Dataset Information

Multi-View Convolutional Neural Networks in Rupture Risk Assessment of Small, Unruptured Intracranial Aneurysms.

ABSTRACT: Auto-detection of cerebral aneurysms via convolutional neural network (CNN) is being increasingly reported. However, few studies to date have accurately predicted the risk, but not the diagnosis itself. We developed a multi-view CNN for the prediction of rupture risk involving small unruptured intracranial aneurysms (UIAs) based on three-dimensional (3D) digital subtraction angiography (DSA). The performance of a multi-view CNN-ResNet50 in accurately predicting the rupture risk (high vs. non-high) of UIAs in the anterior circulation measuring less than 7 mm in size was compared with various CNN architectures (AlexNet and VGG16), with similar type but different layers (ResNet101 and ResNet152), and single image-based CNN (single-view ResNet50). The sensitivity, specificity, and overall accuracy of risk prediction were estimated and compared according to CNN architecture. The study included 364 UIAs in training and 93 in test datasets. A multi-view CNN-ResNet50 exhibited a sensitivity of 81.82 (66.76-91.29)%, a specificity of 81.63 (67.50-90.76)%, and an overall accuracy of 81.72 (66.98-90.92)% for risk prediction. AlexNet, VGG16, ResNet101, ResNet152, and single-view CNN-ResNet50 showed similar specificity. However, the sensitivity and overall accuracy were decreased (AlexNet, 63.64% and 76.34%; VGG16, 68.18% and 74.19%; ResNet101, 68.18% and 73.12%; ResNet152, 54.55% and 72.04%; and single-view CNN-ResNet50, 50.00% and 64.52%) compared with multi-view CNN-ResNet50. Regarding F1 score, it was the highest in multi-view CNN-ResNet50 (80.90 (67.29-91.81)%). Our study suggests that multi-view CNN-ResNet50 may be feasible to assess the rupture risk in small-sized UIAs.

SUBMITTER: Ahn JH

PROVIDER: S-EPMC8064331 | biostudies-literature | 2021 Mar

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Multi-View Convolutional Neural Networks in Rupture Risk Assessment of Small, Unruptured Intracranial Aneurysms.

Ahn Jun Hyong JH Kim Heung Cheol HC Rhim Jong Kook JK Park Jeong Jin JJ Sigmund Dick D Park Min Chan MC Jeong Jae Hoon JH Jeon Jin Pyeong JP

Journal of personalized medicine 20210324 4

Auto-detection of cerebral aneurysms via convolutional neural network (CNN) is being increasingly reported. However, few studies to date have accurately predicted the risk, but not the diagnosis itself. We developed a multi-view CNN for the prediction of rupture risk involving small unruptured intracranial aneurysms (UIAs) based on three-dimensional (3D) digital subtraction angiography (DSA). The performance of a multi-view CNN-ResNet50 in accurately predicting the rupture risk (high vs. non-hig ...[more]

PMID: 33805171

Similar Datasets

Project description:ObjectivesArterial calcification is thought to protect against rupture of intracranial aneurysms, but studies in a representative population of intracranial aneurysm patients have not yet been performed. The aim was to compare the prevalence of aneurysm wall calcification and intracranial carotid artery calcification (ICAC) between patients with an unruptured intracranial aneurysm (UIA) and a ruptured intracranial aneurysm (RIA).Materials and methodsWe matched 150 consecutive UIA patients to 150 RIA patients on age and sex. Aneurysm wall calcification and ICAC were quantified on non-contrast enhanced computed tomography images with the modified Agatston score. We compared the prevalence of aneurysm wall calcification, ICAC, and severe ICAC (defined as a modified Agatston score in the fourth quartile) between UIA and RIA patients using univariate and multivariate conditional logistic regression models adjusted for aneurysm characteristics and cardiovascular risk factors.ResultsAneurysm wall calcification was more prevalent in UIA compared to RIA patients (OR 5.2, 95% CI: 2.0-13.8), which persisted after adjustment (OR 5.9, 95% CI: 1.7-20.2). ICAC prevalence did not differ between the two groups (crude OR 0.9, 95% CI: 0.5-1.8). Severe ICAC was more prevalent in UIA patients (OR 2.0, 95% CI: 1.1-3.6), but not after adjustment (OR 1.0, 95% CI: 0.5-2.3).ConclusionsAneurysm wall calcification but not ICAC was more prevalent in UIAs than in RIAs, which corresponds to the hypothesis that calcification may protect against aneurysmal rupture. Aneurysm wall calcification should be further assessed as a predictor of aneurysm stability in prospective cohort studies.Clinical relevance statementCalcification of the intracranial aneurysm wall was more prevalent in unruptured than ruptured intracranial aneurysms after adjustment for cardiovascular risk factors. Calcification may therefore protect the aneurysm against rupture, and aneurysm wall calcification is a candidate predictor of aneurysm stability.Key pointsAneurysm wall calcification was more prevalent in patients with unruptured than ruptured aneurysms, while internal carotid artery calcification was similar. Aneurysm wall calcification but not internal carotid artery calcification is a candidate predictor of aneurysm stability. Cohort studies are needed to assess the predictive value of aneurysm wall calcification for aneurysm stability.

