Abstract
100
Objectives: Semiquantitative determination of PET activity, based upon lesion to background activity, has been shown to be a clinically useful parameter, especially when comparing lesions between scans. Evaluation utilizing SUV values alone may be problematic due to inherent intrinsic limitations. We evaluated the inter and intra-patient reproducibility and variance of multiple reference tissues in 68Ga-PSMA-11 PET/CT scans in order to choose the most reliable for use in semiquantitative determination of PSMA-11 activity.
Methods: We performed a retrospective chart review of 100 total 68Ga-PSMA-11 PET/CTs completed at Indiana University Health from June 2017 through November 2019. Multiple PET/CT scans were completed on 29 patients (27 had 2 scans, 3 had 3 scans). Single PET/CT scans were obtained on 38 patients. Mean SUVs of background organs were acquired with standard regions of interest drawn in the parotid gland, liver, spleen, lumbar paraspinal musculature, and mediastinal blood pool at the aortic arch. SUV mean, standard deviation, and coefficients of variation were calculated for all patients and all organs.
Results: Average SUVs are highest in the parotid gland (13.09) and lowest in the lumbar paraspinal musculature (0.46). Interpatient variation is lowest in the blood pool with a coefficient of variation (CV) of 0.20 and highest in the spleen at 0.38. Intra-patient variation is lowest in the spleen and liver at 0.10 though similar to blood pool at 0.11. The lumbar paraspinal musculature and parotid gland show higher and similar intra-patient variation 0.13.
Conclusions: Based on our analysis, the mediastinal blood pool is the best reference tissue for semiquantitative determination of PSMA-11 activity between patients. In patients with multiple PSMA PET scans, the liver, spleen, and mediastinal blood pool may all be reasonably considered for semiquantitative PSMA-11 activity.
Table 1. 68Ga-PSMA-11 PET/CT SUV Data. SUV: Standardized uptake value; CV: coefficient of variation
