目的·基于心脏磁共振（cardiac magnetic resonance，CMR）构建新型冠状病毒感染（coronavirus disease 2019，COVID-19）轻症患者心肌损伤的诊断模型。方法·纳入2024年1月—2025年6月就诊于上海交通大学医学院附属第六人民医院的COVID-19核酸检测阳性并伴有心血管症状的患者64例；患者均于感染后6个月内完成肌钙蛋白I（cardiac troponin I，cTnI）检测及CMR检查。依据cTnI水平将患者分为cTnI升高组［cTnI（+）组，n=26］、cTnI未升高组［cTnI（-）组，n=38］，并纳入年龄、性别匹配的37例未感染COVID-19的人群作为健康对照组。收集并比较3组受试者的基线临床资料、CMR参数（包括心功能与容积参数、左心室整体及节段心肌应变参数、心肌组织学参数）。采用Spearman秩相关分析评估cTnI与基线临床资料及CMR参数的相关性。通过单因素和多因素Logistic回归分析构建能够早期识别COVID-19相关心肌损伤的诊断模型。采用受试者操作特征（receiver operator characteristic，ROC）曲线及Delong检验验证该模型的判别效能。结果·心功能与容积参数对比分析显示，左心室射血分数（left ventricular ejection fraction，LVEF）等指标的组间差异无统计学意义。与健康对照组相比，cTnI（+）组患者的左心室整体周向应变（global circumferential strain，GCS）、左室基底段周向应变（left ventricular basal-level circumferential strain，CS_Basal）、左室中段周向应变（left ventricular mid-level circumferential strain，CS_Mid）有所降低（均P<0.001），Native T1 mapping值有所升高（P<0.001）；cTnI（-）组患者的Native T1 mapping值也有升高（P=0.007）。相关性分析显示，cTnI与CS_Mid呈显著负相关（r=‒0.600，P<0.001）。多因素Logistic回归分析显示，CS_Mid和Native T1 mapping均为COVID-19相关心肌损伤的独立预测因子（均P<0.001）。ROC曲线显示，基于CS_Mid与Native T1 mapping构建的联合诊断模型具有良好的判别效能（AUC=0.950）。结论·基于CS_Mid与Native T1 mapping的CMR联合诊断模型能够有效识别COVID-19轻症患者的潜在心肌损伤，或可为COVID-19相关心肌损伤的早期识别和临床评估提供客观影像学依据。

Objective ·To establish a cardiac magnetic resonance (CMR)-based diagnostic model for detecting myocardial injury in patients with mild coronavirus disease 2019 (COVID-19). Methods ·A total of 64 patients with positive nucleic acid test results for COVID-19 and accompanying cardiovascular symptoms who visited Shanghai Sixth People's Hospital, Shanghai Jiao Tong University School of Medicine, between January 2024 and June 2025 were enrolled. All patients underwent cardiac troponin I (cTnI) testing and CMR examination within 6 months after infection. According to cTnI levels, patients were divided into an elevated cTnI group [cTnI(+), n=26] and a non-elevated cTnI group [cTnI(-), n=38]. In addition, 37 age- and gender- matched individuals without prior COVID-19 infection were included as healthy controls. Baseline clinical characteristics and CMR parameters, including cardiac function and volumetric parameters, left ventricular global and segmental strain parameters, and myocardial tissue characterization parameters, were collected and compared among the three groups. Spearman rank correlation analysis was performed to evaluate the associations between cTnI and baseline clinical characteristics as well as CMR parameters. Univariate and multivariate Logistics regression analyses were used to establish a diagnostic model for early identification of COVID-19-related myocardial injury. Receiver operator characteristic (ROC) curve analysis and the DeLong test were used to assess the discriminative performance of the model. Results ·Comparative analysis of cardiac function and volumetric parameters showed no significant differences in indices such as left ventricular ejection fraction (LVEF) among the three groups. Compared with the healthy control group, the cTnI(+) group showed decreased left ventricular global circumferential strain (GCS), left ventricular basal-level circumferential strain (CS_Basal), and left ventricular mid-level circumferential strian (CS_Mid) (P<0.001), as well as increased Native T1 mapping values (P<0.001); the cTnI(-) group also showed elevated Native T1 mapping values (P=0.007). Correlation analysis showed that cTnI was significantly negatively correlated with CS_Mid (r=‒0.600, P<0.001). Multivariate Logistic regression analysis revealed that CS_Mid and Native T1 mapping were independent predictors of COVID-19-related myocardial injury (P<0.001). ROC curve analysis showed that the combined diagnostic model based on CS_Mid and Native T1 mapping demonstrated good discriminative performance (AUC=0.950). Conclusion ·The CMR-based diagnostic model combining CS_Mid and Native T1 mapping can effectively identify potential myocardial injury in patients with mild COVID-19. It may provide objective imaging evidence for the early detection and clinical evaluation of COVID-19-related myocardial injury.