Johns Hopkins AI预测手术并发症 85%的时间 使用ECGs和记录 打破传统方法。 Johns Hopkins AI predicts surgery complications 85% of the time using ECGs and records, beating traditional methods.

约翰·霍普金斯大学的研究人员开发了一个AI模型,通过分析标准的心电图测试和病人医疗记录,预测手术后并发症的准确率达到85%,超过60%左右的准确率的传统风险分数。 Researchers at Johns Hopkins University have developed an AI model that predicts post-surgical complications with 85% accuracy by analyzing standard electrocardiogram (ECG) tests and patient medical records, outperforming traditional risk scores that are accurate in about 60% of cases. 该模型利用深层学习发现临床医生先前忽视的ECG数据的微妙模式,有可能改善外科决策和患者结果。 The model uses deep learning to detect subtle patterns in ECG data previously overlooked by clinicians, potentially improving surgical decision-making and patient outcomes. 这项研究以37,000名患者的数据为基础,建议AI可以改变如何评估外科手术风险的方法,并计划进一步测试,以验证和扩大外科手术的使用。 The study, based on data from 37,000 patients, suggests AI could transform how surgical risks are assessed, with further testing planned to validate and expand its use.