Artificial Intelligence Prediction of Cardiovascular Events Using Opportunistic Epicardial Adipose Tissue Assessments From Computed Tomography Calcium Score
Authors' conclusion
Preliminary findings indicate an opportunity to use finely tuned, explainable assessments on EAT for improved cardiovascular risk prediction.
Study context
- Population
- Patients with low-dose cardiac CT calcium score examinationsMACE-enriched cohort
- Participants
- 400 patients
- Comparison
- Fat-omics epicardial adipose tissue (EAT) model vs traditional EAT features (EAT volume, EAT mean HU) and BMI for MACE prediction
- Outcome
- MACE prediction (C-index)
Methodological detailsretrospective cohort, MACE-enriched cohort (56% event rate), training/testing sets (80%/20%), deep-learning EAT segmentation, Cox elastic-net, bootstrap validation, 148 initial EAT features extracted; Cox elastic-net for feature reduction
- Retrospective cohort
- MACE-enriched cohort (56% event rate)
- Training/testing sets (80%/20%)
- Deep-learning EAT segmentation
- Cox elastic-net
- Bootstrap validation
- 148 initial EAT features extracted; Cox elastic-net for feature reduction
Performance
As reported in the abstract
MACE prediction, C-index
Fat-omics
0.69
C-index
EAT-volume
0.53
C-index
EAT-mean-HU
0.55
C-index
BMI
0.57
C-index
Further findings
| Outcome | HR | P value |
|---|---|---|
| MACE, fat-omics high- vs low-risk (split at median risk) | 2.4 | P < 0.001 |
Original publication
Artificial Intelligence Prediction of Cardiovascular Events Using Opportunistic Epicardial Adipose Tissue Assessments From Computed Tomography Calcium Score. JACC Adv. Published 28 August 2024. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →