Article Summary
A new study found that measuring muscle quality on coronary CT scans can predict the risk of heart attack and death over 10 years, with lower muscle density independently associated with significantly higher cardiac risk.
- Machine learning analysis of 1,722 coronary CT scans identified skeletal muscle attenuation as a key predictor of heart disease outcomes
- Patients with low muscle density had 1.85 times higher mortality risk and 1.58 times higher heart attack risk over 10 years
- This is the first study to assess long-term prognostic implications of body composition measured directly from coronary CT images
- Skeletal muscle attenuation remained an independent predictor even after adjusting for coronary calcium score and other risk factors
- Standardized protocols and integration with electronic health records are needed before clinical implementation
Measuring muscle quality on coronary CT angiography (CCTA) can predict both death and heart attack risk over 10 years, according to a study published in Radiology.
Researchers used a machine-learning model to analyze body composition from CCTA scans of 1,722 patients in the SCOT-HEART trial and found that lower skeletal muscle attenuation independently predicted higher rates of all-cause mortality and myocardial infarction, noted lead author Alan Guimaraes, PhD, of the University of Edinburgh in Scotland, and colleagues.
“To our knowledge, this is the first study to assess the prognostic implications of multiorgan body composition on CCTA images with long-term follow-up,” the group wrote.
In the SCOT-HEART (Scottish Computed Tomography of the Heart) trial, CCTA with standard care lowered the risk of myocardial infarction or death from coronary artery disease over 10 years by 41% compared to standard care alone. Previous studies of the trial's imaging data have linked coronary plaque characteristics, adipose tissue attenuation, and liver attenuation to cardiovascular risk, the authors explained, and in this study they further evaluated associations between machine learning-derived multiorgan body composition and outcomes.
The researchers processed wide field-of-view CCTA images from 1,722 patients using a machine-learning model called TotalSegmentator. The model segmented 25 organ structures, with volume and mean attenuation calculated for each structure. The researchers then built multivariable Cox proportional hazards models to assess associations with all-cause mortality and myocardial infarction (MI) over a median follow-up of 10 years, adjusting for age, sex, and scan length.
Over the follow-up period, 133 patients (7.72%) died and 106 (6.16%) experienced a fatal or nonfatal MI. Increased skeletal muscle attenuation was associated with lower all-cause mortality (HR, 0.61; p < 0.001) after multivariable adjustment. MI was associated with increased myocardial volume (HR, 1.09; p = 0.018) and decreased rib (HR, 0.98; p = 0.043) and skeletal muscle (HR, 0.69; p = 0.002) attenuation after multivariable adjustment.
However, when further adjusted for coronary calcium score, only skeletal muscle attenuation was associated with MI (HR, 0.72; p = 0.007). Patients with skeletal muscle attenuation below the median had a higher risk of mortality (HR, 1.85; p < 0.001) and MI (HR, 1.58; p = 0.022), the researchers reported.
“Multiorgan body composition analysis using coronary CT angiography provided additional prognostic information, among which skeletal muscle attenuation was particularly important,“ the group wrote.
Ultimately, to facilitate future clinical integration, the authors suggested that standardized acquisition and analysis protocols must be established, automated measurements must be evaluated across diverse populations and scanner types, clinically relevant thresholds for intervention must be defined, and such protocols should be integrated with electronic health records.
"Further research is needed to understand how to use this additional information for individual risk stratification and management guidance," the authors wrote.
The full study is available here.
