Investigation of iterative deconvolution in ¹⁷⁷Lu-SPECT imaging for lesion dosimetry

Area of Science:

• Nuclear Medicine
• Radiotherapy Physics
• Medical Imaging

Background:

• Quantitative ¹⁷⁷Lu-SPECT enables patient-specific dosimetry for targeted radionuclide therapy.
• Limited spatial resolution in SPECT can lead to underestimation of absorbed doses (AD).
• Lucy-Richardson deconvolution (LRD) is a spill-over correction algorithm previously explored in PET imaging.

Purpose of the Study:

• To evaluate the applicability of the Lucy-Richardson deconvolution (LRD) algorithm for spill-over correction in ¹⁷⁷Lu-SPECT tumor dosimetry.
• To assess the impact of LRD on absorbed dose (AD) estimates in patients undergoing ¹⁷⁷Lu-based therapy.

Main Methods:

• The LRD algorithm's parameters (sigma and iterations) were optimized using the NEMA IEC Body Phantom.
• LRD was applied to reconstructed ¹⁷⁷Lu-SPECT images from 20 patients treated with [¹⁷⁷Lu]Lu-DOTA-TATE or [¹⁷⁷Lu]Lu-PSMA-I&T.
• Tumor absorbed doses (AD) from original reconstruction (OR), OR + LRD, and OR + recovery correction (RC) were compared.

Main Results:

• Optimized LRD parameters (sigma=6.0 mm, 4 iterations) improved sphere recovery coefficients and signal-to-noise ratios.
• Applying LRD to patient data resulted in an average increase of 22% in tumor AD compared to the original reconstruction.
• Phantom-based recovery correction showed a larger increase in tumor AD (57%) compared to LRD.

Conclusions:

• Lucy-Richardson deconvolution (LRD) increases absorbed dose (AD) estimates in ¹⁷⁷Lu-SPECT dosimetry, irrespective of radiopharmaceutical or lesion location.
• LRD presents a promising approach for image-based spill-over correction in ¹⁷⁷Lu-SPECT dosimetry.
• Further research is needed to compare LRD with other partial volume correction methods and their impact on AD uncertainty.