Scholarlink

<p>Right ventricular (RV) strain is an important factor in evaluating RV function. This study aimed to predict mortality in patients with pulmonary hypertension (PH) and right heart failure (RHF) using an artificial intelligence (AI)-assisted multimodal echocardiography approach. A total of 586 patients with PH and RHF were enrolled. Transthoracic echocardiography (TTE) was performed either before hospital admission in the outpatient department or after admission in the inpatient unit. RV structure and function were assessed using two-dimensional measurements and speckle tracking strain analysis of septal and the free-wall segments. A multimodal deep-learning framework, integrating clinical variables, conventional echocardiographic indices, and AI-extracted RV strain features was developed to predict in-hospital and follow-up mortality. Model performance was evaluated using receiver operating characteristic (ROC) curves and the area under the curve (AUC). The death group had significantly higher NT-proBNP levels, pulmonary artery systolic pressure (PASP), and average RV longitudinal strain (RVLS) than the survival group. Logistic regression identified PASP and average RVLS as independent predictors of in-hospital mortality, while Cox regression confirmed average RVLS as an independent predictor of follow-up mortality. The testing AUC for the AI model was 0.823 (0.741–0.905). The AI-guided TTE framework provides an effective and efficient method for mortality prediction in patients with PH and RHF, supporting personalized clinical decision-making and risk stratification.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI–assisted multimodal assessment for right ventricular function from echocardiography predicts mortality in patients with pulmonary hypertension and right heart failure

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月16日

查看全文
  • Han Mou,
  • Ganjun Zhang,
  • Lili Xiu,
  • Qi Zhao,
  • Ruoxi Zhang,
  • Guang-An Liu

摘要

Right ventricular (RV) strain is an important factor in evaluating RV function. This study aimed to predict mortality in patients with pulmonary hypertension (PH) and right heart failure (RHF) using an artificial intelligence (AI)-assisted multimodal echocardiography approach. A total of 586 patients with PH and RHF were enrolled. Transthoracic echocardiography (TTE) was performed either before hospital admission in the outpatient department or after admission in the inpatient unit. RV structure and function were assessed using two-dimensional measurements and speckle tracking strain analysis of septal and the free-wall segments. A multimodal deep-learning framework, integrating clinical variables, conventional echocardiographic indices, and AI-extracted RV strain features was developed to predict in-hospital and follow-up mortality. Model performance was evaluated using receiver operating characteristic (ROC) curves and the area under the curve (AUC). The death group had significantly higher NT-proBNP levels, pulmonary artery systolic pressure (PASP), and average RV longitudinal strain (RVLS) than the survival group. Logistic regression identified PASP and average RVLS as independent predictors of in-hospital mortality, while Cox regression confirmed average RVLS as an independent predictor of follow-up mortality. The testing AUC for the AI model was 0.823 (0.741–0.905). The AI-guided TTE framework provides an effective and efficient method for mortality prediction in patients with PH and RHF, supporting personalized clinical decision-making and risk stratification.