Estimating 3-D whole-body composition from a chest CT scan.

Area of Science:

• Radiology
• Medical Imaging
• Biomedical Engineering

Background:

• Estimating whole-body composition from limited computed tomography (CT) scans presents clinical challenges.
• Accurate body composition assessment is crucial for various medical applications.

Purpose of the Study:

• To determine if whole-body composition, including visceral adipose tissue (VAT), subcutaneous adipose tissue (SAT), intermuscular adipose tissue (IMAT), skeletal muscle (SM), and bone, can be reliably estimated using only chest CT scans.

Main Methods:

• A cohort of 97 lung cancer patients with both chest and whole-body CT scans was analyzed.
• In-house software automatically segmented and quantified tissue volumes (VAT, SAT, IMAT, SM, bone).
• Multivariate linear regression models were developed using chest CT data, incorporating demographics and lung volume, and validated with ten-fold cross-validation.

Main Results:

• High R-squared values were achieved for estimating whole-body SAT (0.901), VAT (0.929), IMAT (0.900), total fat (0.933), SM (0.928), and bone (0.918).
• Mean absolute differences (MADs) indicated accurate estimations across all tissue types, with percentages ranging from 4.32% ± 3.38% for bone to 29.68% ± 61.99% for VAT.

Conclusions:

• The developed algorithm demonstrates significant potential for estimating whole-body composition from standard chest CT scans.
• Body composition measurements derived from chest CT scans are more accurate than those obtained from limited vertebral regions.