Multinuclear MRI and PET system
All MR experiments were carried out on a home-integrated small animal dedicated 9.4 T MRI scanner. This system operates on clinical software from Siemens SYNGO [29]. Multinuclear MR acquisitions of 1H (399.7 MHz, corresponding Larmor frequency of 1H at 9.4T), 7Li (155.3 MHz), 13C (100.5 MHz), 17O (54.2 MHz), 19F (376.0 MHz), 23Na (105.8 MHz) and 31P (161.9 MHz) were achieved using the exchangeable multinuclear coil set (Figure 1a), including a correspondingly tuned RF coil, a transmit/receive (T/R) switch and a preamplifier. The performance of the coil only on a test phantom for the individual nucleus was reported elsewhere [28]. The anaesthetised animal was positioned from the front side of the magnet while these coil sets can be accessed and swapped from the opposite side. This ensures that the animal is not disturbed and remains in the same position during coil change. All PET measurements were performed on a commercial preclinical INVEON PET imager (Siemens Healthineers, Erlangen, Germany). An 18F-FET-PET tracer was used which was synthesised in-house with a specific radioactivity (< 200 GBq/μmol) [30].
MR-PET common bed and animal transport system
A 3D printer (Fortus 400mc, Stratasys, Minnesota, USA) using biocompatible polycarbonate was utilised to design and construct the dedicated bed for the rats and the common MR-PET bed adapter. This purpose-built bed contains: a freely adjustable bite bar, a nose corn, gas in/out tubes for anaesthesia, a respiratory gating pad and a temperature-monitoring unit (SA Instruments, USA). Figure 1b shows the Perspex box (pre-filled with Isoflurane) to transport the animal bed along with the anaesthetised animal from the MR unit to the PET site (< 3 minutes) without disturbance.
Animal handling and tumour inoculation
All animal measurements were approved by the Animal Protection Committee of the local government (LANUV, North-Rhine-Westphalia, Germany) according to the German Animal Welfare Act and the European Community Council directives regarding the protection of animals used for experimental and scientific purposes (2010/63/EU). All rats weighed between 250-310g and were handled under standard housing conditions (12/12 hour light-dark cycle, approximately 22°C room temperature and 54% humidity, free access to water and food) in the animal facility of Forschungszentrum Juelich. Four Fischer-344 rats purchased from Charles River Laboratories (Germany) were included in this study.
Rodent 9L gliosarcoma cells were implanted into the right striatum of two male rats, as previously described [31]. 9L cells invade contiguous brain tissue and develop neovascularisation, which are a hallmark of aggressive gliomas, but tumour margins are still well delineated. The other two healthy female rats were also used as a control. Tumours were allowed to grow for 11 to 12 days followed by MR and PET measurements. Rats were anaesthetised with 2 to 5% isoflurane in oxygen and placed into the animal scanner, in which they were measured under continuous isoflurane anaesthesia. During both MR and PET measurements, the temperature and respiratory rate of the rats were maintained at around 37.8°C and between 48 to 55 beats per minute, respectively. At the time of transportation, isoflurane pre-filled in the box was used to maintain anaesthesia. For the PET scans, a venous catheter was inserted into a tail vein of the rat to deliver the 18F-FET tracer.
Data acquisition scheme and analysis
1H MRI, 23Na MRI and 31P MRS
As described in Figure 1b, standard adjustment, such as shimming and RF power calibration was performed using the proton coil prior to multinuclear experiments and the optimised shim values were employed for all other nuclei during the MR data acquisition. Experiments were carried out using standard MR sequences cloned from the clinical platform from Siemens since the scanner uses the same hardware apart from coils and magnets. The high-resolution structure T2-weighted 1H images were obtained using a turbo spin echo (TSE) sequence [32] in all axes.
23Na images were collected using either a 3D FLASH [33] or a simultaneous single-quantum and triple-quantum-filtered MRI of 23Na (SISTINA) sequence. Using the advanced SISTINA sequence with multiple phase cycling [10, 34], ultrashort TE (UTE), single-quantum and triple-quantum-filtered 23Na images can be generated in a single acquisition, as shown in Figure 3. This allows for the concurrent acquisition of images resulted by the presence of restricted/non-restricted sodium and sodium density-weighted images, which is not possible with the conventional method.
A localised image-selected in vivo spectroscopy (ISIS) sequence [35] was used to acquire 31P spectra. 31P MR spectra were processed with the use of the advanced method for accurate, robust and efficient spectral fitting (AMARES) provided by jMRUI [36], and the 31P metabolites were analysed, such as a PCr/Pi ratio as an example. The detailed MR scan parameters for each experiment are summarised in Table 1. The acquired MR images were then analysed and co-registered using in-house MATLAB (Mathworks, Natick, Na, USA) and ITK-SNAP software [37].
18F-FET-PET
A transmission scan (~10 minutes) with a retractable 57Co point source for attenuation and scattering corrections was initially carried out prior to PET data acquisition. The acquisition was then performed in the 3D list mode (~65 minutes) starting with an injection of 40 ± 3 (mean ± range) MBq 18F-FET in saline into the tail vein (bolus injection of 0.5 ml in 1 minute). Emission data were framed into a dynamic sequence of 6 × 10 second, 5 × 60 second, 5 × 3 minute, 10 × 4 minute frames. Filtered back-projection with ramp filter (cut-off = 0.5) was employed for the reconstruction of 159 slices, with a voxel size of 0.8 × 0.8 × 0.8 mm3. All images were corrected for random coincidences, scatter and attenuation. 18F-FET uptakes in the tumour and a normal tissue were expressed as standardised uptake values by dividing the radioactivity (kBq/mL) in the tumour and a tissue, respectively, by the radioactivity injected per gram of body weight. Tumour-to-brain ratios (TBRs) were also calculated by dividing the mean value of the tumour by the mean value of a normal brain tissue. Summed PET images (18-60 minutes post injection) were used for co-registration of MR and PET images. The acquired PET data were analysed using PMOD (PMOD Technologies LLC., Switzerland).