Accuracy of Automated Computer-Aided Diagnosis for Stroke Imaging: A Critical Evaluation of Current Evidence

Abstract

There is increasing interest in computer applications, using artificial intelligence methodologies, to perform health care tasks previously performed by humans, particularly in medical imaging for diagnosis. In stroke, there are now commercial artificial intelligence software for use with computed tomography or MR imaging to identify acute ischemic brain tissue pathology, arterial obstruction on computed tomography angiography or as hyperattenuated arteries on computed tomography, brain hemorrhage, or size of perfusion defects. A rapid, accurate diagnosis may aid treatment decisions for individual patients and could improve outcome if it leads to effective and safe treatment; or conversely, to disaster if a delayed or incorrect diagnosis results in inappropriate treatment. Despite this potential clinical impact, diagnostic tools including artificial intelligence methods are not subjected to the same clinical evaluation standards as are mandatory for drugs. Here, we provide an evidence-based review of the pros and cons of commercially available automated methods for medical imaging diagnosis, including those based on artificial intelligence, to diagnose acute brain pathology on computed tomography or magnetic resonance imaging in patients with stroke.

Predictive value of clot imaging in acute ischemic stroke: A systematic review of artificial intelligence and conventional studies

“…Various AI methods were developed based on information provided by computed tomography (CT) and magnetic resonance (MR) neuroimaging. The emerging applications of AI in stroke are examined in numerous recent reviews [4] [5] [6] [7] [8] [9] [10] [11] [12]. Often covered are AI developments that aim to differentiate between hemorrhagic stroke, ischemic stroke and stroke mimics, measure the Alberta Stroke Program Early CT Score (ASPECTS), detect and segment parenchyma, analyze the collateral flow status, detect large vessel occlusion (LVO) from computed tomography angiography (CTA), or analyze biomarkers of corticomotor structure and func-tion.…”

“…The image-based AI methods aim to have a powerful impact in clinical management of stroke, by improving the diagnostic, predictive and prognostic value of clinical neuroimaging. Various AI software received FDA (Food and Drug Administration) approval and CE (Conformité Européene) mark, although more evidence is necessary to prove in what degree they are clinically beneficial [12].…”

Diagnostic Performance of Artificial Intelligence-Based Computer-Aided Detection and Diagnosis in Pediatric Radiology: A Systematic Review

“…DL is a subset of ML, and its models have more layers than those of ML. The DL algorithms are capable of modeling high-level abstractions in medical images without predetermined inputs [ 5 , 8 , 9 ].…”

Accuracy of artificial intelligence software for CT angiography in stroke

“…However, in this rapidly evolving field of radiology AI, standardised testing methods do not yet exist, and peer-reviewed evidence of AI software efficacy is only beginning to emerge. 5 e-CTA is one software application developed using AI to automate the identification of distal ICA or proximal MCA occlusion and to score anterior circulation arterial collaterals on CTA. The software is cleared for clinical use (currently only for collateral quantification in the US), but there is limited peer-reviewed testing of e-CTA.…”

“…Software developed using artificial intelligence (AI) designed to assist clinicians interpret stroke CTA and detect LVO are increasingly available. However, in this rapidly evolving field of radiology AI, standardised testing methods do not yet exist, and peer‐reviewed evidence of AI software efficacy is only beginning to emerge 5 …”