Abstract
Aim: To investigate the reliability of multicompartmental volumetric mesenchymal segmentations on MRI and their correlations with anthropomorphic and clinical parameters. Materials and methods: A consecutive series of middle-age (35–50 year old) female volunteers with variable body mass index (BMI) and MRI scans performed as a part of the Dallas Heart Study were included. A semi-automatic segmentation tool was used to partition different mesenchymal tissues- fat, muscle, and bone on MRI of pelvis. Total volumes of each compartment were calculated and compared between overweight/obese (BMI> = 25 kg/m 2 ) and non-obese (BMI < 25 kg/m 2 ) groups, and with physical performance measurements, i.e. mean activity counts per minute (MVPA) and cardiorespiratory fitness (CRF) estimated by submaximal treadmill test (TT). Kruskal Wallis, Mann-Whitney U test, intraclass correlation coefficient (ICC) and Spearman correlations were used. P value <0.05 was considered statistically significant. Results: There were statistically significant positive correlations between fat volume and BMI (p < 0.0001), muscle volume and height (p = 0.03), and bone volume and height (p < 0.0001). Significant inverse correlations were found between bone volume and BMI (p = 0.002). Fair to good interobserver reliability was seen with muscle and fat volumes (ICC = 0.43–0.64) and excellent reliability was seen with bone volumes (ICC = 0.78–0.79). Statistically significant inverse correlations were found between MVPA and age (p = 0.01), and TT with BMI and weight (p = 0.01, 0.03). Conclusion: Multi-compartment mesenchymal tissue volume quantification on pelvic MRI is reliable in females. Inverse correlation of bone volume with BMI has potential implications for future risk of fracture.
Original language | English (US) |
---|---|
Pages (from-to) | 37-43 |
Number of pages | 7 |
Journal | European Journal of Radiology |
Volume | 112 |
DOIs | |
State | Published - Mar 2019 |
Keywords
- Bone
- Fat
- MRI
- Muscle
- Obesity
- Segmentation
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging