Multicenter study of quantitative SPECT imaging: reproducibility of ^99mTc quantitation using a conjugated gradient minimization reconstruction algorithm

Abstract

Objective: This multicenter study aimed to determine the reproducibility of quantitative single-photon emission computed tomography (SPECT) images generated by a commercially available, ordered-subset conjugate gradient minimization (OSCGM) reconstruction engine. Methods: A common cylindrical phantom containing 100 kBq/ml technicium-99m (^99mTc) pertechnetate solution in a volume of 7L of was scanned under standard imaging conditions at six institutions and under local clinical protocols at each. Inter-institutional variation was evaluated with the coefficient of variation (CV) among institutions in the quantitative SPECT images. The dose calibrator accuracy was also investigated by measuring the same lot of commercially available ^99mTc vials. Results: The respective radioactive concentrations under standard and clinical conditions ranged from 95.71±0.60 (mean ± standard deviation) to 108.35±0.36 kBq/mL, and 96.78±0.64 to 108.49±0.11 kBq/mL, respectively. The inter-institutional variation in the radioactive concentration was 4.20%. The bias in the radioactive concentrations in SPECT images was associated with each institutional dose calibrator accuracy. Conclusion: We concluded that the reproducibility of the commercially available quantitative SPECT application using an OSCGM reconstruction engine was high, and comparable to that of positron emission tomography (PET), for comparatively large (~7L) homogeneous objects.