Study shows accuracy of AI for detecting brain haemorrhage in non-contrast head CT

Provider of teleradiology services Medica Reporting has collaborated with Qure.ai on a study that highlights the effectiveness of AI in detecting intracranial haemorrhage (ICH) in non-contrast CT head scans.

Titled ‘A retrospective audit of an artificial intelligence software for the detection of intracranial haemorrhage used by a tele-radiology company in the UK’, the study was published in the British Journal of Radiology.

This research evaluated the performance of Qure.ai’s AI-based software across a subset of 1,315 CT head scans from patients referred for urgent reporting across 44 UK hospitals. The results showed a strong overall agreement between the AI and radiologists in detecting ICH, with a notable accuracy rate of 93.5 per cent.

The study also revealed insights into scenarios where AI can either miss subtle haemorrhages or generate false positives due to motion artefacts.

“AI continues to show great promise in assisting radiologists with timely and accurate diagnoses, and this study demonstrates its real-world applicability,” said co-author and lead investigator Dr Garry Pettet. “This collaboration and study with Qure.ai reflects our commitment to improving lives through excellence in diagnostics and research.”

Read this report on page 24 of the November 2024 issue of RAD Magazine.