Key result
CNN-automated measures of mid-thigh cross-sectional area from MRI were highly accurate compared to manual segmentation and strongly predicted appendicular lean mass (r = 0.87) and knee extensor strength (r = 0.76) in obese adults.
Why the study?
To train and test a machine learning model to automatically measure mid-thigh muscle cross-sectional area to provide rapid estimation of appendicular lean mass and predict knee extensor torque in obese adults.
Population
47 obese adults (BMI 30–40 kg/m 2, age 30–50 years)
Comparison
Automated CNN quantification vs manual segmentation of mid-thigh CSA
Authors
Loading...
CNN automation of mid-thigh CSA may aid lean mass and strength evaluation in obesity; leaves open clinical adoption pending prospective validation.
Cohort (n=47)
No
Effect estimate: r = 0.87
p-value: p=<0.001
A convolutional neural network can accurately automate the measurement of mid-thigh cross-sectional area from MRI scans to estimate appendicular lean mass and knee extensor strength in obese adults.
Bodkin et al. (2022) conducted a cohort in Obesity (n=47). CNN-automated measurement of mid-thigh cross-sectional area from MRI vs. Manual segmentation and DXA imaging was evaluated on Correlation between CNN-measured mid-thigh cross-sectional area and appendicular lean mass (ALM) (r = 0.87, p=<0.001). CNN-automated measures of mid-thigh cross-sectional area from MRI were highly accurate compared to manual segmentation and strongly predicted appendicular lean mass (r = 0.87) and knee extensor strength (r = 0.76) in obese adults.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: