AI-based medical e-diagnosis for fast and automatic ventricular volume measurement in patients with normal pressure hydrocephalus

Abstract

Based on CT and MRI images acquired from normal pressure hydrocephalus (NPH) patients, using machine learning methods, we aim to establish a multimodal and high-performance automatic ventricle segmentation method to achieve an efficient and accurate automatic measurement of the ventricular volume. First, we extract the brain CT and MRI images of 143 definite NPH patients. Second, we manually label the ventricular volume (VV) and intracranial volume (ICV). Then, we use the machine learning method to extract features and establish automatic ventricle segmentation model. Finally, we verify the reliability of the model and achieved automatic measurement of VV and ICV. In CT images, the Dice similarity coefficient (DSC), intraclass correlation coefficient (ICC), Pearson correlation, and Bland–Altman analysis of the automatic and manual segmentation result of the VV were 0.95, 0.99, 0.99, and 4.2 ± 2.6, respectively. The results of ICV were 0.96, 0.99, 0.99, and 6.0 ± 3.8, respectively. The whole process takes 3.4 ± 0.3 s. In MRI images, the DSC, ICC, Pearson correlation, and Bland–Altman analysis of the automatic and manual segmentation result of the VV were 0.94, 0.99, 0.99, and 2.0 ± 0.6, respectively. The results of ICV were 0.93, 0.99, 0.99, and 7.9 ± 3.8, respectively. The whole process took 1.9 ± 0.1 s. We have established a multimodal and high-performance automatic ventricle segmentation method to achieve efficient and accurate automatic measurement of the ventricular volume of NPH patients. This can help clinicians quickly and accurately understand the situation of NPH patient's ventricles.

Automatic Ventriculomegaly Detection in Fetal Brain MRI: A Step-by-Step Deep Learning Model for Novel 2D-3D Linear Measurements

“…It could also provide both right and left ventricles in one cut (compared to two cuts in manual measurement). The following sections describe our AI model's novelty, advantages, and limitations [18,19].…”

Deep learning-based segmentation of brain parenchyma and ventricular system in CT scans in the presence of anomalies

“…With this aim, we have demonstrated that a robust and fully automatic segmentation of brain parenchyma and ventricular system can be achieved by deep learning. Unlike previous publications on the topic, that have either excluded subjects with severe anomalies (e.g., Huff et al, 2019 ; Cai et al, 2020 ) or focused on a particular kind of disease (e.g., Zhou et al, 2020 , 2022 ), our developed algorithm is robust for two different kinds of anomalies (intracranial hemorrhage and tumors), despite being trained only on normal and hemorrhage data with partially incomplete annotations. In addition, the higher generalizability and robustness to anatomical changes does not come with a decrease in segmentation performance on normal anatomy, which is equally important.…”

“…Multiple DL-based methods for segmentation of brain parenchyma and ventricles have been proposed: either for direct quantification (Huff et al, 2019 ; Zhou et al, 2020 ) or as a pre-processing step for registration purposes (Dubost et al, 2020 ; Walluscheck et al, 2023 ). In further studies, substructures of the brain parenchyma are segmented (Cai et al, 2020 ; Zopes et al, 2021 ) or a cross-modality approach for CT and MRI is proposed (Zopes et al, 2021 ; Zhou et al, 2022 ). However, these studies often focus on specific diseases such as hydrocephalus or are evaluated only on data of healthy or elderly patients.…”

Assessing CT-based Volumetric Analysis via Transfer Learning with MRI and Manual Labels for Idiopathic Normal Pressure Hydrocephalus

“…Interestingly, while numerous studies have successfully demonstrated brain tissue segmentation and volumetric assessment using MR images 26,27 , these techniques have encountered challenges when attempted using CT scans with inferior soft tissue contrast. Recent research has thus explored robust automated and semi-automated segmentation(s) 28,29 and volumetric assessments of CT images using state-of-the-art image processing techniques centred on deep learning; specifically, segmentation models employing deep learning architectures, such as fully connected convolutional neural networks and 2D/3D U-Nets have been utilized [30][31][32][33] . However, most deep-learningbased studies have thus far been conducted using limited datasets and relied on manually annotated structural labels.…”