Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
August 15, 2025Frontiers in MedicineOpen Access

A new integrated machine learning model: application to improve the accuracy of predicting left atrial appendage thrombus in patients with non-valvular atrial fibrillation

View Full Paper
Ask AI
Bookmark
Share

Authors

PMPeipei MaiLuoyang Central Hospital Affiliated to Zhengzhou UniversityHHhuanhuan huoSecond Affiliated Hospital of Xi'an Jiaotong UniversityXLXiaona LiLuoyang Central Hospital Affiliated to Zhengzhou University

Discussion

Loading...

Member takes

Overview

Machine learning improves left atrial appendage thrombus detection in patients with non-valvular atrial fibrillation, suggesting a non-invasive alternative to transesophageal echocardiography.

Key Points

  • The Logistic Regression model provides a reliable, non-invasive approach for predicting left atrial appendage thrombus in high-risk patients.
  • With an AUC of 80.9%, the machine learning model significantly outperformed independent models based on transthoracic echocardiography.
  • Testing involved 698 patients with non-valvular atrial fibrillation from a hospital between January 2021 and May 2024.
  • Integrating clinical data with echocardiographic features may reduce reliance on transesophageal echocardiography for thrombus detection.

Cite This Study

Mai et al. (2025) studied this question.

synapsesocial.com/papers/68af50a1ad7bf08b1ead899ahttps://doi.org/10.3389/fmed.2025.1661696
View Full Paper
Ask AI
Bookmark
Share