跳到正文
arXiv cs.AI· Ziqing Yu, Yuhui Tao, Jiayu Huo, Lei Pan, Zilong Xiao, Juecheng Chen, Xiao Li, Jianxuan Li, You Zhou, Zhixing Li, Cong Wang, Beijian Zhang, Chen Chen, Hongyang Lu, Konstantinos Patlatzoglou, Daniel B. Kramer, Jonathan W. Waks, Yangang Su, Fu Siong Ng, Shuo Wang, Yixiu Liang, Junbo Ge·· 2 小时前AI 评分41

CardioLoom:报告引导的心电图基础模型用于跨域心血管表型分析

CardioLoom: A Report-Guided Electrocardiogram Foundation Model for Cross-Domain Cardiovascular Phenotyping

AI 导读

研究者发布报告引导的心电图基础模型 CardioLoom,在 1,298,359 名患者的 2,781,289 份 ECG-报告对上完成预训练。模型在 45 项内部 ECG 诊断中 macro-AUPRC 达 0.624,优于 DeepECG-SSL 的 0.615;仅用 10% 患者标签即取得 0.559,超过使用全部标签的监督训练(0.549)。

正文 · 原文

Authors:Ziqing Yu, Yuhui Tao, Jiayu Huo, Lei Pan, Zilong Xiao, Juecheng Chen, Xiao Li, Jianxuan Li, You Zhou, Zhixing Li, Cong Wang, Beijian Zhang, Chen Chen, Hongyang Lu, Konstantinos Patlatzoglou, Daniel B. Kramer, Jonathan W. Waks, Yangang Su, Fu Siong Ng, Shuo Wang, Yixiu Liang, Junbo Ge

View PDF

Abstract:Routine electrocardiogram (ECG) reports offer scalable supervision for cardiovascular representation learning. In this retrospective study, we pretrained CardioLoom, a report-guided ECG foundation model, on 2,781,289 ECG-report pairs from 1,298,359 patients. Downstream development used a separate cohort of 721,317 ECGs from 233,798 patients. Evaluation covered 89 targets, with external testing in nine cohorts. Across 45 internal ECG diagnoses, macro area under the precision-recall curve (AUPRC) was 0.624 versus 0.615 for DeepECG-SSL. Across 203 external ECG evaluations, corresponding means were 0.425 and 0.424. Using labels from 10% of training patients, CardioLoom achieved macro-AUPRC 0.559, exceeding supervised training with all labels (0.549). For heart failure with reduced ejection fraction, AUPRCs were 0.690, 0.641 and 0.677 in Zhongshan, Shiyuan and BIDMC. Cardiac amyloidosis screening yielded 4.5% positive predictive value using a threshold targeting 99% specificity in validation. Report-guided pretraining supported label-efficient ECG diagnosis, with variable transfer across echocardiographic phenotypes.
Comments: 135 pages, 5 main figures, 7 supplementary figures and 31 supplementary tables. Substantially revised; ECGCLIP renamed CardioLoom. Updated pretraining methods, cohort accounting, baseline comparisons, statistical analyses and supplementary materials
Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)
Cite as: arXiv:2605.25446 [cs.AI]
  (or arXiv:2605.25446v2 [cs.AI] for this version)
  https://doi.org/10.48550/arXiv.2605.25446

arXiv-issued DOI via DataCite

Submission history

From: Yuhui Tao [view email]
[v1] Mon, 25 May 2026 05:51:44 UTC (33,058 KB)
[v2] Tue, 17 Nov 2026 11:42:21 UTC (10,067 KB)

来源:arXiv cs.AI · arxiv.org