.png){kind=icon}
.png){kind=icon}
Return
CNN-MVM: An Interactive Dual-Stream Network for Meningioma Subtype Classification
.png){kind=icon}
PRE
AI
Y
S
Y
M
Y
H
DOI：10.1002/ima.70407.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Meningiomas constitute a common type of intracranial tumor, and accurate preoperative classification is of great importance for clinical treatment planning and prognostic assessment. Non-invasive imaging-based prediction of meningioma subtypes may assist clinical decision-making in treatment planning. To this end, this study proposes a hybrid CNN-MVM network model. The model combines the local feature extraction capability of a convolutional neural network (CNN) with a Multi-scale Variable State Space (MVSS) module for efficient global contextual modeling. By incorporating a Multi-scale Edge Enhancement Module (MEEM) into the MVSS module, lesion boundary features are further enhanced, thereby enabling accurate classification of meningioma subtypes. Validated on T1CE MRI data from 177 meningioma patients, the model achieved an overall accuracy of 82.34%, outperforming several representative deep learning models. The proposed CNN-MVM network provides an effective approach for improving the accuracy of preoperative meningioma subtype classification.
Keywords:
convolutional neural networks
Mamba
meningioma classification
MRI images
Journal
IF:
2.5
Papers:
2.1K
Citations:
2.3K
Organization
Cited Papers