Partial volume correction strategies for quantitative FDG PET in oncology

Area of Science:

• Medical imaging
• Nuclear medicine
• Quantitative PET imaging

Background:

• Quantitative accuracy in PET is compromised by partial volume effects.
• Decreasing tumor volume leads to underestimation of standardized uptake value (SUV).
• Partial volume correction (PVC) methods aim to improve SUV accuracy.

Purpose of the Study:

• To assess the accuracy and precision of different PVC methods.
• To evaluate PVC performance in simulations, phantom experiments, and clinical studies.
• To determine the impact of PVC on SUV measurements and treatment response assessment.

Main Methods:

• Evaluated three PVC methods: reconstruction-based, iterative deconvolution, and spill-in/spill-out factors.
• Used mathematical phantoms and NEMA NU2 phantoms for simulations and experiments.
• Conducted clinical studies on non-small cell lung cancer and breast cancer patients.

Main Results:

• All PVC methods accurately recovered SUV within 10% for volumes >= 1 ml.
• Reconstruction-based PVC demonstrated superior precision compared to image-based methods.
• Clinical PVC increased SUV by 5-80%, with minor effects on test-retest variability and response assessment.

Conclusions:

• PVC enhances SUV accuracy in PET imaging.
• Reconstruction-based PVC is more precise but requires specialized software.
• Image-based PVC methods are practical alternatives with comparable clinical performance.