Hemorrhage Enlargement Is More Frequent in the First 2 Hours: A Prehospital Mobile Stroke Unit Study.

BACKGROUND: Hematoma enlargement (HE) after intracerebral hemorrhage (ICH) is a therapeutic target for improving outcomes. Hemostatic therapies to prevent HE may be more effective the earlier they are attempted. An understanding of HE in first 1 to 2 hours specifically in the prehospital setting would help guide future treatment interventions in this time frame and setting. METHODS: Patients with spontaneous ICH within 4 hours of symptom onset were prospectively evaluated between May 2014 and April 2020 as a prespecified substudy within a multicenter trial of prehospital mobile stroke unit versus standard management. Baseline computed tomography scans obtained <1, 1 to 2, and 2 to 4 hours postsymptom onset on the mobile stroke unit in the prehospital setting were compared with computed tomography scans repeated 1 hour later and at 24 hours in the hospital. HE was defined as >6 mL if baseline ICH volume was <20 mL and 33% increase if baseline volume >20 mL. The association between time from symptom onset to baseline computed tomography (hours) and HE was investigated using Wilcoxon rank-sum test when time was treated as a continuous variable and using Fisher exact test when time was categorized. Kruskal-Wallis and Wilcoxon rank-sum tests evaluated differences in baseline volumes and HE. Univariable and multivariable logistic regression analyses were conducted to identify factors associated with HE and variable selection was performed using cross-validated L1-regularized (Lasso regression). This study adhered to STROBE guidelines (Strengthening the Reporting of Observational Studies in Epidemiology) for cohort studies. RESULTS: One hundred thirty-nine patients were included. There was no difference between baseline ICH volumes obtained <1 hour (n=43) versus 1 to 2 hour (n=51) versus >2 hours (n=45) from symptom onset (median [interquartile range], 13 mL [6-24] versus 14 mL [6-30] versus 12 mL [4-19]; P=0.65). However, within the same 3 time epochs, initial hematoma growth (volume/time from onset) was greater with earlier baseline scanning (median [interquartile range], 17 mL/hour [9-35] versus 9 mL/hour [5-23]) versus 4 mL/hour [2-7]; P<0.001). Forty-nine patients had repeat scans 1 hour after baseline imaging (median, 2.3 hours [interquartile range. 1.9-3.1] after symptom onset). Eight patients (16%) had HE during that 1-hour interval; all of these occurred in patients with baseline imaging within 2 hours of onset (5/18=28% with baseline imaging within 1 hour, 3/18=17% within 1-2 hour, 0/13=0% >2 hours; P=0.02). HE did not occur between the scans repeated at 1 hour and 24 hours. No association between baseline variables and HE was detected in multivariable analyses. CONCLUSIONS: HE in the next hour occurs in 28% of ICH patients with baseline imaging within the first hour after symptom onset, and in 17% of those with baseline imaging between 1 and 2 hours. These patients would be a target for ultraearly hemostatic intervention.

Machine Learning Automated Detection of Large Vessel Occlusion From Mobile Stroke Unit Computed Tomography Angiography.

BACKGROUND: Prehospital automated large vessel occlusion (LVO) detection in Mobile Stroke Units (MSUs) could accelerate identification and treatment of patients with LVO acute ischemic stroke. Here, we evaluate the performance of a machine learning (ML) model on CT angiograms (CTAs) obtained from 2 MSUs to detect LVO. METHODS: Patients evaluated on MSUs in Houston and Los Angeles with out-of-hospital CTAs were identified. Anterior circulation LVO was defined as an occlusion of the intracranial internal carotid artery, middle cerebral artery (M1 or M2), or anterior cerebral artery vessels and determined by an expert human reader. A ML model to detect LVO was trained and tested on independent data sets consisting of in-hospital CTAs and then tested on MSU CTA images. Model performance was determined using area under the receiver-operator curve statistics. RESULTS: Among 68 patients with out-of-hospital MSU CTAs, 40% had an LVO. The most common occlusion location was the middle cerebral artery M1 segment (59%), followed by the internal carotid artery (30%), and middle cerebral artery M2 (11%). Median time from last known well to CTA imaging was 88.0 (interquartile range, 59.5-196.0) minutes. After training on 870 in-hospital CTAs, the ML model performed well in identifying LVO in a separate in-hospital data set of 441 images with area under receiver-operator curve of 0.84 (95% CI, 0.80-0.87). ML algorithm analysis time was under 1 minute. The performance of the ML model on the MSU CTA images was comparable with area under receiver-operator curve 0.80 (95% CI, 0.71-0.89). There was no significant difference in performance between the Houston and Los Angeles MSU CTA cohorts. CONCLUSIONS: In this study of patients evaluated on MSUs in 2 cities, a ML algorithm was able to accurately and rapidly detect LVO using prehospital CTA acquisitions.

Mobile Stroke Unit Computed Tomography Angiography Substantially Shortens Door-to-Puncture Time.

Background and Purpose- Endovascular thrombectomy (ET) door-to-puncture time (DTPT) is a modifiable metric. One of the most important, yet time-consuming steps, is documentation of large vessel occlusion by computed tomography angiography (CTA). We hypothesized that obtaining CTA on board a Mobile Stroke Unit and direct alert of the ET team shortens DTPT by over 30 minutes. Methods- We compared DTPT between patients having CTA onboard the Mobile Stroke Unit then subsequent ET from September 2018 to November 2019 and patients in Mobile Stroke Unit from August 2014 to August 2018, when onboard CTA was not yet being used. We also correlated DTPT with change in National Institutes of Health Stroke Scale between baseline and 24 hours. Results- Median DTPT was 53.5 (95% CI, 35-67) minutes shorter with onboard CTA and direct ET team notification: 41 minutes (interquartile range, 30.0-63.5) versus 94.5 minutes (interquartile range, 69.8-117.3; P<0.001). Median on-scene time was 31.5 minutes (interquartile range, 28.8-35.5) versus 27.0 minutes (interquartile range, 23.0-31.0) (P<0.001). Shorter DTPT correlated with greater improvement of National Institutes of Health Stroke Scale (correlation=-0.2, P=0.07). Conclusions- Prehospital Mobile Stroke Unit management including on-board CTA and ET team alert substantially shortens DTPT. Registration- URL: https://clinicaltrials.gov; Unique identifier: NCT02190500.