Clinical Scorecard: Machine Learning-Based Scoring System to Assess Diagnostic Efficacy of Endomyocardial Biopsy
At a Glance
Category
Detail
Condition
Cardiomyopathies and heart failure requiring diagnostic endomyocardial biopsy
Key Mechanisms
Predictive modeling using non-invasive clinical and imaging data to estimate diagnostic yield of EMB
Target Population
Heart failure patients undergoing evaluation for cardiomyopathy diagnosis
Care Setting
Cardiology and heart failure specialty clinics with access to cardiac MRI and biomarker testing
Key Highlights
Endomyocardial biopsy (EMB) diagnostic yield is low (~20%) but remains gold standard for definitive diagnosis in cardiomyopathies.
A machine learning random forest model using non-invasive predictors (notably right ventricular late gadolinium enhancement on cardiac MRI) predicts EMB diagnostic success with high accuracy (AUC ~0.9).
The scoring system may guide clinical decision-making to optimize patient selection for EMB, potentially reducing unnecessary invasive procedures.
Guideline-Based Recommendations
Diagnosis
EMB remains the gold standard for definitive diagnosis in many cardiomyopathies despite low yield.
Non-invasive predictors including cardiac MRI late gadolinium enhancement and biomarkers (NTproBNP, renal function) should be considered to assess likelihood of diagnostic EMB.
Management
Use of a validated machine learning-based predictive score may support decision-making on whether to perform EMB.
Consider EMB particularly when non-invasive data suggest high probability of diagnostic yield, such as presence of right ventricular LGE.
Monitoring & Follow-up
Monitor clinical status and biomarker levels (e.g., NTproBNP) alongside imaging findings to reassess diagnostic strategy.
Follow-up should integrate EMB results with clinical and imaging data to guide therapy.
Risks
EMB is invasive with associated procedural risks; patient selection should balance diagnostic benefit against procedural risk.
Use of predictive scoring may reduce unnecessary EMB procedures and associated complications.
Patient & Prescribing Data
Heart failure patients undergoing evaluation for cardiomyopathy with EMB
Machine learning score incorporating cardiac MRI and biomarkers can stratify patients by likelihood of diagnostic EMB, aiding personalized clinical decisions.
Clinical Best Practices
Incorporate cardiac MRI assessment of late gadolinium enhancement, especially in the right ventricle, when considering EMB.
Utilize non-invasive biomarkers such as NTproBNP and renal function tests to complement imaging data in predictive modeling.
Apply validated machine learning tools to improve patient selection for EMB and optimize diagnostic yield.
Ensure multidisciplinary evaluation including cardiology, imaging, and pathology expertise when interpreting EMB results.
by Christian Basile, Christian L. Polte, Piero Gentile, Entela Bollano, Araz Rawshani, Anders Oldfors, Charlotta Ljungman, Sven-Erik Bartfay, Pia Dahlberg, Clara Hjalmarsson, Marie Björkenstam, Elena Gualini, Antonio Cannatá, Patrizia Pedrotti, Andrea Garascia, Gianluigi Savarese, Aldo Pietro Maggioni, Kristjan Karason, Emanuele Bobbio