Project description:BACKGROUND:Ruptured aneurysms, the commonest cause of nontraumatic subarachnoid hemorrhage, can be catastrophic; the mortality and morbidity of affected patients being very high. Some risk factors, such as smoking, hypertension and female sex have been identified, whereas others, such as hemodynamics, imaging, and genomics, remain unclear. Currently, no accurate model that includes all factors for predicting such rupture is available. We plan to use data from a large cohort of Chinese individuals to set up a multidimensional model for predicting risk of rupture of unruptured intracranial aneurysms (UIAs). METHODS:The China Intracranial Aneurysm Project-2 (CIAP-2) will comprise screening of a cohort of 500 patients with UIA (From CIAP-1) and focus on hemodynamic factors, high resolution magnetic resonance imaging (HRMRI) findings, genetic factors, and biomarkers. Possible risk factors for rupture of UIA, including genetic factors, biomarkers, HRMRI, and hemodynamic factors, will be analyzed. The first project of the China Intracranial Aneurysm Project (CIAP-1; chaired by the Department of Neurosurgery, Tangdu Hospital, Fourth Military Medical University, Xi'an, Shaanxi, China) will prospectively collect a cohort of 5000 patients with UIA from 20 centers in China, and collect baseline information for each patient. Multidimensional data will be acquired in follow-up assessments. Statistically significant clinical features in the UIA cohort will also be analyzed and integrated into the model for predicting risk of UIA rupture. After the model has been set up, the resultant evidence-based prediction will provide a preliminary theoretical basis for treating aneurysms at high risk of rupture. DISCUSSION:This study will explore the risk of rupture of aneurysms and develop a scientific multidimensional model for predicting rupture of unruptured intracranial aneurysms. Clinical Trials registration A Study on a Multidimensional Prediction Model for Rupture Risk of Unruptured Intracranial Aneurysms (CIAP-2), NCT03133624. Registered: 16 April 2017. https://clinicaltrials.gov/ct2/show/NCT03133624.

Project description:PurposeTo investigate whether multi-view convolutional neural networks can improve a fully automated lymph node detection system for pelvic MR Lymphography (MRL) images of patients with prostate cancer.MethodsA fully automated computer-aided detection (CAD) system had been previously developed to detect lymph nodes in MRL studies. The CAD system was extended with three types of 2D multi-view convolutional neural networks (CNN) aiming to reduce false positives (FP). A 2D multi-view CNN is an efficient approximation of a 3D CNN, and three types were evaluated: a 1-view, 3-view, and 9-view 2D CNN. The three deep learning CNN architectures were trained and configured on retrospective data of 240 prostate cancer patients that received MRL images as the standard of care between January 2008 and April 2010. The MRL used ferumoxtran-10 as a contrast agent and comprised at least two imaging sequences: a 3D T1-weighted and a 3D T2*-weighted sequence. A total of 5089 lymph nodes were annotated by two expert readers, reading in consensus. A first experiment compared the performance with and without CNNs and a second experiment compared the individual contribution of the 1-view, 3-view, or 9-view architecture to the performance. The performances were visually compared using free-receiver operating characteristic (FROC) analysis and statistically compared using partial area under the FROC curve analysis. Training and analysis were performed using bootstrapped FROC and 5-fold cross-validation.ResultsAdding multi-view CNNs significantly (p < 0.01) reduced false positive detections. The 3-view and 9-view CNN outperformed (p < 0.01) the 1-view CNN, reducing FP from 20.6 to 7.8/image at 80% sensitivity.ConclusionMulti-view convolutional neural networks significantly reduce false positives in a lymph node detection system for MRL images, and three orthogonal views are sufficient. At the achieved level of performance, CAD for MRL may help speed up finding lymph nodes and assessing them for potential metastatic involvement.

Dataset Information

Multi-View Convolutional Neural Networks in Rupture Risk Assessment of Small, Unruptured Intracranial Aneurysms.

Publications

Multi-View Convolutional Neural Networks in Rupture Risk Assessment of Small, Unruptured Intracranial Aneurysms.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets