Kang Dong-Wan, Kim Museong, Park Gi-Hun, Kim Yong Soo, Han Moon-Ku, Lee Myungjae, Kim Dongmin, Ryu Wi-Sun, Jeong Han-Gil
Division of Intensive Care Medicine, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam, Republic of Korea.
Department of Neurosurgery, Seoul National University Bundang Hospital, Seoul National University College of Medicine, Seongnam, Republic of Korea.
Neuroradiology. 2025 Mar 21. doi: 10.1007/s00234-025-03560-x.
Intracranial hemorrhage (ICH) requires urgent treatment, and accurate and timely diagnosis is essential for improving outcomes. This pivotal clinical trial aimed to validate a deep learning algorithm for ICH detection and assess its clinical utility through a reader performance test.
Retrospective CT scans from patients with and without ICH were collected from a tertiary hospital. Two experts evaluated all scans, with a third expert reviewing disagreements for the final diagnosis. We analyzed the performance of the deep learning algorithm, JLK-ICH, for all cases and ICH subtypes. Additional external validation was performed using a multi-ethnic U.S.
A reader performance study included six non-expert readers who evaluated 800 CT scans, with and without JLK-ICH assistance, following a washout period. ICH presence and five-point scale confidence level for decisions were rated.
A total of 1,370 CT scans were evaluated. The deep learning model showed 98.7% sensitivity (95% confidence interval [CI] 97.8-99.3%), 88.5% specificity (95% CI, 83.6-92.3%), and an area under the receiver operating characteristic curve (AUROC) of 0.936 (95% CI, 0.915-0.957). The model maintained high accuracy across all ICH subtypes, and additional external validation confirmed these results. In the reader performance study, AUROC with JLK-ICH assistance (0.967 [0.953-0.981]) surpassed that without assistance (0.953 [0.938-0.957]; P = 0.009). JLK-ICH particularly improved performance when readers were highly uncertain.
The JLK-ICH algorithm demonstrated high accuracy in detecting all ICH subtypes. Non-expert readers significantly improved diagnostic accuracy for brain CT scans with deep learning assistance.
颅内出血（ICH）需要紧急治疗，准确及时的诊断对于改善预后至关重要。这项关键的临床试验旨在验证一种用于ICH检测的深度学习算法，并通过读者性能测试评估其临床效用。
从一家三级医院收集有和没有ICH患者的回顾性CT扫描图像。两名专家评估所有扫描图像，第三名专家审查分歧以做出最终诊断。我们分析了深度学习算法JLK-ICH对所有病例和ICH亚型的性能。使用多民族的美国数据集进行了额外的外部验证。
一项读者性能研究包括六名非专家读者，他们在洗脱期后在有和没有JLK-ICH辅助的情况下评估了800例CT扫描图像。对ICH的存在情况和决策的五点量表置信水平进行评分。
共评估了1370例CT扫描图像。深度学习模型显示出98.7%的敏感性（95%置信区间[CI]97.8-99.3%），88.5%的特异性（95%CI，83.6-92.3%），以及受试者操作特征曲线下面积（AUROC）为0.936（95%CI，0.915-0.957）。该模型在所有ICH亚型中均保持较高的准确性，额外的外部验证证实了这些结果。在读者性能研究中，有JLK-ICH辅助时的AUROC（0.967[0.953-0.981]）超过了无辅助时的AUROC（0.953[0.938-0.957]；P = 0.009）。当读者高度不确定时，JLK-ICH尤其提高了性能。
JLK-ICH算法在检测所有ICH亚型方面显示出高准确性。非专家读者在深度学习辅助下显著提高了脑CT扫描的诊断准确性。
