医学
动脉瘤
数字减影血管造影
放射科
血管造影
核医学
科恩卡帕
计算机断层血管造影
前瞻性队列研究
内科学
计算机科学
机器学习
作者
Naying He,Haiying Lyu,Youmin Zhang,Robert K.Y. Li,Zhihan Xu,E. Mark Haacke,Ying Cui,Jiqiang Li,Haipeng Dong,Wentao Han,Rui Chang,Zhen Hu,Chengcheng Zhu,Zhangsheng Yu,Yong Lü,H. X. Jiang,Fuhua Yan
标识
DOI:10.1136/jnis-2025-023094
摘要
Background Accurate detection and morphology evaluation of unruptured intracranial aneurysms (UIAs) are essential for patient management. This study aimed to assess the efficacy of ultra-high-resolution (UHR) photon-counting detector-CT angiography (PCD-CTA) in detecting UIAs and characterizing their morphological features. Methods This prospective study recruited consecutive participants suspected of cerebral vascular diseases who underwent PCD-CTA and subsequent digital subtraction angiography (DSA) within 1 month from September 2023 to May 2024. Performance of UIA diagnosis using UHR PCD-CTA images (slice thickness 0.2 mm) and standard resolution (SR, reflecting clinical protocols) reconstructed images were compared with DSA as reference on a per-participant, per-vessel, and per-aneurysm basis. The inter-rater agreement for UIA detection and aneurysm morphology characterization using UHR/SR PCD-CTA was also evaluated. Results Among 95 participants, 50 UIAs were confirmed in 42 participants using DSA. The inter-rater agreement for UIA detection was: kappa 0.95 for UHR and 0.89 for SR (p<0.05). On a per-aneurysm basis, the sensitivity, specificity, and diagnostic accuracy of UHR (98.0%, 96.7%, 97.3%) were all significantly higher than SR (72.0%, 86.7%, 80.0%) (p<0.05). UHR PCD-CTA accurately identified 13 of the 14 aneurysms (93%) missed by SR PCD-CTA, including 3 cases (21%) larger than 3 mm. Furthermore, UHR identified more aneurysm irregularity (18/50, 36%) than SR (5/50, 10%) (p=0.004). UHR also revealed 4 (8.0%) more cases with wall calcification and 3 (6.0%) more cases with intra-aneurysmal hypointensity (possible thrombus) than SR. Conclusions The advantages of UHR PCD-CTA, including enhanced reliability, improved diagnostic accuracy, and more comprehensive information, have the potential to significantly optimize UIA management.
科研通智能强力驱动
Strongly Powered by AbleSci AI