Comparison of scatter and partial volume correction techniques for quantitative SPECT imaging of ²²⁵Ac

The extreme low-count regime for clinical ²²⁵Ac-SPECT imaging poses a challenge to energy-window based scatter correction (EWSC) methods. Moreover, SPECT imaging suffers from partial volume effects (PVE), which can degrade quantification and lead to an underestimation of the absorbed dose estimations, especially in small structures such as lesions. The aim of this study was to investigate the impact of scatter correction and partial volume correction (PVC) techniques on post-therapeutic imaging of the three imageable photopeaks of ²²⁵Ac.

A phantom with three 3D-printed spheres (191, 100, 48 ml) was imaged to compare transmission-dependent scatter correction (TDSC) to EWSC (440, 218 keV)/no scatter correction (no SC) (78 keV), as well as the impact of iterative Yang (IY)- and Richardson-Lucy (RL)-based PVC techniques, in terms of contrast-to-noise ratios (CNR) and recovery coefficients (RC). These scatter correction and PVC methods were also compared for a patient cohort, with two SPECT/CTs acquired 24 and 48 h after [²²⁵Ac]Ac-PSMA-I&T therapy, to evaluate their impact on kidney and lesion dosimetry.

In the phantom study, TDSC outperformed EWSC/no SC across all energy windows in terms of CNR, and in terms of RC for 218 and 78 keV energy windows under clinically relevant conditions. Application of PVC techniques resulted in a clear increase in RC and CNR across all energy windows. In the patient study, RBE-weighted kidney absorbed doses increased on average across all kidneys by 9 ± 4%, 30 ± 29% and 35 ± 29% for 440, 218 and 78 keV energy windows, respectively, when TDSC was applied. For lesion dosimetry, TDSC resulted in an average increase across all lesions by 16 ± 8% (218 keV) and 31 ± 30% (78 keV), and a decrease by 4 ± 8% (440 keV). In the patient study, IY-based PVC increased kidney absorbed doses by 172 ± 54%, 157 ± 45% and 146 ± 47%, for 440, 218 and 78 keV energy windows, respectively. RL-based PVC increased lesion absorbed doses by 34 ± 6%, 29 ± 8%, and 23 ± 10%, for 440, 218 and 78 keV energy windows, respectively.

The phantom and patient studies demonstrated TDSC superiority over EWSC/no SC. PVC techniques substantially increased kidney (IY) and lesion (RL) absorbed doses, highlighting their value for post-reconstruction enhancement of ²²⁵Ac SPECT images.