Prediction of early clinical response to neoadjuvant chemotherapy in Triple-negative breast cancer: Incorporating Radiomics through breast MRI

Abstract

This study assessed pretreatment breast MRI coupled with machine learning for predicting early clinical responses to neoadjuvant chemotherapy (NAC) in triple-negative breast cancer (TNBC), focusing on identifying non-responders. A retrospective analysis of 135 TNBC patients (107 responders, 28 non-responders) treated with NAC from January 2015 to October 2022 was conducted. Non-responders were defined according to RECIST guidelines. Data included clinicopathologic factors and clinical MRI findings, with radiomics features from contrast-enhanced T1-weighted images, to train a stacking ensemble of 13 machine learning models. For subgroup analysis, propensity score matching was conducted to adjust for clinical disparities in NAC response. The efficacy of the models was evaluated using the area under the receiver-operating-characteristic curve (AUROC) before and after matching. The model combining clinicopathologic factors and clinical MRI findings achieved an AUROC of 0.752 (95% CI 0.644-0.860) for predicting non-responders, while radiomics-based models showed 0.749 (95% CI 0.614-0.884). An integrated model of radiomics, clinicopathologic factors, and clinical MRI findings reached an AUROC of 0.802 (95% CI 0.699-0.905). After propensity score matching, the hierarchical order of key radiomics features remained consistent. Our study demonstrated the potential of using machine learning models based on pretreatment MRI to non-invasively predict TNBC non-responders to NAC.

Fig. 1

Receiver operating characteristic (ROC) curves and bar plots depicting the performances of clinicopathologic factors, clinical MRI findings, their combination (a), and radiomics features extracted from different regions of interest (ROIs): Intratumor-ROI, peritumor-ROI, and combined-ROI (b).

Fig. 2

ROC curve for the integrated model of clinicopathologic factors, clinical MRI findings, and radiomics features.

Fig. 3

Flowchart of the study population.

Fig. 4

Two representative samples of non-responder and responder. Axial fat-saturated, contrast-enhanced MR images before (a) and after (b) neoadjuvant chemotherapy demonstrate disease progression in an irregular, heterogeneously enhancing mass. The predicted rates of non-response according to clinicopathologic + MRI model, radiomics model, and integrated model were 86.4%, 91.9%, and 85.9%, respectively. Surgical specimen image (hematoxylin and eosin (H&E) staining, (c) demonstrates extensive residual cancer burden following neoadjuvant chemotherapy. Axial fat-saturated, contrast-enhanced MR images before (d) and after (e) neoadjuvant chemotherapy demonstrate a radiological complete response in an oval rim-enhancing mass. The predicted rates of non-response according to clinicopathologic + MRI model, radiomics model, and integrated model were 12.9%, 4.8%, and 12.1%, respectively. Surgical specimen image (H&E staining, f) demonstrates predominant fibrous tissue and scattered inflammatory cells, with residual malignant cells discernible.

Fig. 5

Workflow scheme depicting the processes of data acquisition, segmentation, radiomics feature extraction, modeling, and response prediction.