Abstract 14920: CT Attenuation of Abdominal Adipose Tissue and its Correlation to ¹⁸FDG-Activity in PET - A Potential Noninvasive Marker of Metabolic and Inflammatory Activity

Abstract

Background: Abdominal visceral adipose tissue (VAT) is more strongly associated with cardiovascular disease than subcutaneous adipose tissue (SAT), potentially due to differences in metabolic activity and secretion of inflammatory cytokines. We hypothesized that the computed tomography (CT) attenuation of VAT but not of SAT correlates with its 2'[[Unable to Display Character: ‑]][(18)F]fluoro-2'-deoxy-D-glucose (¹⁸FDG) activity as measured by positron emission tomography (PET).

Methods: Male subjects with known human immunodeficiency virus (HIV) infection and lipodystrophy underwent PET/CT scanning (Biograph 64, Siemens Medical Solution). Subjects rested for 2 hours prior to the scan wearing a cooling vest with circulating water at 5°C (Polar Products). At the level of the mid L4 vertebral body, area and mean CT attenuation (mHU) were measured for VAT and SAT, while adipose tissue was defined as voxels with CT attenuation between -195 and -45 Hounsfield units (HU). From co-registered PET data, mean standardized uptake values (mSUV) were determined across adipose tissue voxels. An upper SUV threshold of 1 was set to exclude potential brown fat and/or partial volume artifacts. Spearman correlation was used to determine relationships.

Results: From 11 subjects (age 50 [40-57] yrs, BMI of 24.3 [20.9-28.8] kg/m²), VAT and SAT area were 120.7 [30.8-185.9] cm² and 159.7 [57.6-216.4] cm² respectively. Both, the CT attenuation and the ¹⁸FDG-activity were significantly higher in VAT than in SAT (mHU: -94.7 [-102.4-(-84.0)] vs. -99.9 [-109.7-(-96.9)] HU; p=0.005, and mSUV: 0.65 [0.51-0.78] vs. 0.15 [0.12-0.18]; p=0.001; respectively). The CT attenuation of VAT but not SAT correlated significantly with its ¹⁸FDG-activity (r=0.89, p=0.0002 vs. r=0.40, p=0.23; respectively). For VAT, this correlation remained significant after adjustment for HIV-related differences (r=0.93, p=0.008), Framingham Risk score (r=0.89, p=0.0006), or BMI (r=0.69, p=0.03).

Conclusion: In this small sample of HIV subjects with lipodystrophy, we demonstrate that the mean CT attenuation of VAT but not SAT is highly correlated to its ¹⁸FDG[[Unable to Display Character: ‑]]activity independent of HIV related patient characteristics or BMI. Furthermore, both mean CT attenuation and ¹⁸FDG[[Unable to Display Character: ‑]]activity are significantly higher in VAT than in SAT.