Evaluating an AI Model for Detecting Intracranial Aneurysms in Head CT Angiography and Its Impact on Diagnosis and Patient Outcomes in a Double-Blind Multicenter Trial

What this Trial Is About

Researchers are conducting the IDEAL study, a large, multicenter, prospective, double-blind randomized controlled trial, to evaluate the impact of an artificial intelligence AI model on diagnosing intracranial aneurysms using head CT angiography CTA. The study aims to compare the diagnostic sensitivity and specificity of this AI-assisted approach against a sham AI method, while also assessing patient prognosis and various clinical outcomes. Over 6450 patients scheduled for head CTA will be enrolled across 21 hospitals in China. Participants will be randomly assigned to one of two groups the True-AI group, where radiologists use a deep-learning AI model with high sensitivity to assist in diagnosing intracranial aneurysms, and the Sham-AI group, where radiologists use a similar AI model with very low sensitivity but comparable specificity. Both groups will undergo their usual head CTA scans and diagnoses. The study includes a 12-month follow-up to monitor repeat imaging, clinical follow-up, treatment, and patient outcomes. During the study, patients will have their head CTA images interpreted with either the True-AI or Sham-AI assistance, without knowledge of group assignment by patients or radiologists. Researchers will assess diagnostic accuracy, detection rates of intracranial lesions, workload of CTA interpretation, resource use, hospitalization rates, quality of life measures, functional outcomes, and safety indicators at 3, 6, and 12 months. The primary outcomes focus on diagnostic sensitivity and specificity at 6 months, with secondary outcomes covering broader clinical and patient-centered measures over the follow-up period.

IDEAL Study: Blinded RCT for the Impact of AI Model for Cerebral Aneurysms Detection on Patients' Diagnosis and Outcomes