Deep-Learning for Epicardial Adipose Tissue Assessment With Computed Tomography: Implications for Cardiovascular Risk Prediction

West, HW; Siddique, M; Williams, MC; Volpe, L; Desai, R; Lyasheva, M; Thomas, S; Dangas, K; Kotanidis, CP; Tomlins, P; Mahon, C; Kardos, A; Adlam, D; Graby, J; Rodrigues, JCL; Shirodaria, C; Deanfield, J; Mehta, NN; Neubauer, S; Channon, KM; Desai, MY; Nicol, ED; Newby, DE; Antoniades, C; ORFAN Investigators

doi:10.1016/j.jcmg.2022.11.018

Prevention & lipids

Deep-Learning for Epicardial Adipose Tissue Assessment With Computed Tomography: Implications for Cardiovascular Risk Prediction

JACC Cardiovasc ImagingPublished 8 Feb 2023Diagnostic and prognostic study

Read the original article ↗PubMed

Cite

Authors' conclusion

Automated assessment of EAT volume is possible in CCTA, including in patients who are technically challenging; it forms a powerful marker of metabolically unhealthy visceral obesity

Study context

Population: Patients undergoing coronary computed tomography angiograms (CCTA) (ORFAN cohort; post-cardiac surgery cohort; SCOT-HEART trial)3,720 training scans

Comparison: Deep-learning automated epicardial adipose tissue (EAT) volume quantification vs human
Outcome: Machine vs human agreement for EAT volume (external validation)

Methodological detailsdeep-learning development and validation, external validation, longitudinal cohorts

Deep-learning development and validation
External validation
Longitudinal cohorts

Performance

As reported in the abstract

Concordance correlation coefficient, machine vs human

0.970

Further findings

Outcome	← Lower ←→ Higher →	Estimate	P value
CAD and AF (association with EAT volume, adjusted for risk factors incl. BMI)EAT volume (per SD increase) · CAD · after correction for risk factors (including BMI)	0.512	OR 1.13 (95% CI 1.04–1.30)	P = 0.01
CAD and AF (association with EAT volume, adjusted for risk factors incl. BMI)EAT volume (per SD increase) · AF · after correction for risk factors (including BMI)	0.512	OR 1.25 (95% CI 1.08–1.40)	P = 0.03
All-cause mortality, MI and stroke in SCOT-HEART (EAT volume, independent of risk factors) · 5-year follow-upEAT volume (per SD) · all-cause mortality	0.512	HR 1.28 (95% CI 1.10–1.37)	P = 0.02
All-cause mortality, MI and stroke in SCOT-HEART (EAT volume, independent of risk factors) · 5-year follow-upEAT volume (per SD) · MI	0.512	HR 1.26 (95% CI 1.09–1.38)	P = 0.001
All-cause mortality, MI and stroke in SCOT-HEART (EAT volume, independent of risk factors) · 5-year follow-upEAT volume (per SD) · stroke	0.512	HR 1.20 (95% CI 1.09–1.38)	P = 0.02
	0.512Ratio (log scale)

Parts of this Brief are withheld: the automatic checks could not confirm them against the abstract.

Original publication

Deep-Learning for Epicardial Adipose Tissue Assessment With Computed Tomography: Implications for Cardiovascular Risk Prediction. JACC Cardiovasc Imaging. Published 8 February 2023. doi:10.1016/j.jcmg.2022.11.018

Read the original article ↗

Composed by AI from the published abstract and checked against it. How Briefs are made →