Background and purpose Endovascular thrombectomy is an evidence‐based treatment for large vessel occlusion (LVO) stroke. Commercially available artificial intelligence has been designed to detect the presence of an LVO on computed tomography angiogram (CTA). We compared Viz.ai‐LVO (San Francisco, CA, USA) to CTA interpretation by board‐certified neuroradiologists (NRs) in a large, integrated stroke network. Methods From January 2021 to December 2021, we compared Viz.ai detection of an internal carotid artery (ICA) or middle cerebral artery first segment (MCA‐M1) occlusion to the gold standard of CTA interpretation by board‐certified NRs for all code stroke CTAs. On a monthly basis, sensitivity, specificity, accuracy, positive predictive value (PPV), and negative predictive value (NPV) were calculated. Trend analyses were conducted to evaluate for any improvement of LVO detection by the software over time. Results 3851 patients met study inclusion criteria, of whom 220 (5.7%) had an ICA or MCA‐M1 occlusion per NR. Sensitivity and specificity were 78.2% (95% CI 72%–83%) and 97% (95% CI 96%–98%), respectively. PPV was 61% (95% CI 55%–67%), NPV 99% (95% CI 98%–99%), and accuracy was 95.9% (95% CI 95.3%–96.5%). Neither specificity or sensitivity improved over time in the trend analysis. Conclusions Viz.ai‐LVO has high specificity and moderately high sensitivity to detect an ICA or proximal MCA occlusion. The software has the potential to streamline code stroke workflows and may be particularly impactful when emergency access to NRs or vascular neurologists is limited.

BACKGROUND: Distinguishing features of our stroke network include routine involvement of a telestroke nurse (TSRN) for code stroke activations at nonthrombectomy centers and immediate availability of neuroradiologists for imaging interpretation. On May 1, 2021, we implemented a new workflow for code stroke activations presenting beyond 4.5 hours from last known well that relied on a TSRN supported by a neuroradiologist for initial triage. Patients without a target large vessel occlusion (LVO) were managed without routine involvement of a teleneurologist, which represented a change from the preimplementation period. METHODS: We collected data 6 months before and after implementation of the new workflow. We compared preimplementation process metrics for patients managed with teleneurologist involvement with the postimplementation patients managed without teleneurologist involvement. RESULTS: With the new workflow, teleneurologist involvement decreased from 95% (n = 953) for patients presenting beyond 4.5 hours from last known well to 37% (n = 373; P < .001). Compared with patients in the preimplementation period, postimplementation patients without teleneurologist involvement experienced less inpatient hospital admission and observation (87% vs 90%; unadjusted P = .038, adjusted P = .06). Among the preimplementation and postimplementation admitted patients, there was no statistically significant difference in follow-up neurology consultation or nonstroke diagnoses. A similar percentage of LVO patients were transferred to the thrombectomy center (54% pre vs 49% post, P = .612), whereas more LVO transfers in the postimplementation cohort received thrombectomy therapy (75% post vs 39% pre, P = .014). Among LVO patients (48 pre and 41 post), no statistical significance was observed in imaging and management times. CONCLUSION: Our work shows the successful teaming of a TSRN and a neuroradiologist to triage acute stroke patients who present beyond an eligibility window for systemic thrombolysis, without negatively impacting care and process metrics. This innovative partnering may help to preserve the availability of teleneurologists by limiting their involvement when diagnostic imaging drives decision making.

No abstract