Computer-Aided Diagnosis Improves Detection of Small Intracranial Aneurysms on MRA in a Clinical Setting

Abstract

BACKGROUND AND PURPOSE:MRA is widely accepted as a noninvasive diagnostic tool for the detection of intracranial aneurysms, but detection is still a challenging task with rather low detection rates. Our aim was to examine the performance of a computer-aided diagnosis algorithm for detecting intracranial aneurysms on MRA in a clinical setting.

Deep neural network‐based computer‐assisted detection of cerebral aneurysms in MR angiography

“…We adopted this approach by employing an MIP method to obtain 2D representations. Small aneurysms are sometimes difficult to detect in MIP images because of vessel overlap . We applied an MIP method not to the whole MRA volume but to small image patches, which might decrease the effect of vessel overlap.…”

“…However, most systems produce far more false positives (FPs) than radiologists to achieve such sensitivities . For example, for intracranial aneurysm detection in MRA, radiologists have been reported to have sensitivities of 64–70% in clinical settings . On the other hand, recently reported CAD systems evaluated using large datasets generate about 0.8 or more FPs per case (FPs/case) at a sensitivity of 70%.…”

Blob Enhancement and Visualization for Improved Intracranial Aneurysm Detection

“…This is because all these filters are proportional to the response of one or more Hessian eigenvalues, which makes them sensitive to size and intensity (contrast) of blobs. Conversely, the proposed filter is a ratio of Hessian eigenvalues (8) and enhances the blob-like structures independently of their size or intensity. Another advantage of the proposed blob enhancement filter is its nonlinear response characteristic (Fig.…”

“…For small aneurysms, the sensitivity of the detection based on CTA and MRA images is still lower compared to DSA [5], [6]. Although currently the detection of aneurysms is performed visually by experienced neuroradiologists, especially the detection sensitivity of smaller aneurysms is not satisfactory [7].Štepán-Buksakowska et al recently showed that computer-aided diagnosis improves the detection of small intracranial aneurysms [8]. Enhancement of possible aneurysm locations in cerebrovascular images may shorten visual inspection and increase the detection sensitivity by directing the attention of a diagnostician to suspicious locations in the cerebral vasculature.…”

Aneurysm Recurrence Volumetry Is More Sensitive than Visual Evaluation of Aneurysm Recurrences

“…Recently, an automated computer-aided method for aneurysm screening was reported to improve the detection rate for small intracranial aneurysm [38]. Additional applications for computer-aided techniques could be the use of volumetric methods like ARV to detect growth of untreated aneurysms early on.…”