Large vessel occlusion detected in 3 minutes by automated CT screening


By Michael Woodhead

11 Nov 2019

An automated machine algorithm for screening of CT angiograms can accurately detect anterior circulation large vessel occlusion (LVO) within 3 minutes, according to neuro-radiologists in Victoria.

While not intended to replace the judgement of a radiologist, the algorithm could help identify high risk stroke patients who could then be triaged and fast tracked for thrombectomy, say researchers from Monash Medical Centre.

And the algorithm would be particularly useful for regional and smaller hospitals with limited radiologist and telehealth resources, to take advantage of the 24 hour window for transfer for thrombectomy, they suggest.

In a study published in Stroke led by Dr Shalini Amukotuwa of the Monash Diagnostic Imaging department, report results from a trial of the automated software-based approach for LVO detection based on in CT angiograms from 926 patients. Of these, 395 patients had an anterior circulation LVO or M2-MCA occlusion.

The average turnaround time for the process was 160 seconds. The sensitivity and specificity of the algorithm were 97% and 74%, for LVO detection, and 95% and 79%, respectively, when M2 occlusions were included.

On analysis by occlusion site, sensitivities were 90% (M2-MCA), 97% (M1-MCA), and 97% (intracranial internal carotid artery) with corresponding area-under-the-ROC-curves of 0.874 (M2), 0.962 (M1), and 0.997 (intracranial internal carotid artery).

Dr Amukotuwa said the lower specificity results and 11 false positive reports should be seen in the context of  the diagnostic capability of neuroradiologists, for whom studies showed can detect LVOs with 89% to 98% sensitivity and 95% to 98% specificity.

“Automation, which does not achieve this high specificity, cannot replace radiologists; rather, its strength and utility lie in the high sensitivity, which allows expedited diagnosis of LVOs by flagging and prioritising these scans as requiring urgent radiologist review,” she wrote.

The algorithm would help expedite early diagnosis in smaller institutions where CT angiograms may be read first by a trainee or general radiologist, particular in the after hours period when there is no access to a neuro-radiologist, said Dr Amukotuwa.

“An automated tool that draws attention to a positive finding would … help avoid situations where emergent CTAs are buried in a worklist and positive findings are communicated to the care providers with substantial delay.”

She said it should also be noted that that the algorithm does not directly detect the clot but rather the resultant loss of vessel opacification, therefore false positives result from chronic occlusions.

“The purpose of this software is to serve as a triage tool that alerts radiologists to a patient with a potential LVO, and in turn trigger evaluation of the patient’s multimodal CT by the human reader who can then use all available information (not just the CTA) to make a judgement call.”

“Precise localisation of the occlusion site and differentiation of chronic occlusions by the algorithm is, therefore, not critical.”

Already a member?

Login to keep reading.

Email me a login link