The goal of this observational and diagnostic study is to develop and validate an artificial intelligence assisted approach for coronary computer tomography angiography-(CCTA)-based screening and diagnosis of cardiomyopathies in patients with suspected coronary artery diseases. This study aims to develop a computerized CCTA interpretation using artificial intelligence for multi-label classification task to assist cardiomyopathy diagnosis in the clinical workflow.
A clinical diagnosis of cardiomyopathies, including hypertrophic cardiomyopathy, dilated cardiomyopathy, restrictive cardiomyopathy, cardiac amyloidosis, myocarditis, arrhythmogenic right ventricular cardiomyopathy, and coronary artery disease/ischemic heart disease.
At least one CCTA before surgery or implantable device treatment.
No recorded diagnosis of cardiomyopathy or undetermined type of cardiomyopathy.
A clinical diagnosis of secondary cardiac abnormalities due to other organic or systemic diseases.
Surgery or implantable device treatment before CCTA examination.
Inclusion Criteria: participants with at least one CCTA examination.
Consecutive candidates have at least one CCTA between 1/1/2014 and 31/12/2024 will be collected.
Case-control
Cross-sectional
CCTAI model
Diagnostic testUsing a derivative sub-cohort, the investigators aim to first develop an CCTA-based AI-assisted (CCTAI) screening model to distinguish patients with cardiac abnormalities from those normal controls. Second, the investigators target at developing a CCTAI diagnostic model with multi-classification output of cardiomyopathy diagnosis. Both models will be tested in internal validation cohort and external validation cohort.
Diagnostic performance
The performance of the AI models is evaluated by assessing their sensitivity, specificity, precision and F1 score (harmonic mean of the predictive positive value and sensitivity), with two-sided 95% CIs, as well as the AUC of the ROC with two-sided CIs. The F1 score is complementary to the AUC, which is particularly useful in the setting of multiclass prediction and less sensitive than the AUC in settings of class imbalance. For an aggregate measure of model performance, the investigators compute the class frequency-weighted mean for the F1 score and the AUC. Other diagnostic performance assessing metrics include true-positive rate, true-negative rate, false-positive rate, false-negative rate, precision, sensitivity (recall), specificity, positive predictive value, and negative predictive value.
Click on the lead sponsor to view all of their trials.
This trial is not recruiting at the moment. You can still explore other options:
