Abstract P137: Large Vessel Occlusion Prediction Scale Thresholds That Are Sensitive for DAWN Trial Participants

Introduction: Large Vessel Occlusion (LVO) prediction scales are used to triage prehospital suspected stroke patients with a high probability of LVO stroke to endovascular therapy centers. The sensitivities of these scales in the 6 to 24 hour window are unknown. Higher scale score thresholds are typically less sensitive and more specific. Knowing the highest scale score thresholds that remain sensitive could inform threshold selection for clinical use. Sensitivities may also vary between left and right sided LVOs. Methods: LVO prediction scale scores were retrospectively calculated using the NIHSS sub-item scores of patients enrolled in the DAWN Trial. All patients had last known well times between 6 to 24 hours, NIHSS scores ≥ 10, intracranial ICA or proximal MCA occlusions, and mismatches between their exam severities and infarct core volumes. Scale thresholds with sensitivities ≥ 85% were identified. Scores ≥ 5% more sensitive for left or right sided LVOs were identified. Specificities could not be calculated because all DAWN Trial patients had LVOs. Results: 201 out of 206 patients had the required NIHSS sub-item scores. The highest score thresholds that maintained sensitivities ≥ 85% are bolded in the table. Conclusions: CPSS = 3, C-STAT ≥ 2, FAST-ED ≥ 4, G-FAST ≥ 3, RACE ≥ 5, and SAVE ≥ 3 are likely the highest thresholds that can be selected for extended window LVO triage without missing more than 15% of DAWN Trial eligible LVO strokes. For CPSS and SAVE, these are higher than the thresholds suggested by prior studies. CPSS = 3 and RACE ≥ 5 were more sensitive for right sided LVOs. These findings represent the maximum anticipated sensitivities of LVO prediction scales since the NIHSS scores were documented in hospitals during a clinical trial rather than in the prehospital setting. Inclusion of lower NIHSS or more distal LVO patients would lower sensitivities further. Selecting even higher scale thresholds for LVO triage would lead to many missed LVO strokes.

Feasibility study for use of angiographic parametric imaging and deep neural networks for intracranial aneurysm occlusion prediction

BackgroundAngiographic parametric imaging (API), based on digital subtraction angiography (DSA), is a quantitative imaging tool that may be used to extract contrast flow parameters related to hemodynamic conditions in abnormal pathologies such as intracranial aneurysms (IAs).ObjectiveTo investigate the feasibility of using deep neural networks (DNNs) and API to predict IA occlusion using pre- and post-intervention DSAs.MethodsWe analyzed DSA images of IAs pre- and post-treatment to extract API parameters in the IA dome and the corresponding main artery (un-normalized data). We implemented a two-step correction to account for injection variability (normalized data) and projection foreshortening (relative data). A DNN was trained to predict a binary IA occlusion outcome: occluded/unoccluded. Network performance was assessed with area under the receiver operating characteristic curve (AUROC) and classification accuracy. To evaluate the effect of the proposed corrections, prediction accuracy analysis was performed after each normalization step.ResultsThe study included 190 IAs. The mean and median duration between treatment and follow-up was 9.8 and 8.0 months, respectively. For the un-normalized, normalized, and relative subgroups, the DNN average prediction accuracies for IA occlusion were 62.5% (95% CI 60.5% to 64.4%), 70.8% (95% CI 68.2% to 73.4%), and 77.9% (95% CI 76.2% to 79.6%). The average AUROCs for the same subgroups were 0.48 (0.44–0.52), 0.67 (0.61–0.73), and 0.77 (0.74–0.80).ConclusionsThe study demonstrated the feasibility of using API and DNNs to predict IA occlusion using only pre- and post-intervention angiographic information.