Deep learning based detection of intracranial aneurysms on digital subtraction angiography: A feasibility study

Abstract

<jats:sec><jats:title>Background</jats:title><jats:p> Digital subtraction angiography is the gold standard for detecting and characterising aneurysms. Here, we assess the feasibility of commercial-grade deep learning software for the detection of intracranial aneurysms on whole-brain anteroposterior and lateral 2D digital subtraction angiography images. </jats:p></jats:sec><jats:sec><jats:title>Material and methods</jats:title><jats:p> Seven hundred and six digital subtraction angiography images were included from a cohort of 240 patients (157 female, mean age 59 years, range 20–92; 83 male, mean age 55 years, range 19–83). Three hundred and thirty-five (47%) single frame anteroposterior and lateral images of a digital subtraction angiography series of 187 aneurysms (41 ruptured, 146 unruptured; average size 7±5.3 mm, range 1–5 mm; total 372 depicted aneurysms) and 371 (53%) aneurysm-negative study images were retrospectively analysed regarding the presence of intracranial aneurysms. The 2D data was split into testing and training sets in a ratio of 4:1 with 3D rotational digital subtraction angiography as gold standard. Supervised deep learning was performed using commercial-grade machine learning software (Cognex, ViDi Suite 2.0). Monte Carlo cross validation was performed. </jats:p></jats:sec><jats:sec><jats:title>Results</jats:title><jats:p> Intracranial aneurysms were detected with a sensitivity of 79%, a specificity of 79%, a precision of 0.75, a F1 score of 0.77, and a mean area-under-the-curve of 0.76 (range 0.68–0.86) after Monte Carlo cross-validation, run 45 times. </jats:p></jats:sec><jats:sec><jats:title>Conclusion</jats:title><jats:p> The commercial-grade deep learning software allows for detection of intracranial aneurysms on whole-brain, 2D anteroposterior and lateral digital subtraction angiography images, with results being comparable to more specifically engineered deep learning techniques. </jats:p></jats:sec>

An international bibliometric study of scientific articles on intracranial aneurysms

“…The literature review revealed a total number of 21,673 publications. Of all these publications, 13,371 (61.7%) were articles, 2140 (9.9%) were Meeting Abstracts, 1208 (5.6%) were Letters, 965 (4.5%) were Reviews, 716 (3.3%) were Proceedings Papers, and the rest were in the "other publications" category (Editorial Material (3062), Note (388), Correction (93), Book Chapter (55), Early Access (30), News _ Item (29), Reprint (32), Book (1), Discussion (3), Retracted Publication (1), Retraction (1), Bibliography (1)). This study performed the bibliometric analysis of only 13,371 publications that were articles; the other publication types were excluded.…”

“…Intracranial aneurysms are the number one cause of non-traumatic subarachnoid hemorrhage. 1 High mortality and morbidity rates of ruptured intracranial aneurysms are still a serious health problem. 2 The survival rate after aneurysmal subarachnoid hemorrhage (SAH) has increased by 17% in the past few decades, probably owing to better diagnosis, early aneurysms repair, prescription of nimodipine, and advanced intensive care support.…”

#RadialFirst and #RadialForNeuro: A descriptive analysis of Twitter conversations regarding transradial access

“…Despite broad use of #RadialFirst within cardiology and extensive literature supporting transradial access over transfemoral for percutaneous coronary intervention, the vast majority (>80%) of proceduralists in the United States remain low radial adopters, and many recent studies continue to be exclusively femoral access. 27,28 Nonetheless, this analysis provides useful information about topic trends and discussion within neurointerventional surgery and demonstrates a rise in conversation about the transradial approach.…”

Deep Learning-Based Magnetic Resonance Imaging in Diagnosis and Treatment of Intracranial Aneurysm

“…Both the U-Net model and the 3D CNN network use full convolutional networks for image detection. The introduction of deep learning in the labeling process can improve the detection and diagnosis effect of cerebral hemangioma [ 16 – 18 ]. CNN can be divided into 2D convolution and 3D convolution according to the dimension of the convolution kernel.…”