hStroke_Suite CT Module is an AI-powered solution for emergency neuroimaging that enables rapid and accurate detection of intracranial hemorrhages in head CT scans, featuring a mobile application and smartwatch integration for real-time review, collaboration, and continuous connectivity for medical professionals.
Assists in detecting the presence, location, and subtypes of hemorrhages (including subdural hematoma (SDH), epidural hematoma (EDH), intraparenchymal hemorrhage (IPH), intraventricular hemorrhage (IVH), and subarachnoid hemorrhage (SAH)) in non-contrast head CT scans.
Non-Contrast head CT, Axial plane, Slice thickness of ranging from 0.75 to 5 mm.
Secondary Capture Report, CT Segmentation.
The joint CNN-RNN with an attention mechanism yielded a diagnostic accuracy of 98.26% (95% CI 98.14–98.37%) with correctly classifying 49,101 out of 49,968 head CT scans on the training set. The sensitivity, specificity, and AUROC of the model on the training set was 97.72% (95% CI 97.58–97.85%), 98.49% (95% CI 98.42–98.55%), and 0.992 (95% CI 0.991–0.993), respectively. The model achieved a diagnostic accuracy of 99.41% (95% CI 99.51–99.84%) with correctly classifying 5180 out of 5211 head CT scans on the validation set. The sensitivity, specificity, and AUROC of the model on the validation set was 99.70% (95% CI 99.51–99.84%), 99.34% (95% CI 99.09–99.58%), and 0.998 (95% CI 0.998–0.999), respectively.
%35 decrease in reading time with hStroke_Suite CT Module.
