Deep Learning Based Software to Identify Large Vessel Occlusion on Noncontrast Computed Tomography.

BACKGROUND AND PURPOSE: Reliable recognition of large vessel occlusion (LVO) on noncontrast computed tomography (NCCT) may accelerate identification of endovascular treatment candidates. We aim to validate a machine learning algorithm (MethinksLVO) to identify LVO on NCCT. METHODS: Patients with suspected acute stroke who underwent NCCT and computed tomography angiography (CTA) were included. Software detection of LVO (MethinksLVO) on NCCT was tested against the CTA readings of 2 experienced radiologists (NR-CTA). We used a deep learning algorithm to identify clot signs on NCCT. The software image output trained a binary classifier to determine LVO on NCCT. We studied software accuracy when adding National Institutes of Health Stroke Scale and time from onset to the model (MethinksLVO+). RESULTS: From 1453 patients, 823 (57%) had LVO by NR-CTA. The area under the curve for the identification of LVO with MethinksLVO was 0.87 (sensitivity: 83%, specificity: 71%, positive predictive value: 79%, negative predictive value: 76%) and improved to 0.91 with MethinksLVO+ (sensitivity: 83%, specificity: 85%, positive predictive value: 88%, negative predictive value: 79%). CONCLUSIONS: In patients with suspected acute stroke, MethinksLVO software can rapidly and reliably predict LVO. MethinksLVO could reduce the need to perform CTA, generate alarms, and increase the efficiency of patient transfers in stroke networks.

Systematic CT perfusion acquisition in acute stroke increases vascular occlusion detection and thrombectomy rates.

BACKGROUND: In patients with stroke, current guidelines recommend non-invasive vascular imaging to identify intracranial vessel occlusions (VO) that may benefit from endovascular treatment (EVT). However, VO can be missed in CT angiography (CTA) readings. We aim to evaluate the impact of consistently including CT perfusion (CTP) in admission stroke imaging protocols. METHODS: From April to October 2020 all patients admitted with a suspected acute ischemic stroke underwent urgent non-contrast CT, CTA and CTP and were treated accordingly. Hypoperfusion areas defined by time-to-maximum of the tissue residue function (Tmax) >6 s, congruent with the clinical symptoms and a vascular territory, were considered VO (CTP-VO). In addition, two experienced neuroradiologists blinded to CTP but not to clinical symptoms retrospectively evaluated non-contrast CT and CTA to identify intracranial VO (CTA-VO). RESULTS: Of the 338 patients included in the analysis, 157 (46.5%) presented with CTP-VO (median Tmax >6s: 73 (29-127) mL). CTA-VO was identified in 83 (24.5%) of the cases. Overall CTA-VO sensitivity for the detection of CTP-VO was 50.3% and specificity was 97.8%. Higher hypoperfusion volume was associated with increased CTA-VO detection (OR 1.03; 95% CI 1.02 to 1.04). EVT was performed in 103 patients (30.5%; Tmax >6s: 102 (63-160) mL), representing 65.6% of all CTP-VO. Overall CTA-VO sensitivity for the detection of EVT-VO was 69.9% and specificity was 95.3%. Among patients who received EVT, the rate of false negative CTA-VO was 30.1% (Tmax >6s: 69 (46-99.5) mL). CONCLUSION: Systematically including CTP in acute stroke admission imaging protocols may increase the diagnosis of VO and rate of EVT.