Harnessing the power of artificial intelligence, researchers are unlocking the potential of whole-body MRI to predict health risks, paving the way for smarter, personalized prevention strategies.
Study: Body composition analysis based on deep learning from whole-body magnetic resonance imaging to predict all-cause mortality in a large western population. Image credit: Juice Flair / Shutterstock
In a recent study published in the journal eBioMedicineresearchers in Germany and the United States developed and validated a deep learning framework for automated volumetric body composition analysis from whole-body Magnetic Resonance Imaging (MRI) and evaluated its prognostic value for predicting all-cause mortality in a large Western population.
Background
Body composition measures, including adipose tissue compartments and skeletal muscle, have shown strong associations with clinical outcomes and are emerging as important imaging biomarkers to improve personalized risk assessment. However, their routine quantification by imaging modalities such as MRI remains limited in clinical workflows due to time and resource constraints. With its superior ability to differentiate tissue types and assess their distribution, MRI offers significant potential for comprehensive analysis of body composition.
The study highlights that manual quantification is labor intensive, while automated approaches could overcome these obstacles. Fully automated volumetric approaches based on Artificial Intelligence (AI) could overcome current limitations, enabling more accurate and scalable assessments. These findings highlight the importance of developing standardized tools to ensure clinical applicability in diverse populations.
About the Study
The study used data from two large population-based cohort studies: the UK Biobank (UKBB), which included participants aged 45-84, and the German National Cohort (NAKO), with participants aged 40-75. Both studies collected comprehensive clinical data and used a detailed MRI protocol, including whole-body T1-weighted Dixon 3D Volumetric Interference Breath-Keeping Interference (3D VIBE) sequences, used to analyze body composition. Ethical approvals were obtained and informed consent was obtained from all participants.
The primary objective was to develop a deep learning framework for the automated quantification of volumetric body composition measures such as subcutaneous adipose tissue (SAT), visceral adipose tissue (VAT), skeletal muscle (SM), skeletal muscle fat fraction (SMFF) and intramuscular adipose tissue (IMAT), using whole-body MRI. The performance of the framework was evaluated in the UKBB, focusing on its predictive value for all-cause mortality. The study also aimed to assess correlations between whole-body volumetric measurements and traditional single-slice body composition assessment at the L3 vertebra.
The deep learning model used Dixon sequence imaging inputs to generate segmentation masks, allowing quantification of the volumetric and somatic composition of a slice. Experienced radiologists performed manual annotations for model training and independently validated them. Statistical analyzes included survival modeling and association assessments, using harmonized data sets to minimize allocation differences.
Study Results
The UKBB cohort included 36,317 participants (18,777 women and 17,540 men) with a mean age of 65.1 ± 7.8 years and a mean body mass index (BMI) of 25.9 ± 4.3 kg/m². Body composition analysis revealed higher volumetric subcutaneous adipose tissue (VSAT), skeletal muscle fat fraction (VSMFF) and intramuscular adipose tissue (VIMAT) in females, while males showed greater visceral adipose tissue volume (VVAT) and skeletal muscle volume (VSM). (all p < 0.0001). Similar trends were seen among the 23,725 NAKO participants, whose mean age was 53.9 ± 8.3 years with a mean BMI of 27 ± 4.7 kg/m², as well as body composition measures of the single-incision area at the L3 vertebra for both cohorts.
During a median follow-up period of 4.77 years in the UKBB, 634 deaths (1.7%) were recorded. Kaplan-Meier survival curves showed that participants in the lowest 10th percentile of VSM and the highest 10th percentile of VSMFF and VIMAT showed significantly higher mortality rates (log-rank p <0.0001). Adjusted Cox regression analyzes revealed that lower VSM (aHR: 0.86, 95% CI [0.81–0.91]p < 0.0001) was associated with a reduced risk of mortality, whereas higher VSMFF (aHR: 1.07, 95% CI [1.04–1.11]p < 0.0001) and VIMAT (aHR: 1.28, 95% CI [1.05–1.35]p < 0.0001) were associated with increased risk. In contrast, volumetric VSAT and VVAT measurements showed no substantial association with mortality after adjustment for traditional risk factors.
Analysis of single-slice area measurements at L3 yielded results consistent with volumetric measurements, with lower skeletal muscle area (ASM) and higher fat fraction (ASMFF) and intramuscular adipose tissue (AIMAT) associated with mortality. However, after full adjustment, these associations weakened for ASM and AIMAT. Reclassification analyzes showed that volumetric measurements were more effective in identifying high-risk individuals than single-slice measurements, as evidenced by a significant net improvement in reclassification for skeletal muscle (NRI = 0.053, 95% CI [0.016–0.089]).
Correlation analysis between whole-body and single-slice volumetric measurements showed strong agreement at specific vertebral levels, such as L3 for VAT (R = 0.892) and SM (R = 0.944). These findings were replicated in the NAKO cohort, although the association differed significantly by BMI and gender strata. The deep learning framework demonstrated high accuracy, with Dice coefficients exceeding 0.86 and strong agreement between manual and automated segmentation results (r > 0.97).
conclusions
This study developed an automated deep learning framework for whole-body MRI-based body composition analysis and evaluated its prognostic value for predicting mortality in more than 30,000 individuals. Volumetric measures, including SM, SMFF, and IMAT, were independent predictors of mortality, outperforming traditional single-section approaches, which showed variable associations influenced by gender and BMI. Despite these strengths, the study acknowledged limitations, such as cohort demographics representing predominantly Western populations and limited follow-up duration, which could affect generalizability.
Future research should investigate the clinical complementarity of volumetric analysis with MRI in various populations and imaging protocols.
