A custom 13-layer convolutional neural network achieved 86.17% accuracy and an ROC-AUC of 0.86 in classifying carotid plaques from MRI scans, outperforming Inception V3.
Does a hybrid deep learning framework improve carotid plaque segmentation and classification on MRI compared to conventional methods?
A hybrid deep learning framework integrating Mask R-CNN and a custom CNN enables automated, accurate segmentation and classification of carotid plaques on MRI.
Effect estimate: ROC-AUC 0.86
Absolute Event Rate: 86.17% vs 84.21%
p-value: p=0.0001
Background: Accurate segmentation and classification of carotid plaques are critical for assessing stroke risk. However, conventional methods are hindered by manual intervention, inter-observer variability, and poor generalizability across heterogeneous datasets, limiting their clinical utility. Methods: We propose a hybrid deep learning framework integrating Mask R-CNN for automated plaque segmentation with a dual-path classification pipeline. A dataset of 610 expert-annotated MRI scans from Xiangya Hospital was processed using Plaque Texture Analysis Software (PTAS) for ground truth labels. Mask R-CNN was fine-tuned with multi-task loss to address class imbalance, while a custom 13-layer CNN and Inception V3 were employed for classification, leveraging handcrafted texture features and deep hierarchical patterns. The custom CNN was evaluated via K10 cross-validation, and model performance was quantified using Dice Similarity Coefficient (DSC), Intersection over Union (IoU), accuracy, and ROC-AUC. Results: = 0.0001), outperforming Inception V3 (84.21% accuracy). Both models significantly surpassed conventional methods in plaque characterization, with the custom CNN showing superior discriminative power for high-risk plaques. Conclusion: This study establishes a fully automated, hybrid framework that synergizes segmentation and classification to advance stroke risk stratification. By reducing manual dependency and inter-observer variability, our approach enhances reproducibility and generalizability across diverse clinical datasets. The statistically significant ROC-AUC and high accuracy underscore its potential as an AI-driven diagnostic tool, paving the way for standardized, data-driven cerebrovascular disease management.
Alregal et al. (Fri,) conducted a other in Carotid plaque (n=106). Custom 13-layer Convolutional Neural Network (CNN) vs. Inception V3 was evaluated on Classification accuracy for vulnerable vs stable carotid plaque (ROC-AUC 0.86, p=0.0001). A custom 13-layer convolutional neural network achieved 86.17% accuracy and an ROC-AUC of 0.86 in classifying carotid plaques from MRI scans, outperforming Inception V3.