Hypertrophic Cardiomyopathy Federated Learning Implementation Platform
- Conditions
- Hypertrophic Cardiomyopathy
- Registration Number
- NCT06461468
- Lead Sponsor
- American Heart Association
- Brief Summary
HCM FLIP study is a two-phase protocol focusing on the detection of Hypertrophic Cardiomyopathy using Electrocardiograms and Echocardiograms through Federated Learning.
- Detailed Description
HCM FLIP (Hypertrophic Cardiomyopathy Federated Learning Implementation Platform) aim to build and test a model's system impact to detect hypertrophic cardiomyopathy (HCM) by training a machine learning (ML) model with electrocardiograms (ECGs) and echocardiograms (ECHOs). Approximately 10-1000 HCM cases and 30-10,000 age/sex-matched controls per institution, depending on size, will be included in the study. We hypothesize that a federated ML model will discriminate cases of HCM from those without HCM in a real-world setting.
Recruitment & Eligibility
- Status
- ENROLLING_BY_INVITATION
- Sex
- All
- Target Recruitment
- 1000
- Patients with maximum left ventricular wall thickness exceeding 15 mm (including the right ventricular component of the septum) without any other explanation for ventricular hypertrophy (e.g., severe hypertension, cardiac amyloidosis, severe AS, as determined by local investigators). The measurement could be made in an ECHO or on magnetic resonance imaging (MRI).
- Patients must have > one (1) ECG and/or > one (1) ECHO available that meet minimum compatibility requirements. If multiple ECGs and ECHOs are available per patient, then all available data meeting compatibility requirements will be used for model training purposes.
HCM-Labeled Case
- Any sign of infiltration found in cardiac MRI, if performed.
Control Case (Non-HCM) Inclusion Criteria:
- No diagnosis of HCM
- Age/sex are matched to HCM cases (+/- 5 years, if possible; +/- 10 years if numbers do not permit).
- Patient must have > one (1) ECG and/or > one (1) ECHO available that meet minimum compatibility requirements. If multiple ECGs and ECHOs are available per patient, then all available data meeting compatibility requirements will be used for model training purposes.
Control Case (Non-HCM) Exclusion Criteria
- Suggestion of HCM in a clinically obtained ECHO or cardiac MRI report unless subsequently confirmed no diagnosis of HCM. Any new clinical information discovered during the study will be left to the discretion of the local investigator.
Study & Design
- Study Type
- OBSERVATIONAL
- Study Design
- Not specified
- Primary Outcome Measures
Name Time Method Diagnosis of HCM Through study completion, an average of 2 years The number/instances of HCM diagnoses as identified by the ML model as compared to clinical diagnosis confirmation. Due to model training and efficacy goals, HCM diagnosis determined clinically via EKG/ECHO reading will be compared to the ML model's capacity to identify HCM correctly and efficiently.
- Secondary Outcome Measures
Name Time Method Diagnosis of different types of HCM Through study completion, an average of 2 years Diagnosis of different types of HCM (i.e., apical, obstructive), HCM without hypertrophy, genetic positive/negative indicators, among others, as identified by the ML model as compared to clinical diagnosis confirmation. Due to model training and efficacy goals, HCM diagnosis determined clinically via EKG/ECHO reading will be compared to the ML model's capacity to identify HCM correctly and efficiently.
Trial Locations
- Locations (6)
Riverside Medical Center
🇺🇸Kankakee, Illinois, United States
University of Michigan Medical Center
🇺🇸Ann Arbor, Michigan, United States
Wooster Community Hospital
🇺🇸Wooster, Ohio, United States
Thomas Hospital
🇺🇸Fairhope, Alabama, United States
The University of Texas Southwestern Medical Center
🇺🇸Dallas, Texas, United States
Johns Hopkins University School of Medicine
🇺🇸Baltimore, Maryland, United States