Evaluation of two software programs for cross-sectional body composition analysis on abdominal computed tomography scans of patients with cirrhosis.

Interobserver intraclass correlation coefficients (ICCs) were highly equivalent for all body composition measures analyzed.

• ICC range for interobserver comparisons was 0.984-0.999 (P < 0.001)
• Measures included CSMA, SMI, MA, VAT, VATI, SAT, and SATI
• Fifty patients were randomly selected from two centers
• Two observers independently evaluated each CT scan

Intersoftware intraclass correlation coefficients were highly equivalent for all body composition measures.

• ICC range for intersoftware comparisons was 0.984-0.999 (P < 0.001)
• The two software programs compared were SliceOmatic V.5.0 (Magog, Canada) and Synapse 3D, Fujifilm
• All seven body composition parameters showed excellent agreement between the two software programs

Pearson's correlation coefficients were excellent for all interobserver and intersoftware comparisons.

• Pearson's correlation coefficient range was 0.970-0.998 across all comparisons
• Excellent correlations were found for both interobserver and intersoftware comparisons
• All body composition parameters including muscle and adipose measures demonstrated this level of correlation

The study evaluated cross-sectional body composition analysis using abdominal CT scans in patients with cirrhosis from two centers.

• Fifty patients were randomly selected from two centers
• Body composition measures analyzed included cross-sectional muscle area (CSMA), skeletal muscle index (SMI), muscle attenuation (MA), visceral adipose tissue area (VAT), visceral adipose tissue index (VATI), subcutaneous adipose tissue area (SAT), and subcutaneous adipose tissue area index (SATI)
• Abdominal CT evaluation was identified as the technique that is better evaluated for body composition in this population
• Both interobserver and intersoftware agreements were assessed