文献速递神经MR

商用深度学习算法在非综合卒中中心疑似卒中患者脑MRI缺血性病灶检测中的诊断准确性研究

Diagnostic test accuracy study of a commercially available deep learning algorithm for ischemic lesion detection on brain MRIs in suspected stroke patients from a non-comprehensive stroke center.
PubMed 文献 · Eur J Radiol · 2023-09-29

回顾性纳入2020年1月6日至2022年4月1日疑似卒中的1030例脑MRI患者,以神经放射科医生重新判读为参考,评估CE认证AI工具Apollo 2.1.1检测急性缺血性病灶的能力。最终995例分析显示基于病例的敏感度89%(95%CI 85%-91%)、特异度90%(95%CI 87%-92%)。AI工具具有高敏感度和特异度,敏感度取决于缺血病灶特征,特异度取决于图像质量。

要点

  • 基于病例分析,AI检测急性缺血性病灶敏感度89%、特异度90%。
  • DWI伪影病例中特异度降低,阅片需关注图像质量。
  • 缺血病灶越大、呈碎片化时敏感度更高,非急性期病灶敏感度降低。

核心结论

  • 与经验丰富的神经放射科医生相比,该AI工具在MRI上检测急性缺血性病灶具有高敏感度和特异度。

研究设计与样本

  • 回顾性纳入2020年1月6日至2022年4月1日疑似卒中且年龄≥18岁、扫描前4周内有症状的1030例脑MRI患者,排除因伪影或序列缺失导致判读困难的扫描。

关键影像表现/征象

  • AI工具检测急性缺血性病灶的敏感度受缺血病灶大小和碎片化影响,病灶越大、越碎片化敏感度越高。
  • 非急性期病灶年龄会降低AI检测的敏感度。
  • DWI伪影病例中AI检测的特异度降低。

诊断标准与关键数值

  • 基于病例分析检测急性缺血性病灶:敏感度89%(95%CI 85%-91%),特异度90%(95%CI 87%-92%)。
  • 按性别、年龄或合并症分层,敏感度和特异度无显著差异。

局限性

  • 研究为回顾性设计。
  • 前800例连续纳入,其余病例针对较少见病灶进行富集。
  • 使用同一厂商的四种MRI扫描仪型号。

以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。

英文原文(摘要)· 点击展开

PURPOSE: To estimate the ability of a commercially available artificial intelligence (AI) tool to detect acute brain ischemia on Magnetic Resonance Imaging (MRI), compared to an experienced neuroradiologist.

METHODS: We retrospectively included 1030 patients with brain MRI, suspected of stroke from January 6th, 2020 to 1st of April 2022, based on these criteria: Age ≥ 18 years, symptoms within four weeks before the scan. The neuroradiologist reinterpreted the MRI scans and subclassified ischemic lesions for reference. We excluded scans with interpretation difficulties due to artifacts or missing sequences. Four MRI scanner models from the same vendor were used. The first 800 patients were included consecutively, remaining enriched for less frequent lesions. The index test was a CE-approved AI tool (Apollo version 2.1.1 by Cerebriu).

RESULTS: The final analysis cohort comprised 995 patients (mean age 69 years, 53 % female). A case-based analysis for detecting acute ischemic lesions showed a sensitivity of 89 % (95 % CI: 85 %-91 %) and specificity of 90 % (95 % CI: 87 %-92 %). We found no significant difference in sensitivity or specificity based on sex, age, or comorbidities. Specificity was reduced in cases with DWI artifacts. Multivariate analysis showed that increasing ischemic lesion size and fragmented lesions were independently associated with higher sensitivity, while non-acute lesion ages lowered sensitivity.

CONCLUSIONS: The AI tool exhibits high sensitivity and specificity in detecting acute ischemic lesions on MRI compared to an experienced neuroradiologist. While sensitivity depends on the ischemic lesions' characteristics, specificity depends on the image quality.

来源:Eur J Radiol,查看原文。版权归原作者所有。

引用

  1. Krag CH, Müller FC, Gandrup KL, et al. Diagnostic test accuracy study of a commercially available deep learning algorithm for ischemic lesion detection on brain MRIs in suspected stroke patients from a non-comprehensive stroke center. Eur J Radiol. 2023. PMID: 37804650. doi:10.1016/j.ejrad.2023.111126

供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。