Abstract
205
Objectives: Inflammation is a common mechanism in lung diseases, including those seen in nicotine dependence, a leading cause of morbidity and mortality worldwide. Both Electronic Nicotine Delivery Systems (ENDS; e-cigarettes) and cigarettes deliver nicotine to lungs, but measuring any resulting inflammation is a challenge. Inducible nitric oxide synthase (iNOS) is an enzyme constitutively expressed in lung epithelium that is specific for inflammation. A radiotracer targeting iNOS, [18F]6-(2-fluoro-propyl)-4-methylpyridin-2-amine ([18F]NOS), has shown promise in measuring cardiopulmonary inflammation. To guide selection of a blood pool for kinetic analysis, we evaluated impacts of pool selection (right atrium (RA), pulmonary artery (PA), or left ventricle (LV)) on subsequent model fits of lung tissue uptake of [18F]NOS.
Methods: Eleven subjects (mean 28yo, 20-41yo), including 5 healthy controls, 3 ENDS users, and 3 smokers underwent a 60-minute dynamic scan of their lungs in a Philips Ingenuity PET/CT scanner following injection of [18F]NOS (mean 192 MBq, 141-234 MBq). Patients provided informed consent according to the UPenn Institutional Review Board. Venous blood was sampled to assess [18F]NOS plasma to whole blood activity ratios and parent fractions. Blood pool RA, PA, and LV time activity curves (TACs) were measured using 1 cm3 peak VOIs while lung uptake TACs were extracted from all lung tissue in PET field of view. Kinetic analyses using a 2-compartment model with a fixed 0.15 lung blood volume fraction were performed to estimate the total volume of distribution (VT), non-displaceable binding potential (BPND), and transport into first compartment (K1) (Pmod v3.7). Blood input function (IF) area under the curve (AUC) and peak to tail ratio were assessed excluding one abbreviated scan. Estimate comparisons using each pool were based on χ2, R2, VT, BPND, and K1 and assessed using repeated measures ANOVA with post-hoc test of means (IBM SPSS 25).
Results: Blood IFs AUC varied significantly between each pool (mean ± SD: RA 7370±480, PA 6655±502, LV 7000±522 (g/mL)s, n=10, p < 0.02). Peak to tail ratios were similar between RA and PA p=0.09, while RA and PA were each significantly different from LV at p < 0.001 (RA 43.8±14.8, PA 35.4±6.9, LV 16.9±4.8, n=10). The PA IF had the lowest 2-compartment modeling χ2 (p=0.01) while RA and LV had similar levels of error (p=0.11). The PA IF also had the highest R2 at 0.99±0.01 versus RA and LV, both at 0.97±0.02, p < 0.02. VT was significantly greater using PA than RA (p=0.04), but similar to LV (p >equal 0.05) (RA 1.20±0.87, PA 1.41±1.14, LV 1.34±0.92 mL/cm3). BPND was lowest using PA (p < 0.02), and similar between RA and LV (p=0.09), (RA 2.11±1.17, PA 1.21±0.30, LV 1.54±0.51). K1 varied by blood pool selection (RA 1.14±0.72, PA 1.97±0.63, LV 26.34±20.42 ml/cm3/min, p < 0.003).
Conclusions: Pulmonary artery blood pool VOIs are better than either right atrium or left ventricle blood pools for extraction of an image-derived blood IF for quantitating pulmonary [18F]NOS uptake based on having both the lowest χ2 and highest R2values (p < 0.02). The importance of blood pool selection was confirmed by the observed significant differences in many model estimates based on choice of blood pool. Research Support: NCATS/TBIC: UL1TR001878 (RW, JD), R21HL144673 (RW), K01DA040023 (RD), K23DA038726-02 (JD) and P30DA046345 (RM).
