AI-enhanced Echocardiography In Cardiac Amyloidosis: A Systematic Review And Meta-analysis Of Diagnostic Performance
HFSA ePoster Library. Mylavarapu M. 10/11/26; 4235143; 513
Dr. Maneeth Mylavarapu
REGULAR CONTENT
Login now to access Regular content available to all registered users.
Abstract
Discussion Forum (0)
Rate & Comment (0)
General ePoster - EP.6.A: On 2026-10-11 12:45 (Monitor 45)
Topic: Infiltrative Cardiomyopathies (Sarcoidosis, Amyloidosis) and Myocarditis
Maneeth Mylavarapu1, Madiha Kiyani2, Lakshmi Sai Meghana Kodali3, Niharika Tanwar4, Mina Kerolos1. 1Endeavor Health Cardiovascular Institute, Glenview, IL; 2MedStar Health Internal Medicine, Baltimore, Baltimore, MD; 3University of Michigan-Flint, Flint, MI; 4Advocate Illinois Masonic Medical Center, Chicago, IL
Introduction:
Cardiac amyloidosis (CA) is a progressive and often underdiagnosed condition due to its heterogeneous clinical presentation and the specialized expertise required for echocardiographic interpretation. While artificial intelligence (AI) models have been under development to enhance the detection of CA through echocardiography, their collective diagnostic performance have not been robustly studied.
Hypothesis:
We hypothesized that AI-enhanced echocardiographic models would demonstrate high diagnostic accuracy in identifying patients with cardiac amyloidosis.
Methods:
Per PRISMA guidelines, a comprehensive search was conducted in PubMed/MEDLINE, Google Scholar, and Cochrane and studies that evaluated the efficacy of AI-enhanced echocardiography in identifying CA were included. The primary outcome was the pooled area under the curve (AUC) for the diagnostic accuracy of AI in identifying overall CA. Subgroup analyses were conducted to differentiate performance between transthyretin (ATTR) and light-chain (AL) amyloidosis. Invariance-variance random effects model, pooling data for AUC and 95% confidence intervals were used for the analysis.
Results:
A total of 9 studies with 4419 patients (validation cohorts) were included in the analysis. AI-enhanced echocardiography demonstrated high diagnostic accuracy with a pooled AUC of 0.96 (95% CI: 0.94-0.98; p<0.0001), though significant heterogeneity was observed (I2 = 81.1%). In subgroup analyses, AI performance remained robust across subtypes, yielding an AUC of 0.98 (95% CI: 0.96-0.99; p<0.0001) for ATTR and 0.96 (95% CI: 0.92-0.99; p<0.0001) for AL amyloidosis.
Conclusion:
AI-enhanced echocardiography demonstrates high diagnostic accuracy for cardiac amyloidosis across both ATTR and AL subtypes. While these results support the use of AI as a powerful screening tool to reduce underdiagnosis, the observed heterogeneity highlights the need for standardized validation in diverse clinical settings. Future research should focus on integrating these models into routine clinical workflows to improve early detection and patient outcomes.
Topic: Infiltrative Cardiomyopathies (Sarcoidosis, Amyloidosis) and Myocarditis
Maneeth Mylavarapu1, Madiha Kiyani2, Lakshmi Sai Meghana Kodali3, Niharika Tanwar4, Mina Kerolos1. 1Endeavor Health Cardiovascular Institute, Glenview, IL; 2MedStar Health Internal Medicine, Baltimore, Baltimore, MD; 3University of Michigan-Flint, Flint, MI; 4Advocate Illinois Masonic Medical Center, Chicago, IL
Introduction:
Cardiac amyloidosis (CA) is a progressive and often underdiagnosed condition due to its heterogeneous clinical presentation and the specialized expertise required for echocardiographic interpretation. While artificial intelligence (AI) models have been under development to enhance the detection of CA through echocardiography, their collective diagnostic performance have not been robustly studied.
Hypothesis:
We hypothesized that AI-enhanced echocardiographic models would demonstrate high diagnostic accuracy in identifying patients with cardiac amyloidosis.
Methods:
Per PRISMA guidelines, a comprehensive search was conducted in PubMed/MEDLINE, Google Scholar, and Cochrane and studies that evaluated the efficacy of AI-enhanced echocardiography in identifying CA were included. The primary outcome was the pooled area under the curve (AUC) for the diagnostic accuracy of AI in identifying overall CA. Subgroup analyses were conducted to differentiate performance between transthyretin (ATTR) and light-chain (AL) amyloidosis. Invariance-variance random effects model, pooling data for AUC and 95% confidence intervals were used for the analysis.
Results:
A total of 9 studies with 4419 patients (validation cohorts) were included in the analysis. AI-enhanced echocardiography demonstrated high diagnostic accuracy with a pooled AUC of 0.96 (95% CI: 0.94-0.98; p<0.0001), though significant heterogeneity was observed (I2 = 81.1%). In subgroup analyses, AI performance remained robust across subtypes, yielding an AUC of 0.98 (95% CI: 0.96-0.99; p<0.0001) for ATTR and 0.96 (95% CI: 0.92-0.99; p<0.0001) for AL amyloidosis.
Conclusion:
AI-enhanced echocardiography demonstrates high diagnostic accuracy for cardiac amyloidosis across both ATTR and AL subtypes. While these results support the use of AI as a powerful screening tool to reduce underdiagnosis, the observed heterogeneity highlights the need for standardized validation in diverse clinical settings. Future research should focus on integrating these models into routine clinical workflows to improve early detection and patient outcomes.
