Rapid Standardized CT-Based Method to Determine Lean Body Mass SUV for PET—A Significant Improvement Over Prediction Equations

In (18)F-fluorodeoxyglucose ((18)F-FDG) positron emission tomography (PET) studies, maximum standardized uptake value (SUV(max)) is the parameter commonly used to provide a measurement of the metabolic activity of a tumor. SUV normalized by body mass is affected by the proportions of body fat and lean tissue, which present high variability in patients with cancer. SUV corrected by lean body mass (LBM), denoted as SUL, is recommended to provide more accurate, consistent, and reproducible SUV results; however, LBM is frequently estimated rather than measured. Given the increasing importance of a quantitative PET parameter, especially when comparing PET studies over time to evaluate disease response clinically, and its use in oncological clinical trials, we set out to evaluate the commonly used equations originally derived by James (1976) and Janmahasatian et al. (2005) against computerized tomography (CT)-derived measures of LBM. METHODS: Whole-body (18)F-FDG PET images of 195 adult patients with cancer were analyzed retrospectively. Representative liver SUV(mean) was normalized by total body mass. SUL was calculated using a quantitative determination of LBM based on the CT component of the PET/CT study (LBM(CT)) and compared against the equation-estimated SUL. Bland and Altman plots were generated for SUV-SUL differences. RESULTS: This consecutive sample of patients undergoing usual care (men, n = 96; women, n = 99) varied in body mass (38–127 kg) and in Body Mass Index (BMI) (14.7–47.2 kg/m2). LBM(CT) weakly correlated with body mass (men, r(2) = 0.32; women, r(2) = 0.22), and thus SUV and SUL(CT) were also weakly correlated (men, r(2) = 0.24; women, r(2) = 0.11). Equations proved inadequate for the assessment of LBM. LBM estimated by James’ equation showed a mean bias (overestimation of LBM compared with LBM(CT)) in men (+6.13 kg; 95% CI 4.61–7.65) and in women (+6.32 kg; 95% CI 5.26–7.39). Janmahasatian’s equation provided similarly poor performance. CONCLUSIONS: CT-based LBM determinations incorporate the patient’s current body composition at the time of a PET/CT study, and the information garnered can provide care teams with information with which to more accurately determine FDG uptake values, allowing comparability over multiple scans and treatment courses and will provide a robust basis for the use of PET Response Criteria in Solid Tumors (PERCIST) in clinical trials.