Department of Medical Imaging, Chi Mei Medical Center, Tainan, Taiwan.
Department of Medical Research, Chi Mei Medical Center, Tainan, Taiwan.
Eur J Radiol. 2023 Oct;167:111034. doi: 10.1016/j.ejrad.2023.111034. Epub 2023 Aug 11.
This study aimed to develop preprocedural real-time artificial intelligence (AI)-based systems for predicting individualized risks of contrast-associated acute kidney injury (CA-AKI) and dialysis requirement within 30 days following contrast-enhanced computed tomography (CECT).
This single-center, retrospective study analyzed adult patients from emergency or in-patient departments who underwent CECT; 18,895 patients were included after excluding those who were already on dialysis, had stage V chronic kidney disease, or had missing data regarding serum creatinine levels within 7 days before and after CECT. Clinical parameters, laboratory data, medication exposure, and comorbid diseases were selected as predictive features. The patients were randomly divided into model training and testing groups at a 7:3 ratio. Logistic regression (LR) and random forest (RF) were employed to create prediction models, which were evaluated using receiver operating characteristic curves.
The incidence rates of CA-AKI and dialysis within 30 days post-CECT were 6.69% and 0.98%, respectively. For CA-AKI prediction, LR and RF exhibited similar performance, with areas under curve (AUCs) of 0.769 and 0.757, respectively. For 30-day dialysis prediction, LR (AUC, 0.863) and RF (AUC, 0.872) also exhibited similar performance. Relative to eGFR-alone, the LR and RF models produced significantly higher AUCs for CA-AKI prediction (LR vs. eGFR alone, 0.769 vs. 0.626, p < 0.001) and 30-day dialysis prediction (RF vs. eGFR alone, 0.872 vs. 0.738, p < 0.001).
The proposed AI prediction models significantly outperformed eGFR-alone for predicting the CA-AKI and 30-day dialysis risks of emergency department and hospitalized patients who underwent CECT.
本研究旨在开发术前实时人工智能（AI）系统，以预测接受增强计算机断层扫描（CECT）后 30 天内的对比相关急性肾损伤（CA-AKI）和透析需求的个体风险。
这是一项单中心、回顾性研究，分析了在急诊或住院病房接受 CECT 的成年患者；排除已经接受透析、慢性肾脏病 5 期或 CECT 前后 7 天内血清肌酐水平缺失数据的患者后，共纳入 18895 例患者。选择临床参数、实验室数据、药物暴露和合并症作为预测特征。患者按 7:3 的比例随机分为模型训练组和测试组。采用逻辑回归（LR）和随机森林（RF）建立预测模型，并通过接受者操作特征曲线进行评估。
CECT 后 30 天内 CA-AKI 和透析的发生率分别为 6.69%和 0.98%。对于 CA-AKI 的预测，LR 和 RF 的性能相似，曲线下面积（AUC）分别为 0.769 和 0.757。对于 30 天透析预测，LR（AUC，0.863）和 RF（AUC，0.872）也表现出相似的性能。与 eGFR 相比，LR 和 RF 模型对 CA-AKI 预测（LR 与 eGFR 相比，0.769 与 0.626，p<0.001）和 30 天透析预测（RF 与 eGFR 相比，0.872 与 0.738，p<0.001）的 AUC 显著提高。
对于接受 CECT 的急诊和住院患者，所提出的 AI 预测模型在预测 CA-AKI 和 30 天透析风险方面明显优于 eGFR 单独预测。
