Quantitative and Qualitative Improvement of Low-Count [⁶⁸Ga]Citrate and [⁹⁰Y]Microspheres PET Image Reconstructions...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiochemistry

Background:

• Positron emission tomography (PET) imaging often requires low photon statistics, particularly when reducing radiopharmaceutical radiation exposure.
• Scenarios like Y-90 microsphere radioembolization and Ga-68 citrate cancer imaging present challenges due to low positron emission fractions or extended uptake times.

Purpose of the Study:

• To investigate the efficacy of the block sequential regularized expectation maximization (BSREM) algorithm for improving image quality and quantitative accuracy in low-count PET imaging.
• To evaluate BSREM's performance in specific low-count scenarios, including Y-90 and Ga-68 imaging.

Main Methods:

• Utilized a time-of-flight (TOF) PET/MRI system to scan NEMA/IEC Body phantoms mimicking low-count conditions for Y-90 and Ga-68.
• Compared TOF-BSREM with conventional TOF ordered subset expectation maximization (TOF-OSEM), assessing contrast recovery, background variation, and signal-to-noise ratio.
• Conducted visual quality assessments on patient images ([68Ga]citrate, n=6) and preliminary assessments for [90Y]microspheres.

Main Results:

• TOF-BSREM demonstrated superior quantitative performance (contrast recovery, background variation, SNR) over TOF-OSEM in phantom studies.
• Patient studies with [68Ga]citrate showed a 17% improvement in visual analogue scale and a 1-point increase in Likert score with TOF-BSREM (beta=500) compared to TOF-OSEM.
• Optimal regularization parameter (beta) selection was crucial for balancing image noise and resolution, with beta=500 for [68Ga]citrate and beta=2000 for [90Y]microspheres.

Conclusions:

• The TOF-BSREM algorithm significantly enhances image quality and quantitative accuracy in low-count PET imaging scenarios.
• Algorithm performance is dependent on the radiopharmaceutical and specific counting statistics, necessitating adjustment of the beta parameter for optimal results.