Deep learning-based recognition and segmentation of intracranial aneurysms under small sample size

Abstract

The manual identification and segmentation of intracranial aneurysms (IAs) involved in the 3D reconstruction procedure are labor-intensive and prone to human errors. To meet the demands for routine clinical management and large cohort studies of IAs, fast and accurate patient-specific IA reconstruction becomes a research Frontier. In this study, a deep-learning-based framework for IA identification and segmentation was developed, and the impacts of image pre-processing and convolutional neural network (CNN) architectures on the framework’s performance were investigated. Three-dimensional (3D) segmentation-dedicated architectures, including 3D UNet, VNet, and 3D Res-UNet were evaluated. The dataset used in this study included 101 sets of anonymized cranial computed tomography angiography (CTA) images with 140 IA cases. After the labeling and image pre-processing, a training set and test set containing 112 and 28 IA lesions were used to train and evaluate the convolutional neural network mentioned above. The performances of three convolutional neural networks were compared in terms of training performance, segmentation performance, and segmentation efficiency using multiple quantitative metrics. All the convolutional neural networks showed a non-zero voxel-wise recall (V-Recall) at the case level. Among them, 3D UNet exhibited a better overall segmentation performance under the relatively small sample size. The automatic segmentation results based on 3D UNet reached an average V-Recall of 0.797 ± 0.140 (3.5% and 17.3% higher than that of VNet and 3D Res-UNet), as well as an average dice similarity coefficient (DSC) of 0.818 ± 0.100, which was 4.1%, and 11.7% higher than VNet and 3D Res-UNet. Moreover, the average Hausdorff distance (HD) of the 3D UNet was 3.323 ± 3.212 voxels, which was 8.3% and 17.3% lower than that of VNet and 3D Res-UNet. The three-dimensional deviation analysis results also showed that the segmentations of 3D UNet had the smallest deviation with a max distance of +1.4760/−2.3854 mm, an average distance of 0.3480 mm, a standard deviation (STD) of 0.5978 mm, a root mean square (RMS) of 0.7269 mm. In addition, the average segmentation time (AST) of the 3D UNet was 0.053s, equal to that of 3D Res-UNet and 8.62% shorter than VNet. The results from this study suggested that the proposed deep learning framework integrated with 3D UNet can provide fast and accurate IA identification and segmentation.

Compare deep learning model and conventional logistic regression model for the identification of unstable saccular intracranial aneurysms in computed tomography angiography

“…Yang et al ( 18 ) trained a CNN-based algorithm that achieved 97.5% accuracy in detecting IAs. Zhu et al ( 19 ) demonstrated the fast and accurate identification and segmentation of IAs using a deep-learning-based framework from CTA images: 3D-UNet exhibited a better overall segmentation performance under a relatively small sample size. Wang et al ( 20 ) proved that the multiphase fusion DLM with automatic phase selection can automatically detect IAs more sensitively.…”

Staged cluster transformers for intracranial aneurysms segmentation from structure fused 3D MRA

“…3D U‐Net 22 is one of its variants which uses 3D convolutional operations to achieve voxel‐level classification for MRI images. Based on this, Zhu et al 43 evaluated the performance of 3D U‐Net, 22 V‐Net, 27 and 3D Res‐UNet 24 for IAs segmentation. A network 44 was proposed for IAs segmentation combining 3D U‐Net, residual connection and attention gate.…”

Advances in research and application of artificial intelligence and radiomic predictive models based on intracranial aneurysm images

“…At higher levels, CNNs can combine these simple features into complex patterns, such as specific shapes or objects ( 13 ). CNNs can also enrich data composition by extracting deep learning features through convolutional layers and combining them with radiomic features ( 14 ). Radiomics based on convolutional neural networks (CNNs) differs from traditional radiomics in that the models can automatically learn to extract and select image features.…”

“…Zhu et al ( 14 ) used small sample datasets to compare the detection of aneurysms using 3D UNet, VNet, and 3D Res-UNet, with 3D UNet showing the best detection performance. This suggests that the architectural structure of different CNN models could impact detection and segmentation capabilities.…”