New standards for phantom image quality and SUV harmonization range for multicenter oncology PET studies

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Oncology

Background:

• 18F-FDG PET is crucial for cancer lesion detection, staging, characterization, and treatment response assessment.
• Standardizing image quality and SUV harmonization is essential for reliable multicenter oncology PET studies.
• Scanner technology advancements necessitate regular updates to image quality and SUV harmonization standards.

Purpose of the Study:

• To propose new image quality reference levels ensuring small lesion detectability in PET scans.
• To introduce a novel SUV harmonization range and an image noise criterion to minimize SUV variability across scanners.
• To update existing standards for multicenter oncology PET imaging.

Main Methods:

• Utilized the NEMA NU2 image quality phantom with 37 image sets from 23 recent PET/CT scanner models.
• Analyzed PET images (30-1800s acquisition durations) quantitatively and visually.
• Derived metrics including contrast-to-noise ratio (Q_{H,10 exttt{mm}}/N_{10 exttt{mm}}), background variability (CV_BG), and SUVmax/SUVpeak.

Main Results:

• Proposed image quality reference levels: Q_{H,10 exttt{mm}}/N_{10 exttt{mm}} ≥ 2.5 and CV_BG ≤ 14.1% for 10-mm sphere visual detection.
• Defined a new SUV harmonization range using the 10th-90th percentiles of SUV distributions.
• Recommended CV_BG ≤ 10% as an image noise criterion to reduce intra-scanner SUV variability.

Conclusions:

• New image quality reference levels were established to ensure small lesion detectability.
• A new SUV harmonization range and image noise criterion were proposed to minimize SUV variabilities, considering point-spread function reconstruction.
• Adherence to these proposed standards will enhance the reliability and standardization of multicenter oncology PET studies.