Influence on voxel-based dosimetry: noise effect on absorbed dose dosimetry at single time-point versus sequential...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiotherapy Physics

Background:

• Accurate absorbed dose mapping is crucial for effective radionuclide therapy.
• Statistical noise in single-photon emission computed tomography (SPECT) imaging can influence the accuracy of dose calculations.
• Understanding the propagation of image noise to dose maps is essential for optimizing SPECT-based dosimetry.

Purpose of the Study:

• To quantify how statistical fluctuations in SPECT images affect voxel-level absorbed dose calculations.
• To evaluate the impact of varying acquisition and processing parameters on absorbed dose map accuracy.
• To compare dosimetry outcomes from single-timepoint SPECT versus time-integrated activity analyses.

Main Methods:

• Utilized iodine-131 filled phantoms with varying SPECT/CT acquisition and reconstruction protocols.
• Employed STRATOS software for voxel-level absorbed dose distribution assessment.
• Analyzed absorbed dose values and coefficient of variation (COV) for single time-point and sequential SPECT data.

Main Results:

• Single-timepoint dosimetry showed minimal impact on mean absorbed dose from counts or reconstruction, but dose map uniformity correlated linearly with SPECT noise.
• Absorbed dose COV from sequential SPECT imaging (high and low counts) was slightly lower than that from the highest count single SPECT image.
• The influence of SPECT counts and reconstruction parameters on dosimetry was nearly linear for isolated images but less significant for sequential analyses.

Conclusions:

• SPECT image noise significantly impacts the uniformity of absorbed dose maps, particularly in single-timepoint dosimetry.
• Sequential SPECT imaging provides more robust absorbed dose estimates compared to single time-point analyses.
• Reconstruction parameters and SPECT counts have a less pronounced effect on dosimetry when utilizing time-integrated activity from sequential SPECT data.