Limited performance of calculated attenuation correction for brain PET of cat

Abstract

1722

Objectives: Since post-injection transmission (TX) PET scans are not permitted or provided to microPET scanner users at present, additional time for performing TX scan and awaiting FDG uptake is required for attenuation and scatter corrections. Increasing probability of subject movement and possible biological effect of long term anesthesia would be the problem in additional TX scan. Therefore, calculated attenuation correction (CAC) in animal PET studies may be useful but should be validated. In this study, we investigated the validity of CAC for cat brain PET data. Methods: FDG brain PET scans of 9 normal cats were acquired using a microPET Focus 120 scanner. For each animal, TX scan using ⁶⁸Ge source was performed for 30 min before the tracer injection. After 30 min uptake time of FDG, EM scan was performed for 30 min. Both the EM and TX PET data were reconstructed using Maximum a posteriori (MAP) algorithm. For the CAC, uniform attenuation map obtained by assigning the linear attenuation coefficient for soft tissue to whole head region was forward projected into the 3D sinogram space. Voxel wise paired t-test using SPM was performed to identify the difference between measured attenuation correction (MAC) and CAC EM data. Results: In SPM analysis, PET counts in CAC data was significantly higher than in MAC data dominantly in thalamus and cerebellum which were close proximities to esophagus and air cavity (p < 0.001). This would be due to the over-correction in CAC since attenuation coefficient of the soft tissue was assigned to esophagus and air cavity. Lower uptake in CAC PET compared to MAC PET was not detected. Conclusions: CAC for cat brain PET has advantage over MAC in terms of scan time. However, CAC yielded overestimated glucose metabolism on bottom of the brain near the air cavity.