Duly Health and Care, Department of Radiology, Downers Grove, IL, USA.
Department of Biomedical and Health Information Sciences, University of Illinois Chicago, Chicago, IL, USA.
Nat Commun. 2023 Jul 7;14(1):4039. doi: 10.1038/s41467-023-39631-x.
Deep learning (DL) models can harness electronic health records (EHRs) to predict diseases and extract radiologic findings for diagnosis. With ambulatory chest radiographs (CXRs) frequently ordered, we investigated detecting type 2 diabetes (T2D) by combining radiographic and EHR data using a DL model. Our model, developed from 271,065 CXRs and 160,244 patients, was tested on a prospective dataset of 9,943 CXRs. Here we show the model effectively detected T2D with a ROC AUC of 0.84 and a 16% prevalence. The algorithm flagged 1,381 cases (14%) as suspicious for T2D. External validation at a distinct institution yielded a ROC AUC of 0.77, with 5% of patients subsequently diagnosed with T2D. Explainable AI techniques revealed correlations between specific adiposity measures and high predictivity, suggesting CXRs' potential for enhanced T2D screening.
深度学习 (DL) 模型可以利用电子健康记录 (EHR) 预测疾病并提取放射学发现用于诊断。由于经常需要进行门诊胸部 X 光检查 (CXR)，我们研究了通过结合放射学和 EHR 数据使用 DL 模型来检测 2 型糖尿病 (T2D)。我们的模型是从 271,065 张 CXR 和 160,244 名患者中开发的，在一个前瞻性的 9,943 张 CXR 数据集上进行了测试。在这里，我们展示了该模型能够有效地检测 T2D，ROC AUC 为 0.84，患病率为 16%。该算法标记了 1381 例（14%）疑似 T2D 的病例。在另一个不同的机构进行的外部验证得到了 ROC AUC 为 0.77，其中 5%的患者随后被诊断为 T2D。可解释人工智能技术揭示了特定肥胖指标与高预测性之间的相关性，这表明 CXR 有可能用于增强 T2D 筛查。
Zotero 插件
