Predicting pathological complete response to neoadjuvant systemic therapy for triple-negative breast cancers using
Summary
This study developed a deep learning model using MRI scans to predict treatment response in triple-negative breast cancer (TNBC) patients undergoing neoadjuvant systemic therapy (NAST). The model shows high accuracy in identifying patients likely to achieve pathologic complete response (pCR).
Area of Science:
- Radiology
- Oncology
- Artificial Intelligence
Background:
- Triple-negative breast cancer (TNBC) presents a significant challenge in treatment selection.
- Predicting response to neoadjuvant systemic therapy (NAST) is crucial for personalized treatment strategies.
- Early identification of treatment responders can optimize therapeutic decisions and improve patient outcomes.
Purpose of the Study:
- To develop and validate a deep learning (DL) model for predicting pathologic complete response (pCR) to NAST in TNBC patients.
- To assess the model's performance using multiparametric MRI data, including dynamic contrast-enhanced (DCE) MRI and diffusion-weighted imaging (DWI).
- To evaluate the model's potential for enabling early, personalized treatment adjustments in TNBC management.
Main Methods:
- A deep learning model was trained and validated using serial DCE-MRI and DWI data from TNBC patients.
- Input data included pre-treatment (baseline) and post-four cycles (C4) of doxorubicin/cyclophosphamide treatment MRI scans.
- Model performance was evaluated using area under the receiver operating characteristic curves (AUCs) across training, validation, testing, and prospective testing cohorts.
Main Results:
- The DL model achieved high AUCs for predicting pCR (standard definition) across all cohorts, reaching 0.88 ± 0.02 in the testing group.
- For breast-only pCR prediction, the retrained model demonstrated strong performance, with AUCs of 0.86 ± 0.03 in the testing group.
- The model showed robust predictive capabilities, indicating its potential for early treatment response assessment.
Conclusions:
- The developed deep learning model, utilizing multiparametric MRI, is highly effective in predicting treatment response to NAST in TNBC patients.
- This AI-driven approach can potentially distinguish between pCR and non-pCR patients early in the treatment course.
- The findings suggest a promising tool for personalized treatment strategies in TNBC, enabling timely therapeutic modifications.
