The key component in deep learning research is the availability of training data sets. With a limited number of publicly available COVID-19 chest X-ray images, the generalization and robustness of deep learning models to detect COVID-19 cases developed based on these images are questionable. We aimed to use thousands of readily available chest radiograph images with clinical findings associated with COVID-19 as a training data set, mutually exclusive from the images with confirmed COVID-19 cases, which will be used as the testing data set. We used a deep learning model based on the ResNet-101 convolutional neural network architecture, which was pretrained to recognize objects from a million of images and then retrained to detect abnormality in chest X-ray images. The performance of the model in terms of area under the receiver operating curve, sensitivity, specificity, and accuracy was 0.82, 77.3%, 71.8%, and 71.9%, respectively. The strength of this study lies in the use of labels that have a strong clinical association with COVID-19 cases and the use of mutually exclusive publicly available data for training, validation, and testing.
The recent diagnostic assessment of cerebrovascular disease makes use of computational fluid dynamics (CFD) to quantify blood flow and determine the hemodynamics factors contributing to the disease from patient-specific models. However, compliant and anatomical patient-specific geometries are generally reconstructed from the medical images with different threshold values subjectively. Therefore, this paper tends to present the effect of extracted geometry with different threshold coefficient, Cthres by using a patient-specific cerebral aneurysm model. A set of medical images, digital subtraction angiography (DSA) images from the real patient diagnosed with internal carotid artery (ICA) aneurysm was obtained. The threshold value used to extract the patient-specific cerebral aneurysm geometry was calculated by using a simple threshold determination method. Several threshold coefficients, Cthres such as 0.2, 0.3, 0.4, 0.5 and 0.6 were employed in the image segmentation creating three-dimensional (3D) realistic arterial geometries that were then used for CFD simulation. As a result, we obtained that the volume of patient-specific cerebral aneurysm geometry decreases as the threshold coefficient, Cthres increases. There is dislocation of artery attached to the ICA aneurysm geometry occurred at a high threshold coefficient, Cthres. Besides, the physical changes also bring remarkable physiological effect on the wall shear stress (WSS) distribution and velocity flow field at patient-specific cerebral aneurysm geometry reconstructed with different threshold coefficient, Cthres.
Nowadays, the cerebral aneurysm is an abnormal focal dilation of a brain artery which is considered as a serious and potentially life-threatening condition. The rupture of an aneurysm causes subarachnoid hemorrhage (SAH) and is associated with high rates of morbidity and mortality. A better understanding of the mechanisms underlying aneurysm pathophysiology is crucial for the development of new preventive procedures and therapeutic strategies. This study focuses on the modeling and simulation of the blood flow analysis using simplified aneurysm models to perform early prediction on the geometrical effects of hemodynamics. The investigation involves three simplified models of aneurysms reconstructed using Solidworks 2019, in which the aneurysms are developed at the bifurcation. The qualitative comparison of the hemodynamics between three models was obtained and the geometrical effects were evaluated. The results show that the differences in shape and geometry on aneurysms affect the hemodynamics trend and are capable to apply for further understanding of problems regarding hemodynamics in the patient.
Nowadays, the knowledge of precise blood flow patterns in human blood vessels, especially focusing on Carotid Bifurcations Artery (CBA) area by using computational and modern techniques are very important to develop our understanding regarding to human diseases for both essential research and clinical treatment. This paper tends to discuss the progress regarding to the integration between Phase Contrast Magnetic Resonance Imaging (PC-MRI) and Computational Fluid Dynamics (CFD), specifically to the human diseases. We technically define the model geometry reconstruction, review both PC-MRI and CFD methods to create mesh models, obtain boundary conditions, define the governing equations in CFD, define the material properties, and assumptions used in running the CFD simulations. Detailed information on PC-MRI and CFD is provided in tables, such as the MRI setup, software used, CFD model setup, measurement parameter, and summary of the result contribution from each reviewed article. Numerous fusions between PC-MRI and CFD are specified by summarizing the investigation carried out by significant group’s research, reviewing the important outcomes, and discussing the techniques, drawbacks and possibilities for further study. We hope that this perspective analysis will encourage a fusion of PC-MRI and CFD research contributing to continuous advancement of human health with close cooperation and collaboration among clinicians and engineers.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.