A comparative analysis of privacy-preserving large language models for automated echocardiography report analysis
Mahmoudi E, Vahdati S, Chao C, Khosravi B, Misra A, Lopez-Jimenez F, Erickson B. A comparative analysis of privacy-preserving large language models for automated echocardiography report analysis. Journal Of The American Medical Informatics Association 2025, 32: 1120-1129. PMID: 40334045, PMCID: PMC12257941, DOI: 10.1093/jamia/ocaf056.Peer-Reviewed Original ResearchConceptsMean square errorLanguage modelLarge modelsHigher accuracyAutomated data extractionClassification metricsNear-perfect performanceGround truthValvular heart diseaseVHD severityIrrelevant informationComputational efficiencyIncreasing processing timeProcessing timeCOTSSmall modelsAccuracyHigh performancePerformanceConsecutive transthoracic echocardiographiesOptimizationRandom reportsTransthoracic echocardiographyRegistry formatMetrics
This site is protected by hCaptcha and its Privacy Policy and Terms of Service apply