Clinically Applicable Deep Learning for Intracranial Aneurysm Detection in Computed Tomography Angiography Images: A Comprehensive Multicohort Study

Abstract

Intracranial aneurysm is a common life-threatening disease. CTA is recommended as a standard diagnosis tool, while the interpretation is time-consuming and challenging. We presented a novel deep-learning-based framework trained on 1,177 DSA verified bone-removal CTA cases. The framework had excellent tolerance to the influence of occult cases of CTA-negative but DSA-positive aneurysms, image quality, and manufacturers. Simulated real-world studies were conducted in consecutive internal and external cohorts, achieving improved sensitivity and negative predictive value than radiologists. A specific cohort of suspected acute ischemic stroke was employed and found 96.8% predicted-negative cases can be trusted with high confidence, leading to reducing in human burden. A prospective study is warranted to determine whether the algorithm could improve patients' care in comparison to radiologists' assessment.