Intracranial hemorrhage (ICH) has high morbidity and mortality with nearly 50% 30 day mortality for patients admitted to the ICU and as few as 20% of survivors demonstrating full neurologic recovery. Early intervention has been shown to improve clinical outcomes. Given high mortality and morbidity, prompt identification of ICH has high clinical utility. Several applications have emerged using artificial intelligence (AI) for automated detection of ICH, including RAPID ICH (iSchemaView, Menlo Park, CA). We present our initial clinical experience with RAPID ICH in a busy Level 1 trauma center.*Methods and Materials The study was performed under the supervision of the local institutional review board.Patients presenting to the emergency department (ED) receiving CT scans of the brain and inpatients (IP) scanned on the ED scanner at one level 1 trauma center were included in the study. The RAPID ICH output ("no ICH" or "suspected ICH") was recorded for each study. The initial interpreting emergency radiologist or neuroradiologist had access to the RAPID ICH output. Radiology reports reporting ICH were considered positive and those reporting no ICH were considered negative. A board certified neuroradiologist reviewed each case who had access to the initial report, RAPID ICH output, and all subsequent examinations. In cases with disagreement between the readers, a third reader adjudicated the result and their decision was considered final. The expert reads were used as the gold standard and sensitivity, specificity, negative predictive value (NPV) and positive predictive value (PPV) were calculated.*Results A total of 1388 patients were included in the study, 1251 from the ED and 137 IP. For the ED patients, 139 had ICH and 1112 did not, where as in the IP cohort, 100 had ICH and 37 did not. RAPID ICH demonstrated overall sensitivity of 79% (73% ED, 88% IP), overall specificity of 95% (96% ED, 84% IP), overall PPV of 76% (65% ED, 94% IP) and overall NPV of 96% (97% ED, 72% IP).*Conclusions RAPID ICH demonstrated relatively high sensitivity and very high specificity for ICH, with lower sensitivity and higher specificity in the ED, and higher sensitivity but lower specificity in the inpatient setting. Using AI applications such as RAPID ICH for active worklist reprioritization may allow timely triage of potentially positive studies, allowing early intervention and potentially leading to improved clinical outcomes, especially in the inpatient setting with longer average turnaround times for routine studies.*Clinical Relevance/Application RAPID ICH demonstrated relatively high sensitivity and specificity in both IP and ED settings. Active worklist reprioritization using RAPID ICH may improve turnaround times and allow earlier intervention and improved clinical outcomes.

A total of 1388 patients were included in the study, 1251 from the ED and 137 IP. For the ED patients, 139 had ICH and 1112 did not, where as in the IP cohort, 100 had ICH and 37 did not. RAPID ICH demonstrated overall sensitivity of 79% (73% ED, 88% IP), overall specificity of 95% (96% ED, 84% IP), overall PPV of 76% (65% ED, 94% IP) and overall NPV of 96% (97% ED, 72% IP).

RAPID ICH demonstrated relatively high sensitivity and specificity in both IP and ED settings. Active worklist reprioritization using RAPID ICH may improve turnaround times and allow earlier intervention and improved clinical outcomes.