Related Experiment Video
Updated: Jul 26, 2026
Automated Midline Shift and Intracranial Pressure Estimation based on Brain CT Images
Published on: April 13, 2013
A Novel Detection and Classification Framework for Diagnosing of Cerebral Microbleeds Using Transformer and Language
Cong Chen1,2, Lin-Lin Zhao3, Qin Lang2
1School of Clinical Medicine, College of Medicine, Nanjing Medical University, Nanjing 211166, China.
Abstract:
The detection of Cerebral Microbleeds (CMBs) is crucial for diagnosing cerebral small vessel disease. However, due to the small size and subtle appearance of CMBs in susceptibility-weighted imaging (SWI), manual detection is both time-consuming and labor-intensive. Meanwhile, the presence of similar-looking features in SWI images demands significant expertise from clinicians, further complicating this process. Recently, there has been a significant advancement in automated detection of CMBs using a Convolutional Neural Network (CNN) structure, aiming at enhancing diagnostic efficiency for neurologists. However, existing methods still show discrepancies when compared to the actual clinical diagnostic process. To bridge this gap, we introduce a novel multimodal detection and classification framework for CMBs' diagnosis, termed MM-UniCMBs. This framework includes a light-weight detection model and a multi-modal classification network. Specifically, we proposed a new CMBs detection network, CMBs-YOLO, designed to capture the salient features of CMBs in SWI images. Additionally, we design an innovative language-vision classification network, CMBsFormer (CF), which integrates patient textual descriptions-such as gender, age, and medical history-with image data. The MM-UniCMBs framework is designed to closely align with the diagnostic workflow of clinicians, offering greater interpretability and flexibility compared to existing methods. Extensive experimental results show that MM-UniCMBs achieves a sensitivity of 94% in CMBs' classification and can process a patient's data within 5 s.
Insights
This study introduces MM-UniCMBs, a new AI framework for detecting Cerebral Microbleeds (CMBs). It improves diagnostic accuracy and efficiency by combining imaging and patient data, closely mimicking clinical workflows.
Area of Science:
- Neurology
- Medical Imaging
- Artificial Intelligence
Background:
- Cerebral Microbleeds (CMBs) detection is vital for diagnosing cerebral small vessel disease.
- Manual CMB detection in susceptibility-weighted imaging (SWI) is challenging due to subtle features, requiring significant clinical expertise and time.
- Current automated methods using Convolutional Neural Networks (CNNs) show discrepancies with clinical diagnosis.
Purpose of the Study:
- To develop a novel multimodal detection and classification framework (MM-UniCMBs) for CMB diagnosis.
- To enhance the accuracy and efficiency of CMB detection and classification, aligning with clinical diagnostic workflows.
- To improve interpretability and flexibility in CMB diagnosis.
Main Methods:
- Introduced a lightweight detection model, CMBs-YOLO, for capturing salient CMB features in SWI images.
- Developed an innovative language-vision classification network, CMBsFormer (CF), integrating patient textual data with imaging.
- Integrated these components into the MM-UniCMBs framework, designed for clinical workflow alignment.
Main Results:
- MM-UniCMBs achieved a 94% sensitivity in CMB classification.
- The framework can process a patient's data within 5 seconds.
- Demonstrated improved performance compared to existing methods, with greater interpretability and flexibility.
Conclusions:
- The MM-UniCMBs framework offers a promising multimodal approach for accurate and efficient CMB diagnosis.
- This AI-driven system closely mimics clinical diagnostic processes, potentially aiding neurologists.
- The integration of imaging and clinical data enhances diagnostic capabilities for cerebral small vessel disease.
