Body composition (BC) measurements derived using artificial intelligence (AI) are strongly linked to cardiometabolic risk, but after adjusting for BMI and waist circumference (WC), only visceral adipose tissue (VAT) proportion and skeletal muscle fat fraction (SMFF), as well as skeletal muscle (SM) proportion in men only, provide additional prognostic value, according to a prospective cohort study published Sept. 30 in Annals of Internal Medicine.

Using UK Biobank data, Matthias Jung, MD, et al., analyzed whole-body MRIs from 33,432 adults (mean age, 65 years; 53% women; 100% White; mean BMI, 26 kg/m²) without diabetes or history of cardiovascular events. They applied an AI tool to the MRIs to derive 3D BC measures, including subcutaneous adipose tissue, VAT, SM and SMFF, and subsequently calculated their relative distribution to assess cardiometabolic outcomes.

Results showed that after a median 4.2 years of follow-up, 1% of women and 2% of men had incident diabetes mellitus (DM) and 1% of women and 2% of men experienced a major adverse cardiovascular event (MACE), defined as myocardial infarction, ischemic stroke or cardiovascular mortality.

Findings also revealed that as participants aged, adipose tissue compartments and SMFF increased, whereas SM decreased. In sex-stratified analyses where models adjusted for age, smoking and hypertension, greater adiposity measures and low SM proportion were linked to higher rates of DM and MACE in both sexes.

However, after additionally adjusting for BMI and WC, only elevated VAT proportion and high SMFF (top fifth percentile in the cohort for each) were associated with a higher risk for DM (respective adjusted hazard ratios [aHRs], 2.16 and 1.27 in women and 1.84 and 1.84 in men) and MACE (aHR, 1.37 and 1.72 in women and 1.22 and 1.25 in men).

Moreover, in men only, those in the bottom fifth percentile of SM proportion had an increased risk for DM (aHR, 1.96) and MACE (aHR, 1.55).

Jung and colleagues write that "automated MRI-based BC analysis is accurate and feasible," and captures details in only three minutes. "After further validation in diverse populations, this approach may enable opportunistic assessment of BC from routine imaging to identify patients at high cardiometabolic risk."