Related Experiment Video
Updated: Jul 9, 2026

In vitro Assessment of Aortic Regurgitation Using Four-Dimensional Flow Magnetic Resonance Imaging
Published on: February 25, 2022
Artificial Intelligence-Driven Detection, Characterization, and Risk Stratification in Patients with Severe Tricuspid
Luna Varela do Carmo1, Auristela Isabel de Oliveira Ramos1, Dorival Julio Della Togna1
1Instituto Dante Pazzanese de Cardiologia, São Paulo, SP - Brasil.
Artificial intelligence identified complex clinical and prognostic factors in severe tricuspid regurgitation (TR) patients in Brazil. High-risk TRI-SCORE and elevated creatinine predicted mortality in this large cohort.
Area of Science:
- Cardiology
- Artificial Intelligence in Medicine
- Medical Informatics
Background:
- Tricuspid regurgitation (TR) is a prevalent condition with high mortality, especially with isolated surgery or late intervention.
- Transcatheter therapies for TR are expanding, yet Brazil lacks data and validated prognostic tools.
- Effective management of severe TR is hindered by challenges in risk stratification and patient selection.
Purpose of the Study:
- To define clinical, laboratory, and echocardiographic data in severe TR patients using an AI model.
- To investigate prognostic variables within a Brazilian cohort.
- To leverage AI for enhanced understanding and management of severe TR.
Main Methods:
- Retrospective analysis of 71,911 echocardiographic reports (2021-2024).
- Development of a Python-based natural language processing (NLP) model to identify severe TR and extract variables.
- Automated calculation of TRI-SCORE and EuroSCORE II, with subsequent review.
Main Results:
- 803 patients with severe TR identified; mean age 68.8 years, 64.1% female.
- Common comorbidities: hypertension (71.7%), atrial fibrillation (69.4%). Secondary TR predominated (96.8%), often linked to mitral valve disease (68.5%).
- High surgical risk (EuroSCORE II median 13.4%, TRI-SCORE median 7), with 17.1% mortality. Independent predictors of mortality included high-risk TRI-SCORE (RR 2.19) and increased creatinine (RR 3.13).
Conclusions:
- The largest Brazilian severe TR cohort, analyzed with AI, reveals significant clinical and prognostic complexity.
- NLP application to large echocardiographic databases is efficient for data extraction and analysis.
- AI-assisted analysis can support multicenter studies and improve understanding of severe TR.
Related Concept Videos
Aortic Regurgitation II: Clinical Features and Diagnostic Tests
Aortic Regurgitation I: Introduction
Rheumatic Heart Disease II: Clinical Manifestations and Diagnostic Studies
Mitral Regurgitation II: Clinical Features and Diagnostic Tests
Aortic Regurgitation III: Medical Management
Mitral Regurgitation I: Introduction