Key result
Fully automated adipose segmentation showed excellent agreement with manual segmentation and reduced analysis time to 2 seconds per slice compared with 8 minutes per slice.
Why the study?
Does a fully automated adipose segmentation method improve analysis time while maintaining precision compared to manual segmentation in subjects undergoing abdominal MRI?
Does a fully automated adipose segmentation method improve analysis time while maintaining precision compared to manual segmentation in subjects undergoing abdominal MRI?
A fully automated MRI adipose segmentation method is highly accurate and significantly faster than manual segmentation across diverse patient populations.
Answers are grounded in this paper and cite their sources.
May expedite adipose quantification on abdominal MRI; leaves open prospective validation and outcome impact.
PURPOSE: To validate a fully automated adipose segmentation method with magnetic resonance imaging (MRI) fat fraction abdominal imaging. We hypothesized that this method is suitable for segmentation of subcutaneous adipose tissue (SAT) and intra-abdominal adipose tissue (IAAT) in a wide population range, easy to use, works with a variety of hardware setups, and is highly repeatable. MATERIALS AND METHODS: Analysis was performed comparing precision and analysis time of manual and automated segmentation of single-slice imaging, and volumetric imaging (78-88 slices). Volumetric and single-slice data were acquired in a variety of cohorts (body mass index [BMI] 15.6-41.76) including healthy adult volunteers, adolescent volunteers, and subjects with nonalcoholic fatty liver disease and lipodystrophies. A subset of healthy volunteers was analyzed for repeatability in the measurements. RESULTS: The fully automated segmentation was found to have excellent agreement with manual segmentation with no substantial bias across all study cohorts. Repeatability tests showed a mean coefficient of variation of 1.2 ± 0.6% for SAT, and 2.7 ± 2.2% for IAAT. Analysis with automated segmentation was rapid, requiring 2 seconds per slice compared with 8 minutes per slice with manual segmentation. CONCLUSION: We demonstrate the ability to accurately and rapidly segment regional adipose tissue using fat fraction maps across a wide population range, with varying hardware setups and acquisition methods.
No takes yet. Share an insight, caveat, or question.
Addeman et al. (2014) studied Healthy volunteers, nonalcoholic fatty liver disease, and lipodystrophies. Fully automated adipose segmentation method vs. Manual segmentation was evaluated on Precision and analysis time of segmentation. Fully automated adipose segmentation showed excellent agreement with manual segmentation and reduced analysis time to 2 seconds per slice compared with 8 minutes per slice.
Synapse has enriched 10 related papers. Consider them for comparative context: