Page 214 - 18F-FDG PET as biomarker in aggressive lymphoma; technical and clinical validation

Chapter 8
Materials and methods
Database
This IPD meta-analysis included 1692 patients with de novo DLBCL from the PETRA database with I-PET scans after 1 to 4 cycles of chemotherapy for those who were treated with R-CHOP. This database was established by the PETRA consortium and contains patient-level data for 2539 patients with non-Hodgkin lymphoma who were enrolled in the Bologna [7], HOVON-84 [8], IAEA [9], GSTT15 [10], NCRI [11], Nordic-US Intergroup [12], PETAL [13], and SAKK 38/07 [14] studies.
The following are eligibility criteria for the PETRA database: adult patients age 18 years or older who had first-line treatment for non-Hodgkin lymphoma and had received an 18F-FDG I-PET scan. The trial had to have a prospective design or retrospective design with consecutive patients, at least 40 patients with progression-free survival (PFS) and overall survival (OS) data, and a full-ring PET system. Individual study protocols were approved by local institutional review boards, and written informed consent was provided by all participants in each study. After signing a data sharing agreement, data were made available to PETRA. Data remained the property of contributing investigators. The use of all data within the PETRA imaging database has been approved by the institutional review board of the Vrije Universiteit Medical Center ( JR/20140414).
Data collection, harmonization, and re-analysis
Each study was checked for missing data and for data that were consistent with those in published reports. Trial investigators were contacted about discrepancies or missing information. Patient numbers were recoded to PETRA identification numbers that consisted of study-specific and patient-specific parts. Patient data from original studies were merged into an online database and harmonized using the PETRA coding for all studies. All PET images were given a new pseudonym and were uploaded to an online database [15].
Survival rates were recalculated by using the date of baseline PET at the start of follow-up. If the date of baseline PET was not available, we used the baseline CT date or date of diagnosis (supplemental Data). Missing variables were completed, whenever possible, by reviewing scans. Scans were reviewed to provide Deauville
212