Biases in Multicenter Longitudinal PET Standardized Uptake Value Measurements

Area of Science:

• Nuclear medicine
• Medical imaging
• Radiochemistry

Background:

• Longitudinal positron-emission tomography/computed tomography (PET/CT) studies are crucial for assessing treatment response and conducting multicenter trials.
• Instrumentation variability, including human error, can introduce measurement biases in PET/CT studies.
• Accurate estimation of variability between patient scans is essential for reliable clinical trial outcomes.

Purpose of the Study:

• To investigate measurement biases in longitudinal PET/CT studies caused by instrumentation variability.
• To assess the impact of calibration methodology on standardized uptake value (SUV) errors over time.
• To evaluate the effectiveness of cross-calibration kits for dose calibrators and PET scanners.

Main Methods:

• Utilized National Institute of Standards and Technology-traceable calibration methodology for germanium-68/gallium-68 ((68)Ge/(68)Ga) sources.
• Constructed a cross-calibration kit using a (68)Ge/(68)Ga batch for both dose calibrators and PET scanners.
• Conducted repeat measurements across multiple PET imaging sites using six dose calibrators and six PET/CT scanners.

Main Results:

• Dose calibrator biases ranged from -50% to 9% and were generally stable, except when settings were altered.
• PET scanner biases in activity concentration ranged from -27% to 13% over a median of 174 days.
• Standardized uptake value (SUV) measurement errors ranged from -20% to 47%, with longitudinal biases varying significantly between scanners and over time (-11% to 59%).

Conclusions:

• Measurement biases in SUV are not stable over time and vary between PET scanner sites, impacting longitudinal study reliability.
• Attention to PET scanner calibration is critical for accurate longitudinal assessments.
• Cross-calibration kits for dose calibrators and PET scanners can aid in quality assurance and control for multicenter trials.