Preparation of 15O-labeled gases
15O-labeled gases were produced from an N2 gas containing 2.0% CO2 (for 15O-CO2) or 2.0% O2 (for 15O-CO and 15O-O2) using an 14N (d, n) 15O reaction. The cyclotron (CYPRIS HM-12S; Sumitomo Heavy Industries Ltd., Tokyo) was operated with an average beam current of 7 μA and a deuteron acceleration energy of 6 MeV. The gas concentration stabilizing system (CYPRIS G3-A; Sumitomo Heavy Industries Ltd., Tokyo) controlled the flow rates and the radioactivity concentrations of the 15O-labeled gases. 15O-labeled gases were supplied through a gas mixture device with pure oxygen to maintain the oxygen concentration at around 30%.
The animal experiment was approved by the Institutional Animal Care and Use Committee of the Osaka University Graduate School of Medicine (Approval number: 20-144-2). Normal male Sprague-Dawley (SD) rats (Japan SLC Inc., Hamamatsu, Japan) were housed under a 12-h light and dark cycle with free access to food and water. Rats (n = 9, 9 weeks, body weight = 310 ± 19 g) were anesthetized by the inhalation of 2% isoflurane, and a polyethylene tube was set into the femoral artery for arterial blood sampling. The anesthesia was switched to an intramuscular injection of xylazine (4.8 μg/g of body weight), butorphanol (1.6 μg/g of body weight), and midazolam (1.2 μg/g of body weight), and the flexible plastic tube was inserted into the trachea for the inhalation of 15O-labeled gases after the tracheotomy. The respirator (SN-480-7; Shinano Seisakusyo) was connected to the airway tube, and artificial ventilation (Additional file 1: Figure S1) was started (respiratory rate = 60 breaths per min, tidal volume = 3 mL).
The PET measurement was performed using a small-animal PET-CT scanner (Inveon MM; Siemens Medical Solutions, Knoxville, USA). The rats were placed in a supine position on the warming bed, and their rectal temperature, which was kept at 37.0 °C ± 0.5 °C, was monitored. Heart rate, systolic blood pressure (SBP), and diastolic blood pressure (DBP) were measured using a noninvasive system (BP-98A-L; Softron, Japan) by the tail-cuff method before and after each PET measurement. The PET scan was started at the same time as the start of the inhalation of each 15O-labeled gas. Using the steady-state method, 15O-labeled gases were ventilated continuously during the 16-min PET scanning period for the 15O-CO2 gas (200 MBq/min) and the 15O-O2 gas (400 MBq/min) studies (n = 9). In addition, 15O-CO gas (400 MBq/min) inhalation was also performed for 3 min, and the PET measurement was continued for up to 13 min in some animals (n = 6 of 9). Arterial blood samplings were performed at 13 and 16 min after the start of the PET scan in the 15O-CO2 and 15O-O2 studies and at 10 min after the start of the PET scan in the 15O-CO study. Arterial blood gas data (pH, PaCO2, PaO2, SaO2, hematocrit, and hemoglobin) were measured using a blood gas analyzer (i-STAT; FUSO Pharmaceutical Industries, Ltd., Japan) at 13 min after the start of the PET scan in the 15O-CO2 and 15O-O2 studies. The CT scan was performed for scatter and attenuation correction using a tube voltage of 80 kV and a tube current of 140 μA after PET acquisition. The weight and radioactivity count of the whole blood and blood plasma were measured using a NaI scintillation well counter (BeWell; Molecular Imaging Labo, Osaka, Japan).
Evaluation of the influence of 15O-labeled gas radioactivity in the lung on the brain
After overnight fasting, 18F-FDG (57.2 MBq) was administered to a normal SD rat (body weight = 319.43 g) via the tail vein under 2% isoflurane anesthesia to simulate accumulation in the brain during the steady state inhalation of 15O-CO2 gas. Based on previous data for 18F-FDG uptake in the brain under isoflurane anesthesia in our facility, the injected dose was calculated to be equivalent to the radioactivity in the brain during the inhalation of 15O-CO2 gas. Sixty minutes later, the rat was sacrificed with deep anesthesia, and a balloon phantom was inserted into the thoracic cavity after the removal of the lung. The balloon phantom was connected to the airway tube to ventilate 15O-labeled gases. Each 10-min PET scan was started 2 min after the start of the supply of 15O-O2 gas with different activity concentrations (0, 200, 400, and 600 MBq/min).
Reconstruction and data analysis
All the PET images were reconstructed using FBP, OSEM2D with 16 subsets and 4 iterations, and OSEM3D-MAP with 16 subsets, 2 iterations for OSEM3D, and 18 iterations for MAP with scatter and attenuation correction. The reconstruction parameters for iterative reconstruction were decided according to the default values provided by the manufacturer and have often been used in previous studies [7, 9, 11, 15, 16]. The requested resolution of the MAP reconstruction was set to 1.5 mm. The single scatter simulation algorithm was applied as the scatter correction [17, 18]. The image matrix and the voxel size were 128 × 128 × 159 and 0.776 × 0.776 × 0.796 mm, respectively. The energy and the timing window were 350–650 keV and 3.432 ns, respectively. Cross-calibration factors between the PET scanner and the dose calibrator were measured for each reconstruction method using a cylinder phantom of the same size as the NEMA standard rat phantom. The quantitative PET images (CBF, OEF, CMRO2, and CBV) were generated using the steady state inhalation method and in-house software according to the protocol described in a previous study . Time activity curves were obtained by setting spherical volumes of interest (VOIs; 10 mm in diameter) over the brain on dynamic PET images (1 min × 16 frames) reconstructed using OSEM2D to check the steady state for each 15O-labeled gas measurement. The PET images were aligned with the template of a T2-weighted magnetic resonance image using the rigid registration method and PMOD software, version 3.604 (PMOD Technologies). The VOI template (W. Schiffer) was automatically placed on the brain displayed in the PET images . The radioactivity counts for PET in the frontal cortex, somatosensory cortex, visual cortex, striatum, thalamus, pons, cerebellum, hippocampus, midbrain, and whole brain were obtained. The quantitative values (CBF, OEF, CMRO2, and CBV) in each brain region were calculated using each of the reconstruction methods (Additional file 1: Figure S2). For the rat with the lung balloon phantom in the thoracic cavity, VOIs were placed on the brain and the lung balloon phantom. The radioactivity counts were compared among the different activity concentrations of supplied 15O-labeled gases (0, 200, 400, and 600 MBq/min).
The radioactivity concentrations and quantitative values calculated using each reconstruction method were compared using a one-way repeated measures analysis of variance with Bonferroni-corrected pairwise comparisons. The quantitative values reconstructed using FBP in each brain region were compared with that in the cerebellum using a paired t test with Bonferroni’s correction. The cerebellum was used as a reference to understand the overall trend of the distribution in the brain according to previous studies [20, 21]. The radioactivity concentration in the whole brain was compared with that in the lung using a paired t test for the rat with a balloon phantom in the thoracic cavity. Statistical analyses were performed using Microsoft Excel 2013 and SPSS Statistics version 17.0 (SPSS Inc., Chicago). A P value of less than 0.05 was considered statistically significant.