Introduction:
Treatment of spontaneous intracerebral hemorrhage (ICH) requires rapid, accurate estimation of hemorrhage volume to determine appropriate patient care and guide prognosis. ICH volume estimation on Computed Tomography (CT) imaging using the ABC/2 formula is the clinical gold standard, however this method can be inaccurate, suffers from observer scoring variability, and takes time to make the measurement on a workstation. Semi-Autonomous Segmentation (SAS) is the gold standard for hemorrhage volume estimation, however it is not used clinically due to the increased time for analysis. Recently, artificial intelligence (AI) driven segmentation has been developed (Viz.ai, San Francisco, California) to automatically detect ICH and calculate hematoma volume.
Objective:
Our goal is to validate the accuracy of the Viz.ai ICH segmentation algorithm as a tool for determining hemorrhage volume by comparing its performance to both ABC/2 and SAS.
Methods:
Seventy head CTs positive for ICH were analyzed with SAS in 3D Slicer to determine ICH volume as the standard reference volume. The same CT scans were then analyzed using the ABC/2 method. Finally, scans were uploaded to Viz.ai for ICH volume analysis.
Results:
Compared against standard SAS, Viz.ai ICH volumes were more accurate than ABC/2 in 77% of cases. Average difference between Viz.ai ICH volume and SAS ICH volume was 4.9±4.2 mL (R2=0.98). Average difference between ABC/2 ICH volume and SAS ICH volume was 10.6±11.4 mL (R2=0.77).
Conclusion:
This study indicates that Viz.ai more accurately estimates ICH volume than ABC/2 over a broad range of hematoma volumes when compared to standard SAS, which when coupled with significantly faster analysis compared to SAS justifies the use of AI in ICH triage workflow.
BackgroundMinimally invasive evacuation may help ameliorate outcomes after intracerebral hemorrhage (ICH). However, hospital length of stay (LOS) post-evacuation is often long and costly.ObjectiveTo examine factors associated with LOS in a large cohort of patients who underwent minimally invasive endoscopic evacuation.MethodsPatients presenting to a large health system with spontaneous supratentorial ICH qualified for minimally invasive endoscopic evacuation if they met the following inclusion criteria: age ≥18, premorbid modified Rankin Scale (mRS) score ≤3, hematoma volume ≥15 mL, and presenting National Institutes of Health Stroke Scale (NIHSS) score ≥6. Demographic, clinical, radiographic, and operative characteristics were included in a multivariate logistic regression for hospital and ICU LOS dichotomized into short and prolonged stay at 14 and 7 days, respectively.ResultsAmong 226 patients who underwent minimally invasive endoscopic evacuation, the median intensive care unit and hospital LOS were 8 (4–15) days and 16 (9–27) days, respectively. A greater extent of functional impairment on presentation (OR per NIHSS point 1.10 (95% CI 1.04 to 1.17), P=0.007), concurrent intraventricular hemorrhage (OR=2.46 (1.25 to 4.86), P=0.02), and deep origin (OR=per point 2.42 (1.21 to 4.83), P=0.01) were associated with prolonged hospital LOS. A longer delay from ictus to evacuation (OR per hour 1.02 (1.01 to 1.04), P=0.007) and longer procedure time (OR per hour 1.91 (1.26 to 2.89), P=0.002) were associated with prolonged ICU LOS. Prolonged hospital and ICU LOS were in turn longitudinally associated with a lower rate of discharge to acute rehabilitation (40% vs 70%, P<0.0001) and worse 6-month mRS outcomes (5 (4–6) vs 3 (2–4), P<0.0001).ConclusionsWe present factors associated with prolonged LOS, which in turn was associated with poor long-term outcomes. Factors associated with LOS may help to inform patient and clinician expectations of recovery, guide protocols for clinical trials, and select suitable populations for minimally invasive endoscopic evacuation.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.