Why the study?
The prognostic value of automated deep learning-derived CAC scores on low-dose attenuation correction CT images during MPI was unknown.
Does automated coronary calcium scoring derived by deep learning on non-ECG gated CT images predict MACE in patients suspected of CAD?
Population
Chest-pain cohort undergoing RubidiumPET/CT without a history of coronary revascularization
Comparison
High CAC score (≥400) vs low CAC score (<400)
Design
Cohort study
Follow-up
4 years
Key result
Automated deep learning-derived high CAC scores (≥400) on non-ECG gated CT images independently predicted MACE (HR 2.19; 95% CI 1.43-3.35) and improved risk reclassification.
Authors
Loading...
May refine MACE risk stratification beyond perfusion imaging; leaves open need for prospective trials before clinical adoption.
Cohort
Does automated coronary calcium scoring derived by deep learning on non-ECG gated CT images predict MACE in patients suspected of CAD?
Effect estimate: HR 2.19 (95% CI 1.43-3.35)
p-value: p=<0.001
Automated deep learning-derived CAC scores from non-ECG gated PET/CT attenuation correction images provide independent prognostic value for MACE beyond myocardial perfusion imaging alone.
Dekker et al. (2021) conducted a cohort in Suspected coronary artery disease. Automated deep learning-derived CAC score vs. Low CAC score (<400) was evaluated on Major adverse cardiovascular events (all cause death, late revascularization >90 days, or nonfatal myocardial infarction) (HR 2.19, 95% CI 1.43-3.35, p=<0.001). Automated deep learning-derived high CAC scores (≥400) on non-ECG gated CT images independently predicted MACE (HR 2.19; 95% CI 1.43-3.35) and improved risk reclassification.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: