Coronary Artery Disease
Conditions
Keywords
Artificial intelligence, Facial characteristics
Brief summary
The purposes of this study are 1) to explore the association between multi-dimension facial characteristics and the increased risk of coronary artery diseases (CAD); 2) to evaluate the diagnostic efficacy of multi-dimension appearance factors for coronary artery diseases.
Detailed description
Previous study demonstrated the feasibility of using deep learning to detect coronary artery disease based on facial photos. However, several limitations made the algorithm hard to be utilized in clinical practice, including low specificity and lack of external validation. Adding multi-dimension facial characteristics may further increase the algorithm effect. Thus, the investigators designed a single-center, cross-sectional study to explore the association between multi-dimension facial characteristics and CAD and to evaluate the predictive efficacy of multi-dimension appearance factors for CAD. The investigators will recruit patients undergoing coronary angiography or coronary computer tomography angiography. Patients' baseline information and multi-dimension facial images will be collected. The investigators will train and validate a deep learning algorithm based on multi-dimension facial photos.
Interventions
OTHERNo intervention
No intervention
Sponsors
China National Center for Cardiovascular Diseases
Study design
Observational model
OTHER
Time perspective
PROSPECTIVE
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
Yes
Inclusion criteria
* Undergoing coronary angiography or coronary computer tomography angiography * Written informed consent
Exclusion criteria
* Prior percutaneous coronary intervention (PCI) * Prior coronary artery bypass graft (CABG) * Screening coronary artery disease before treating other heart diseases * Without blood biochemistry outcome * With artificially facial alteration (i.e. cosmetic surgery, facial trauma or make-up) * Other situations which make patients fail to be photographed
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Area under receiver operating curve (AUC) | At the end of enrollment (1 mouth) | Area under receiver operating curve of algorithm assessed in test group |
Secondary
| Measure | Time frame | Description |
|---|---|---|
| Sensitivity of algorithm | At the end of enrollment (1 mouth) | Sensitivity of algorithm assessed in test group |
| Specificity of algorithm | At the end of enrollment (1 mouth) | Specificity of algorithm assessed in test group |
| Positive predictive value (PPV) | At the end of enrollment (1 mouth) | PPV of algorithm assessed in test group |
| Negative predictive value (NPV) | At the end of enrollment (1 mouth) | NPV of algorithm assessed in test group |
| Diagnostic accuracy rate | At the end of enrollment (1 mouth) | Diagnostic accuracy rate of algorithm assessed in test group |
Countries
China
Outcome results
None listed