ABSTRACT
OBJECTIVE: To evaluate variability in aneurysm detection and the potential of artificial intelligence (AI) software as a screening tool by comparing conventional computed tomography angiography (CTA) images (standard care) with AI software. METHODS: Neuroradiologists reviewed 770 CTA images and reported the presence or absence of saccular aneurysms. Subsequently, the images were analyzed by AI software. If the software suspected an aneurysm, it flagged the corresponding image. In cases where there was a mismatch between the radiologist's report and the AI findings, an expert neurosurgeon evaluated CTA images providing a definitive conclusion on the presence or absence of an aneurysm. RESULTS: AI software flagged 33 cases as potential aneurysms; 16 cases were positively identified as aneurysms by radiologists, and 17 were dismissed. A total of 737 cases were considered negative by AI software, while in the same group, radiologists identified aneurysms in 28 CTA images. Compared with the radiologist's report, AI performance had a sensitivity of 36%, specificity of 97.6%, and negative predictive value of 96.2%. There were 45 mismatch cases between AI and radiologists. AI flagged 17 images as showing an aneurysm that was unreported by radiologists; the expert neurosurgeon confirmed that 7 of the 17 images showed an aneurysm. In 28 images considered negative by AI, radiologists indicated aneurysms; 17 of those confirmed by the neurosurgeon. CONCLUSIONS: AI has the potential to increase the diagnosis of unruptured intracranial aneurysms. However, it must be used as an adjacent tool within the standard of care due to limited applicability in real-world settings.
ABSTRACT
BACKGROUND: We analyzed the effect of specific optimization steps to reduce treatment delays in a nonacademic stroke hospital setting. METHODS: The data from patients with ischemic stroke who had been treated with intravenous tissue plasminogen activator or endovascular therapy, or both, were analyzed. The metrics were divided into 2 periods: preoptimization period (October 1, 2015 to September 30, 2016) and postoptimization period (October 1, 2016 to September 30, 2017). The key interventions were 1) notification by the emergency medical service to the emergency department and stroke team; 2) division of the stroke alert between level 1 (intravenous/intra-arterial candidate) and level 2; 3) direct transportation of level 1 patients to brain computed tomography; 4) limitation of nonessential interventions; 5) stroke orientation; 6) 24-hour, 7-day code stroke response by a vascular neurologist; 7) earlier notification of the interventional radiology team; 8) direct transportation from computed tomography to angiography suite for large vessel occlusion; and 9) multidisciplinary monthly meetings to discuss delayed cases. RESULTS: A total of 279 patients were identified. No significant differences in any of the baseline characteristics were documented. Almost all metrics favored the postoptimization period, with remarkable improvement in the door-to-puncture time (median, 64 minutes; interquartile range, 36-86; vs. 47 minutes; interquartile range, 20-62; P = 0.001). We observed an increased percentage of good clinical outcomes in the postoptimization group (60.1% vs. 54.8%; P = 0.500). We found an 8.4% increase in patients with good clinical outcomes in the postoptimization group compared with our previously reported work. CONCLUSIONS: For acute reperfusion therapies, significant reductions in workflow intervals can be achieved after simple optimization methods in a nonacademic community-based hospital.
