E2193. Single vs. Multislice Thoracic Body Composition Analysis in Low-Dose CT
Authors
Jeremy Stephan;
Rush University Medical Center
Andrew Neilson;
Rush University Medical Center
Sanjib Basu;
Rush University Medical Center
Karteek Popuri;
Memorial University of Newfoundland
Mirza Beg;
Simon Fraser University
Christopher Seder;
Rush University Medical Center
Palmi Shah;
Rush University Medical Center
Objective:
This study aimed to investigate a sex-specific correlation between single-slice areas and multislice volumes of body composition-based analysis at varying thoracic levels in low-dose CT (LDCT) scans for lung cancer screening.
Materials and Methods:
Patients who underwent lung cancer screening with LDCT scans between 2015 and 2021 with at least 1-year follow-up were identified from an institutional database. Middle vertebral single slice areas (cm<sup>2</sup>) ranging from T4 - T12 and multislice T4 - T12 volumes (cm<sup>3</sup>) of skeletal muscle (SKM), epicardial adipose (EpAT), subcutaneous adipose (SAT), and intramuscular adipose (IMAT) tissue were assessed using an automated multislice segmentation platform. Sex-specific z-scores controlling for body mass index (BMI) were calculated for tissue subtypes across all patients for the single and multislice measurements. Pearson coefficients were used to quantify the degree of correlation for single and multislice measurements, and Fischer’s z-transformation was used to assess the between-sex variance.
Results:
This study included 1,829 LDCT scans from patients undergoing lung cancer screening from 2015 - 2021. The cohort had a median (IQR) age of 66 (IQR 61 - 70), with 53% (974/1829) were women, 49% (900/1829) were current smokers, median BMI of 29 (IQR 25 - 33), and median smoking history of 40 packyears (IQR 34 - 47). Positive correlation between single area and multislice volumes ranged from strong to very strong across tissues for men and women (SKMwomen <em>R</em> = 0.702, <em>p</em> = 0; SKMmen <em>R</em> = 0.769, <em>p</em> = 0; EpATwomen <em>R</em> = 0.798, <em>p</em> = 0; EpATmen <em>R</em> = 0.793, <em>p</em> = 0; SATwomen <em>R</em> = 0.891, <em>p</em> = 0; SATmen <em>R</em> = 0.895, <em>p</em> = 0; IMATwomen <em>R</em> = 0.828, <em>p</em> = 0; IMATmen <em>R</em> = 0.837, <em>p</em> = 0). The vertebral level at which the single slice area most strongly correlated with the multislice volume varied with tissue subtype (SKM = T4, EpAT = T8, SAT = T6, IMAT = T6). Single-slice area SKM at T4 correlated with multislice SKM volume more strongly in men than women (0.769 vs 0.702, <em>p</em> = 0.004).
Conclusion:
Single-slice area and multislice volume measurements of SKM, EpAT, SAT, and IMAT in LDCT scans are strongly correlated. Single-slice areas derived from LDCT to quantify body composition vary with tissue subtype and sex, suggesting that multislice measurements may offer an advantage for tissue analysis. Body composition to assess cancer-associated cachexia in patients with lung cancer is a prognostic factor for mortality. This study reports the optimal vertebral levels for body composition analysis from lung cancer screening scans.
