DCE-MRI based deep learning analysis of intratumoral subregion for predicting Ki-67 expression level in breast cancer

• Published in: Magnetic resonance imaging Vol. 119; p. 110370
• Main Authors: Ding, Zhimin, Zhang, Chengmeng, Xia, Cong, Yao, Qi, Wei, Yi, Zhang, Xia, Zhao, Nannan, Wang, Xiaoming, Shi, Suhua
• Format: Journal Article
• Language: English
• Published: Netherlands Elsevier Inc 01.06.2025
• Subjects: Adult; Aged; Algorithms; Breast - diagnostic imaging; Breast cancer; Breast Neoplasms - diagnostic imaging; Breast Neoplasms - metabolism; Contrast Media; Deep Learning; Dynamic Contrast Enhanced Magnetic Resonance Imaging; Female; Humans; Image Interpretation, Computer-Assisted - methods; Ki-67; Ki-67 Antigen - metabolism; Magnetic resonance imaging; Magnetic Resonance Imaging - methods; Middle Aged; Retrospective Studies; Subregion; Ki-67; DCE-MRI; DCA; CI; ROC; DL; T1WI; Breast cancer; IQR; AUC; Deep learning; Subregion; VOI; IVIM; Magnetic resonance imaging; DSC; T2WI; DWI; LASSO; mRMR; HER2; ITH
• ISSN: 0730-725X; 1873-5894; 1873-5894
• DOI: 10.1016/j.mri.2025.110370

To evaluate whether deep learning (DL) analysis of intratumor subregion based on dynamic contrast-enhanced MRI (DCE-MRI) can help predict Ki-67 expression level in breast cancer. A total of 290 breast cancer patients from two hospitals were retrospectively collected. A k-means clustering algorithm confirmed subregions of tumor. DL features of whole tumor and subregions were extracted from DCE-MRI images based on 3D ResNet18 pre-trained model. The logistic regression model was constructed after dimension reduction. Model performance was assessed using the area under the curve (AUC), and clinical value was demonstrated through decision curve analysis (DCA). The k-means clustering method clustered the tumor into two subregions (habitat 1 and habitat 2) based on voxel values. Both the habitat 1 model (validation set: AUC = 0.771, 95 %CI: 0.642–0.900 and external test set: AUC = 0.794, 95 %CI: 0.696–0.891) and the habitat 2 model (AUC = 0.734, 95 %CI: 0.605–0.862 and AUC = 0.756, 95 %CI: 0.646–0.866) showed better predictive capabilities for Ki-67 expression level than the whole tumor model (AUC = 0.686, 95 %CI: 0.550–0.823 and AUC = 0.680, 95 %CI: 0.555–0.804). The combined model based on the two subregions further enhanced the predictive capability (AUC = 0.808, 95 %CI: 0.696–0.921 and AUC = 0.842, 95 %CI: 0.758–0.926), and it demonstrated higher clinical value than other models in DCA. The deep learning model derived from subregion of tumor showed better performance for predicting Ki-67 expression level in breast cancer patients. Additionally, the model that integrated two subregions further enhanced the predictive performance. The deep learning model derived from subregion of breast cancer showed better performance than the whole tumor for predicting Ki-67 expression level. The model that integrated subregions further enhanced the predictive performance. [Display omitted] •Deep learning model derived from subregion of breast cancer showed better performance in predicting Ki-67 level.•The optimal performance was achieved when subregion signatures were integrated.•The results remained reliable for three different molecular subtype subgroups.