Effect of reconstruction settings and therapy on 18F-FDG PET/CT blood pool and liver SUV in patients with diffuse large B-cell lymphoma

608Objectives: The PET Re-Analysis (PETRA) project is an international consortium building a shared database consisting of individual patient data and 18F-FDG PET/CT scans. PETRA aims to re-analyze interim PET/CT scans in diffuse large B-cell lymphoma (DLBCL) to validate interim PET/CT as a biomarker of response in first-line therapy. The EANM guidelines for tumor imaging 2.0 suggest that liver and blood pool standardized uptake values (SUVs) may be used for quality control to verify e.g. correct FDG administration or technically correct scanner performance. Moreover, the ICML response criteria compare tumor uptake to liver and blood pool uptake. For validation of interim PET/CT as an imaging biomarker, the quality control of PET/CT scans is essential to assure harmonized quality of pooled PETRA data. To this end, we investigated whether reconstruction settings, therapy and use of growth factors influenced liver and blood pool SUVs in DLBCL patients.Methods: PET/CT scans of DLBCL patients were collected from the HOVON 84 study, a multicenter randomized controlled trial. Patients were treated with R-CHOP14 therapy and were randomized to receive rituximab intensification in the first four cycles or not. Granulocyte-colony stimulating factor (G-CSF) was given to oppose neutropenic side effects. Therapy was not adapted based on interim PET/CT results. Liver and mediastinal blood pool (MBP) uptakes were analyzed in 123 baseline PET/CT scans to determine the effect of reconstruction settings. Based on the EANM Research Ltd (EARL) criteria, we created 3 groups; EARL compliant, non-EARL compliant and unknown. Liver and MBP SUVs were analyzed at baseline, interim (after 4 cycles of chemotherapy) and end-of treatment for 34 DLBCL patients to determine the influence of therapy. Interim scan timing was defined as time between the start of the fourth R-CHOP14 cycle and interim PET/CT. Differences in SUV variance between reconstruction settings were analyzed using Levene’s test. Baseline, interim and end-of-treatment SUVs were compared using paired samples t-test (p <0.05 was considered to be significant). Correlation between SUV and interim scan timing was analyzed using linear regression.Results: No significant differences in variances between the various SUVs were seen in EARL compliant baseline scans compared to unknown or non-EARL compliant scans. As is shown in Figure 1, a significant increase was seen in liver SUVmean (+0.414 p =