Uncertainties in calibration for quantitative 90Y SPECT/CT

Objectives Quantitative 90Y SPECT/CT can be used to estimate absorbed doses to tumors and liver following 90Y microsphere therapies (90YMT). Uncertainties in count-to-activity calibration of 90Y SPECT directly affects SPECT activity quantification and absorbed dose estimates. Correction of 90Y bremsstrahlung scatter and attenuation are challenging. The objective of this study is to determine the uncertainties in the calibration factor (CF) for quantitative 90Y SPECT/CT studies.

Methods 90Y activity in patients is largely located in the liver and confined to the SPECT field of view (FOV). 30 clinical studies were selected with 90Y activities in FOV known within 10%; the lung shunt fraction (from MAA scan) was <10% and net administered activity (by measuring exposure rate of prescribed activity and post-therapy waste) was >98%. Error in 90Y activity assay using dose calibrator is <5%. Post-90YMT Planar (ANT & POST) and SPECT/CT images were acquired using Siemens SymbiaT with MELP collimator in 4 energy windows (EW) A:70-90, B:90-125, C:125-161, D:314-415 keV. SPECT was reconstructed from projections in EW B using OSEM with CT attenuation correction, resolution recovery and in-house EW-based scatter correction. Self-CF was defined as net activity in FOV/ total counts. For each patient, self-CF was calculated for Planar counts, SPECT projection counts, and SPECT reconstructed counts. Correlations and least-squares fit of net activity with total counts were performed. Mean and maximum deviation of self-CF from best-fit CF were calculated.

Results Total counts in Planar, SPECT projections, reconstructed SPECT images for patient scans were proportional to the net activity in all EW; R²>0.99, p<0.01. The mean & maximum deviation of self-CF for each patient from the least-squares fit CF for Planar, SPECT projections, reconstructed SPECT images were 10% & 23%, 9% & 27%, 4% & 10%.

Conclusions 90Y SPECT/CT reconstruction in patient scans adequately correct for attenuation and scatter as evident by decrease in variability of self-CF. A single CF can be used to convert the total SPECT/CT counts to activity in the FOV with errors <10%.

Research Support NIH/NCI R01 CA13898