Automated detection and segmentation of intracranial hemorrhage suspect hyperdensities in non-contrast-enhanced CT scans of acute stroke patients

Objectives: Artificial intelligence (AI)–based image analysis is increasingly used in acute stroke care. Its application for detecting and quantifying hemorrhage-suspect hyperdensities in non-contrast-enhanced head CT (NCCT) scans may aid clinical decision-making and expedite stroke management.

Methods: NCCT scans from 160 patients with suspected acute stroke were analyzed for acute intracranial hemorrhages (ICH) using a novel AI-based algorithm. Two blinded neuroradiology residents independently assessed the scans, with an expert neuroradiologist establishing ground truth. Sensitivity, specificity, and AUC were calculated for ICH and intraparenchymal hemorrhage (IPH) detection, and the algorithm automatically segmented and quantified IPH volumes (ICC and Dice coefficient).

Results: Of 160 patients, 79 had acute ICH and 47 had IPH. For ICH detection, the algorithm achieved sensitivity 0.91 and specificity 0.89 (readers 0.99–1.00 / 0.98). For IPH detection, the algorithm achieved sensitivity 0.98 and specificity 0.89. ICC showed excellent agreement (0.98) for IPH volumes, with a mean Dice coefficient of 0.82.

Conclusion: The AI-based algorithm reliably detected acute ICH and accurately quantified IPH volumes in this dataset.