Using Multi-Peak Reconstructions for Dosimetry of Fast and Quantitative Lutetium-177 SPECT/CT 📝

Author: Julia Brosch-Lenz, Munir Ghesani, Francesc Massanes, Michael Morris, Babak Saboury, Eliot Siegel, Alexander Hans Vija 👨‍🔬

Affiliation: Institute of Nuclear Medicine, Siemens Medical Solutions USA Inc., Molecular Imaging 🌍

Abstract:

Purpose: Quantitative imaging of the radiopharmaceutical distribution is crucial for treatment evaluation and dosimetry. However, as patient numbers for Lutetium-177-(177Lu)-labelled therapies continue to grow, imaging duration becomes a significant challenge, and concerns about quantitative reproducibility increase. This study explores a “triple”-peak 177Lu-SPECT reconstruction that include the 54-65 keV X-ray peaks and the 113kev photopeak in addition to the 208keV photopeak to increase image counts.

Methods: Four timepoint 177Lu-PSMA-617 SPECT images were obtained for five patients at a Siemens Symbia Pro.specta™ X3 SPECT/CT with 5/8” crystal and MELP collimator between day 0 and 8 post-therapy. The data was acquired in three bed positions (head-thigh) over 360° with 60 projections per detector with 5s/projection. Besides xSPECT QuantTM reconstruction using the 208keV photopeak (“SinglePeak”), a quantitative prototype reconstruction used a combination of 208keV, 113keV, and 57keV windows (“TriplePeak”) with adapted energy window-based scatter correction and custom calibration. Count statistics, quantitation, and noise characteristics as assessed by coefficient of variation (CoV) were evaluated between SinglePeak and TriplePeak reconstructions. Time-activity-curves were fitted in MIRDfit and absorbed doses were calculated using MIRDcalc.

Results: The scatter-corrected counts ratio in the total field-of-view between TriplePeak and SinglePeak was 2.2. The CoVs of SinglePeak vs. TriplePeak for kidneys, liver, spleen, and lesions across patients and timepoints were 40±21% vs. 39±21%, 54±32% vs. 53±32%, 64±49% vs. 62±49%, and 25±5% vs. 26±5%. Percentage differences in activity between reconstructions were 11±5%, 10±1%, 10±6%, and 8±7% for kidneys, liver, spleen, and lesions. Percentage differences in absorbed dose were 8±4%, 10±1%, 9±2%, and 7±4%.

Conclusion: This study demonstrated an increase in counts for TriplePeak compared to SinglePeak reconstruction that warrants short 177Lu-SPECT acquisition times with improved signal-to-noise level. Small impact on image quantification and dosimetry was found. Next steps will focus on optimization of reconstruction methodology and scatter correction including simulations and more patients.

Back to List