We assessed the potential use of a home-made dual support system compatible with Siemens INVEON PET/CT scanner allowing the imaging of two rats simultaneously. The design of this system was based on existing systems, improving them both in terms of ergonomics and accessibility at a limited cost.
We performed scans in various configuration on phantoms and in vivo scan with rats, to evaluate a potential loss of image quality and a bias in quantitative measurement.
Scanning two rats during the acquisition reduces costs (because animals can be injected with tracer of the same production) and saves time. It potentially doubles the size of cohorts and significant increases statistical power.
To our knowledge, only one study (Cheng et al) has previously demonstrated the feasibility of PET simultaneous multiple scanning and positional reproducibility on rat preclinical model. As the dual rat support system of Cheng et al was designed to be used on a PET-only system (Siemens microPET P4), the animals could be positioned head-to-head since the imaging system allowed to access both side of the FOV. The PET/CT INVEON scanner used in this work has only one side accessible for tracer injection, since the CT scanner side is shielded for radioprotection purposes. For scanning two rats simultaneously, we designed a dual support system where the two rats are positioned in the same directions on bunk beds, allowing the access to both animals’ tail veins. As the positioning of each rat implies an offset from the scanner’s FOV center, one aim of our study was to evaluate its influence on the image quality.
The accuracy of the radioactivity measurements was first assessed in vitro on cylindrical phantoms (Falcon tubes) filled with 50 ml of 151.2 ± 28.9 kBq/g [18F]FDG solution. Experiments were performed with pair of phantoms filled with different radioactivity doses: one higher and one lower (activity ratio of 32 to 43% between phantoms), to quantify the impact on the detection performances despite the position in the scanner (upper, lower or dual). With calibrated activity in phantoms with size equivalent to rat size, we demonstrated that the measurement accuracy is similar when performing scan with one or two emitting objects in the FOV. The 6% difference between conditions is in the range of the uncertainties of the PET measures. We also showed that dispersion of the voxel values in a ROI where activity is supposed to be homogenous, is around 5% and, again, it is not increased when two object are in the FOV. More, we did not observe spillover of activity from one phantom to the other. Finally, the activity contrasts between regions are preserved even when the measurement is simultaneous.
The observations made on the accuracy of the measurements and their independence in the field of view for phantom conditions were favorable to the in vivo experimentation for which the ground truth cannot be so precisely known. In the second experiment, which was performed on living rats, the performance criterion was evaluated by the reproducibility parameters of the measurement. Indeed, the use of the bunk bed could be validated provided that the test-retest reproducibility parameters were the same with one or with two individuals in the FOV.
The results showed that the in vivo reproducibility was excellent with a bias of less than 10%, a coefficient of variation of less than 5%, and, above all, an ICC greater than 0.95, which is exceptional. This latter result means that the inter-individual variability is clearly greater than the intra-individual variability in the measurements, which guarantees that the slight variations between test and retest can be highlighted whether they are systematically reproduced between two groups or between two experimental conditions practiced on individuals from the same group. Knowing these, the test-retest reproducibility performances in SOLO scan conditions, we were able to verify that they were identical, or even better, in DUO conditions, and that it was also as good by comparing scans performed in SOLO with scans performed in DUO. In conclusion, we showed that there is no loss of quality by carrying out a scan with 2 rats simultaneously, and moreover, that the data acquired with one rat in the FOV could be compared to the data acquired when two rats are in the FOV.
Although we took the precaution of performing these measurements at a relatively high radioactivity level, a limitation of our study is that we did not demonstrate whether these results are still valid in any range of radioactivity. In particular, we cannot exclude a saturation effect of the PET detector at high radioactivity.,. To overcome this risk, it is therefore advisable to evaluate the doses injected in each rat and to check on the performance curves of the machine that twice this dose does not reach the peak NEC that can be seen on the curve's scanner performance (Qinan Bao et al). The NEC peak of the INVEON camera being reached for a phantom at approximately 100 MBq, the maximum dose to be injected into an animal would be 50 MBq, which corresponds to a weight dose of 160 kBq/g for an average-sized rat 300g. This weight dose is quite exceptional and probably never practiced. In our study, we injected a weight dose of approximately 35 kBq/g, which provided an image of quite sufficient quality.
Another risk is the saturation of the detector during dynamic acquistions at the first pass of the tracer at a high concentration, concentrated in the arterio-venous space,. This risk also exists for an acquisition with a single animal, but it is effectively doubled when two animals are in the FOV. A proposed solution is to slightly shift the injection time (~ 10 sec) of the two rats so that the peaks of activity in the field of view do not overlap temporally.
Finally, if the technical solution we propose has been applied to a PET/CT Siemens INVEON system, it can be transposed to any other small animal system.