and diameter percentage in predicting breast cancer response to NAT using pre- and early-treatment dynamic enhanced MR (DCE-MRI).
• To develop a tool for early, noninvasive prediction of NAT outcomes.

Main Methods:

• Retrospective analysis of 142 invasive breast cancer patients' DCE-MRI data before and after two NAT cycles.
• Construction of Diameter% and radiomics models (pre-NAT, early-NAT, delta radiomics).
• Development of a fusion model combining significant clinical indicators and radiomics scores.

Main Results:

• The delta radiomics model (AUC=0.87) outperformed pre-NAT, early-NAT radiomics, and Diameter% models (AUC=0.83) in the test set.
• The fusion model achieved the highest efficacy in predicting pathological complete response (pCR) with AUCs of 0.91 in both training and test sets.
• The fusion model demonstrated the greatest clinical benefit and significantly improved pCR prediction compared to the Diameter% model.

Conclusions:

• Radiomics, particularly delta and early-NAT radiomics, shows potential as biomarkers for early, noninvasive prediction of NAT outcomes.
• A fusion model integrating clinicopathological indicators and radiomics effectively predicts NAT response.
• These findings support the use of advanced imaging analysis for personalized breast cancer treatment.