Quantitative implications of the updated EARL 2019 PET-CT performance standards

Area of Science:

• Nuclear Medicine
• Medical Imaging
• Radiomics

Background:

• The European Association of Nuclear Medicine Research Ltd (EARL) provides standardized specifications for quantitative PET-CT imaging.
• Updated EARL specifications (EARL2) have been released, necessitating an evaluation of their impact on quantitative reads.
• Maintaining consistency with current EARL standards (EARL1) is crucial for clinical practice during the transition to new specifications.

Purpose of the Study:

• To investigate the impact of the new EARL2 specifications on quantitative metrics derived from clinical PET-CT studies.
• To test a method for generating quantitative reads compliant with EARL1 standards using EARL2-reconstructed images.

Main Methods:

• Four PET-CT image datasets were created from 13 non-small cell lung cancer (NSCLC) and 17 lymphoma studies: one EARL1 compliant, one EARL2 reconstructed, and two EARL2 reconstructed with 6 mm (EARL2F6) and 7 mm (EARL2F7) Gaussian post-filtering.
• Quantitative metrics including metabolic tumor volume (MATV), standardized uptake values (SUVmax, SUVpeak, SUVmean), and total lesion glycolysis (TLG) were compared across the datasets.
• Tumor-to-background ratios (tumor-to-liver and tumor-to-blood pool) and three volume of interest (VOI) methods (MAX41, A50P, SUV4) were also assessed in 55 lesions.

Main Results:

• The EARL2 specifications, particularly with the MAX41 VOI method, led to a 22% decrease in MATV, unchanged TLG, and a 23-30% increase in SUV values compared to EARL1.
• The EARL2F7 dataset (7 mm Gaussian filter) demonstrated the best alignment with EARL1 quantitative metrics, showing no significant differences for most metrics (p>0.05).
• While SUV metrics were consistent across different VOI methods, MATV and TLG showed variations. No significant differences were observed between NSCLC and lymphoma cohorts.

Conclusions:

• Adoption of EARL2 standards can alter quantitative PET-CT metrics, potentially increasing SUVs and reducing MATV while slightly changing TLG compared to EARL1.
• Applying Gaussian filtering (specifically 7 mm in this study) to EARL2-reconstructed PET images is an effective method to achieve quantitative reads compliant with EARL1 standards.
• This approach allows for the utilization of updated EARL2 protocols while ensuring continuity and comparability with existing EARL1-based quantitative data in clinical research.