Automated Detection of Severe Cerebral Edema Using Explainable Deep Transfer Learning after Hypoxic Ischemic Brain Injury.

IF 6.5 1区医学 Q1 CRITICAL CARE MEDICINE

Resuscitation Pub Date : 2025-05-23 DOI:10.1016/j.resuscitation.2025.110652

Zihao Wang, Annelise M Kulpanowski, William A Copen, Eric S Rosenthal, Jacob A Dodelson, David E McCrory, Brian L Edlow, W Taylor Kimberly, Edilberto Amorim, MBrandon Westover, MingMing Ning, Morteza Zabihi, Pamela W Schaefer, Rajeev Malhotra, Joseph T Giacino, David M Greer, Ona Wu

{"title":"Automated Detection of Severe Cerebral Edema Using Explainable Deep Transfer Learning after Hypoxic Ischemic Brain Injury.","authors":"Zihao Wang, Annelise M Kulpanowski, William A Copen, Eric S Rosenthal, Jacob A Dodelson, David E McCrory, Brian L Edlow, W Taylor Kimberly, Edilberto Amorim, MBrandon Westover, MingMing Ning, Morteza Zabihi, Pamela W Schaefer, Rajeev Malhotra, Joseph T Giacino, David M Greer, Ona Wu","doi":"10.1016/j.resuscitation.2025.110652","DOIUrl":null,"url":null,"abstract":"Background: Substantial gaps exist in the neuroprognostication of cardiac arrest patients who remain comatose after the restoration of spontaneous circulation. Most studies focus on predicting survival, a measure confounded by the withdrawal of life-sustaining treatment decisions. Severe cerebral edema (SCE) may serve as an objective proximal imaging-based surrogate of neurologic injury.Methods: We retrospectively analyzed data from 288 patients to automate SCE detection with machine learning (ML) and to test the hypothesis that the quantitative values produced by these algorithms (ML_SCE) can improve predictions of neurologic outcomes. Ground-truth SCE (GT_SCE) classification was based on radiology reports.Results: The model attained a cross-validated testing accuracy of 87% [95% CI: 84%, 89%] for detecting SCE. Attention maps explaining SCE classification focused on cisternal regions (p<0.05). Multivariable analyses showed that older age (p<0.001), non-shockable initial cardiac rhythm (p=0.004), and greater ML_SCE values (p<0.001) were significant predictors of poor neurologic outcomes, with GT_SCE (p=0.064) as a non-significant covariate.Conclusion: Our results support the feasibility of automated SCE detection. Future prospective studies with standardized neurologic assessments are needed to substantiate the utility of quantitative ML_SCE values to improve neuroprognostication.","PeriodicalId":21052,"journal":{"name":"Resuscitation","volume":" ","pages":"110652"},"PeriodicalIF":6.5000,"publicationDate":"2025-05-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Resuscitation","FirstCategoryId":"3","ListUrlMain":"https://doi.org/10.1016/j.resuscitation.2025.110652","RegionNum":1,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"CRITICAL CARE MEDICINE","Score":null,"Total":0}

引用次数: 0

Abstract

Background: Substantial gaps exist in the neuroprognostication of cardiac arrest patients who remain comatose after the restoration of spontaneous circulation. Most studies focus on predicting survival, a measure confounded by the withdrawal of life-sustaining treatment decisions. Severe cerebral edema (SCE) may serve as an objective proximal imaging-based surrogate of neurologic injury.

Methods: We retrospectively analyzed data from 288 patients to automate SCE detection with machine learning (ML) and to test the hypothesis that the quantitative values produced by these algorithms (ML_SCE) can improve predictions of neurologic outcomes. Ground-truth SCE (GT_SCE) classification was based on radiology reports.

Results: The model attained a cross-validated testing accuracy of 87% [95% CI: 84%, 89%] for detecting SCE. Attention maps explaining SCE classification focused on cisternal regions (p<0.05). Multivariable analyses showed that older age (p<0.001), non-shockable initial cardiac rhythm (p=0.004), and greater ML_SCE values (p<0.001) were significant predictors of poor neurologic outcomes, with GT_SCE (p=0.064) as a non-significant covariate.

Conclusion: Our results support the feasibility of automated SCE detection. Future prospective studies with standardized neurologic assessments are needed to substantiate the utility of quantitative ML_SCE values to improve neuroprognostication.

查看原文本刊更多论文

缺氧缺血性脑损伤后使用可解释的深度迁移学习自动检测严重脑水肿。

背景：在心脏骤停患者恢复自然循环后仍处于昏迷状态的神经预后方面存在实质性的空白。大多数研究的重点是预测生存，这是一项因撤销维持生命的治疗决定而混淆的措施。严重脑水肿（SCE）可以作为神经损伤的客观近端影像学替代指标。方法：我们回顾性分析288例患者的数据，用机器学习（ML）自动检测SCE，并检验这些算法产生的定量值（ML_SCE）可以改善神经系统预后预测的假设。基础真实SCE （GT_SCE）分类基于放射学报告。结果：该模型检测SCE的交叉验证准确率为87% [95% CI: 84%， 89%]。结论：本研究结果支持SCE自动检测的可行性。未来需要标准化神经系统评估的前瞻性研究来证实定量ML_SCE值对改善神经预后的效用。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Resuscitation 医学-急救医学

CiteScore

12.00

自引率

18.50%

发文量

556

审稿时长

21 days

期刊介绍： Resuscitation is a monthly international and interdisciplinary medical journal. The papers published deal with the aetiology, pathophysiology and prevention of cardiac arrest, resuscitation training, clinical resuscitation, and experimental resuscitation research, although papers relating to animal studies will be published only if they are of exceptional interest and related directly to clinical cardiopulmonary resuscitation. Papers relating to trauma are published occasionally but the majority of these concern traumatic cardiac arrest.