Improving accuracy of SUV estimates in paediatric oncology: Recommending against the use of body weight corrected SUV in [¹⁸F]FDG PET.

Abstract

Few studies have assessed body weight dependency of Standardised Uptake Value (SUV) formulations in paediatric patients. This study aims to compare different SUV formulations measured in reference tissues in paediatric patients and determine which correction method shows the least dependency on body weight. A single-centre, retrospective analysis of [18F]FDG PET/CT scans was performed. SUV measurements were taken from liver and blood pool using EARL1 reconstructions. SUV measurements were corrected for body weight (SUVBW), lean body mass (LBM) according to James (SUVLBMJames) and Janmahasatian (SUVLBMJanma), and body surface area (BSA) according to DuBois (SUVBSADuBois) and Haycock (SUVBSAHaycock). The coefficient of determination (r²) was used to assess the correlation between SUV and body weight. In total, 461 scans were analysed, including 185 (40%) from female patients. The median age of patients was 12 years (IQR 8-15.5 years). SUVBW exhibited the strongest correlation with body weight, with r² = 0.65 for the liver and r² = 0.50 for the blood pool. In contrast, SUVBSADuBois and SUVBSAHaycock had the weakest correlation, with r² = 0.09 for the liver and r² = 0.06 for the blood pool. SUVLBMJames and SUVLBMJanma had moderate correlations, with r² = 0.51 and r² = 0.44 for the liver and blood pool, respectively, and r² = 0.47 and r² = 0.42, respectively. In paediatric [18F]FDG PET/CT scans, SUVBW should be avoided due to elevating values in heavier patients. SUVBSA presents the least dependency on body weight and provides the most consistent assessments of metabolic activity.