Abstract
Body composition is increasingly recognized as an important factor in cancer outcomes. Use of computed tomography (CT) in cancer care provides the opportunity to accurately quantify whole-body lean and adipose tissues from images at the third lumbar spine. We sought to substantiate the use of routinely captured, single-slice chest CT images at the thoracic level for evaluation of skeletal muscle, residual lean tissue, and adiposity among pediatric solid tumor patients. We performed a retrospective analysis among children who underwent treatment for a solid tumor at Columbia University Irving Medical Center. Skeletal muscle (SM), residual lean tissue (RLT), and adipose tissue cross-sectional areas (cm(2) ) were analyzed at diagnosis and at first follow-up for disease evaluation (6-14 weeks). Imaging analysis was performed utilizing slice-O-matic image analysis software. Of the 57 patients identified, 39 had chest CT imaging that included intervertebral level T12-L1, and 22 also had concurrent imaging at L3. Correlation coefficients between body composition variables at T12-L1 and L3 were strong (r = 0.93-0.98). Paired t-test showed a significant decrease in SM (-4.2 ± 8.12, p = 0.003) and RLT (-10.7 ± 28.5, p = 0.025) as well as a trend toward a significant increase in visceral adipose tissue (3.10 ± 9.65, p = 0.052). Univariable analysis demonstrated a significant association between increasing age and increased SM loss (β = -0.496 with SE = 0.194, p = 0.011), and a lack of association between body mass index and body composition changes. We provide the first line of evidence that single-slice images from routinely obtained chest CT scans provide a simple, readily available mechanism for assessing body composition in pediatric solid tumor patients. Adverse body composition changes were observed, particularly among adolescents and young adults. Precis: Changes in body composition can be detected via routine CT images in pediatric patients undergoing treatment for solid tumors.