General ePoster - EP.6.A: On 2026-10-11 12:45 (Monitor 45)
Topic: Infiltrative Cardiomyopathies (Sarcoidosis, Amyloidosis) and Myocarditis
Maneeth Mylavarapu1, Madiha Kiyani2, Lakshmi Sai Meghana Kodali3, Niharika Tanwar4, Mina Kerolos1. 1Endeavor Health Cardiovascular Institute, Glenview, IL; 2MedStar Health Internal Medicine, Baltimore, Baltimore, MD; 3University of Michigan-Flint, Flint, MI; 4Advocate Illinois Masonic Medical Center, Chicago, IL
Introduction:
Cardiac amyloidosis (CA) is a progressive and often underdiagnosed condition due to its heterogeneous clinical presentation and the specialized expertise required for echocardiographic interpretation. While artificial intelligence (AI) models have been under development to enhance the detection of CA through echocardiography, their collective diagnostic performance have not been robustly studied.
Hypothesis:
We hypothesized that AI-enhanced echocardiographic models would demonstrate high diagnostic accuracy in identifying patients with cardiac amyloidosis.
Methods:
Per PRISMA guidelines, a comprehensive search was conducted in PubMed/MEDLINE, Google Scholar, and Cochrane and studies that evaluated the efficacy of AI-enhanced echocardiography in identifying CA were included. The primary outcome was the pooled area under the curve (AUC) for the diagnostic accuracy of AI in identifying overall CA. Subgroup analyses were conducted to differentiate performance between transthyretin (ATTR) and light-chain (AL) amyloidosis. Invariance-variance random effects model, pooling data for AUC and 95% confidence intervals were used for the analysis.
Results:
A total of 9 studies with 4419 patients (validation cohorts) were included in the analysis. AI-enhanced echocardiography demonstrated high diagnostic accuracy with a pooled AUC of 0.96 (95% CI: 0.94-0.98; p<0.0001), though significant heterogeneity was observed (I2 = 81.1%). In subgroup analyses, AI performance remained robust across subtypes, yielding an AUC of 0.98 (95% CI: 0.96-0.99; p<0.0001) for ATTR and 0.96 (95% CI: 0.92-0.99; p<0.0001) for AL amyloidosis.
Conclusion:
AI-enhanced echocardiography demonstrates high diagnostic accuracy for cardiac amyloidosis across both ATTR and AL subtypes. While these results support the use of AI as a powerful screening tool to reduce underdiagnosis, the observed heterogeneity highlights the need for standardized validation in diverse clinical settings. Future research should focus on integrating these models into routine clinical workflows to improve early detection and patient outcomes.
Topic: Infiltrative Cardiomyopathies (Sarcoidosis, Amyloidosis) and Myocarditis
Maneeth Mylavarapu1, Madiha Kiyani2, Lakshmi Sai Meghana Kodali3, Niharika Tanwar4, Mina Kerolos1. 1Endeavor Health Cardiovascular Institute, Glenview, IL; 2MedStar Health Internal Medicine, Baltimore, Baltimore, MD; 3University of Michigan-Flint, Flint, MI; 4Advocate Illinois Masonic Medical Center, Chicago, IL
Introduction:
Cardiac amyloidosis (CA) is a progressive and often underdiagnosed condition due to its heterogeneous clinical presentation and the specialized expertise required for echocardiographic interpretation. While artificial intelligence (AI) models have been under development to enhance the detection of CA through echocardiography, their collective diagnostic performance have not been robustly studied.
Hypothesis:
We hypothesized that AI-enhanced echocardiographic models would demonstrate high diagnostic accuracy in identifying patients with cardiac amyloidosis.
Methods:
Per PRISMA guidelines, a comprehensive search was conducted in PubMed/MEDLINE, Google Scholar, and Cochrane and studies that evaluated the efficacy of AI-enhanced echocardiography in identifying CA were included. The primary outcome was the pooled area under the curve (AUC) for the diagnostic accuracy of AI in identifying overall CA. Subgroup analyses were conducted to differentiate performance between transthyretin (ATTR) and light-chain (AL) amyloidosis. Invariance-variance random effects model, pooling data for AUC and 95% confidence intervals were used for the analysis.
Results:
A total of 9 studies with 4419 patients (validation cohorts) were included in the analysis. AI-enhanced echocardiography demonstrated high diagnostic accuracy with a pooled AUC of 0.96 (95% CI: 0.94-0.98; p<0.0001), though significant heterogeneity was observed (I2 = 81.1%). In subgroup analyses, AI performance remained robust across subtypes, yielding an AUC of 0.98 (95% CI: 0.96-0.99; p<0.0001) for ATTR and 0.96 (95% CI: 0.92-0.99; p<0.0001) for AL amyloidosis.
Conclusion:
AI-enhanced echocardiography demonstrates high diagnostic accuracy for cardiac amyloidosis across both ATTR and AL subtypes. While these results support the use of AI as a powerful screening tool to reduce underdiagnosis, the observed heterogeneity highlights the need for standardized validation in diverse clinical settings. Future research should focus on integrating these models into routine clinical workflows to improve early detection and patient outcomes.
Code of conduct/disclaimer available in General Terms & Conditions
{{ help_message }}
{{filter}}