Commentary on: diagnostic accuracy of a decision-support software for the detection of intracranial large vessel occlusion in CT angiography

In the paper by Androlojc et al. 2023, 1 the authors tested the real-world clinical performance of a commercially available clinical decision-support tool “e-CTA”(Brainomix, Oxford UK) for the detection of intracranial large vessel occlusion (LVO) on CT angiography (CTA) performed on acute stroke patients eligible for mechanical thrombectomy (MT). Since the eminent computer scientist Prof. Geoffrey Hintons’ comments alluding to deep-learning algorithms out performing radiologists in 2016, 2 the computer vision industry has committed significant resources to solving narrow radiological tasks using computation, machine learning, and deep-learning-based models. The e-CTA software automates detection of LVO in the anterior circulation with a view to aiding triage of stroke patients to MT. In reality however, the radiological assessment of a CTA and decision to proceed to mechanical thrombectomy (MT) is