Detection of sub-centimeter lesions using digital TOF-PET/CT system combined with Bayesian penalized likelihood...

Area of Science:

• Medical Imaging
• Nuclear Medicine
• Radiology

Background:

• Advances in PET/CT technology aim to enhance image quality and small lesion detection.
• A novel digital Time-of-Flight PET/CT (TOF-PET/CT) scanner utilizes silicon photomultipliers (SiPM) and Bayesian penalized likelihood (BPL) reconstruction.

Purpose of the Study:

• To quantify the lesion detection capability of the digital TOF-PET/CT scanner with BPL reconstruction.
• To determine the optimal penalization factor (β) for accurate detection of small lesions.

Main Methods:

• PET/CT data acquired from a NEMA body phantom with spheres (4-13 mm) at varying target-to-background ratios (TBR).
• Images reconstructed using OSEM and BPL algorithms with different parameters (TOF, PSF, β values 50-400).
• Quantitative accuracy (Recovery Coefficient) and detectability (Detectability Index, visual assessment) evaluated.

Main Results:

• Digital TOF-PET/CT with BPL demonstrated improved Recovery Coefficient (RC) and Detectability Index (DI) for sub-centimeter spheres compared to OSEM.
• BPL reconstruction with a β value of 200 yielded maximal visual scores and accurate quantitative values for ~8-mm spheres.
• Optimal β value selection was crucial for accurate lesion detection, with detectability dependent on TBR.

Conclusions:

• BPL reconstruction on the digital TOF-PET/CT scanner significantly enhances quantitation and detectability of sub-centimeter lesions.
• An optimized β value of 200 is recommended for detecting sub-centimeter lesions, with performance influenced by TBR.
• This technology holds promise for improved detection of small lesions, potentially aiding in earlier disease diagnosis.