AI model maps body composition from whole-body MRI

Researchers used AI to analyze whole-body MRI scans from more than 66,000 participants and map how fat and muscle are distributed by age, sex, and height, according to RSNA.

The study, published in Radiology, found that skeletal muscle quality and quantity, not only visceral fat, were associated with future diabetes, major cardiovascular events, and all-cause mortality.

The retrospective study included 66,608 individuals from the UK Biobank and the German National Cohort who underwent whole-body MRI between April 2014 and May 2022. The cohort had a mean age of 57.7 years, included 34,443 men, and had a mean body mass index of 26.2.

Using an open-source, fully automated deep-learning framework, the team calculated age-, sex-, and height-normalized body composition metrics from MRI scans. These included subcutaneous adipose tissue, visceral adipose tissue, skeletal muscle, skeletal muscle fat fraction, and intramuscular adipose tissue.

The measures were expressed as z-scores to show how far each participant differed from expected values for their age, sex, and height. The researchers then assessed whether low, middle, or high z-score categories predicted diabetes, major adverse cardiovascular events, and all-cause mortality.

High visceral fat was associated with a 2.26-fold higher risk of future diabetes. High intramuscular fat was associated with a 1.54-fold higher risk of future major cardiovascular events, while low skeletal muscle was associated with a 1.44-fold higher risk of all-cause mortality beyond cardiometabolic risk factors.

The authors also generated age-, sex-, and height-normalized reference curves for body composition measures. An open-source z-score calculator is available for researchers and clinicians to normalize body composition data against the reference values.

RSNA said the approach could allow body composition data to be extracted from routine CT or MRI body scans when imaging is already available, rather than requiring dedicated whole-body MRI.

Next steps include validating the reference curves in clinical populations, including patients with cancer, and developing disease-specific reference values for other patient groups.