Reproducibility and comparability of preclinical PET imaging data: A multicenter small-animal PET study

Abstract

The standardization of preclinical imaging is a key factor to ensure the reliability, reproducibility, validity, and translatability of preclinical data. Preclinical standardization has been slowly progressing in recent years and has mainly been performed within a single institution, whereas little has been done in regards to multicenter standardization between facilities. This study aimed to investigate the comparability among preclinical imaging facilities in terms of PET data acquisition and analysis. In the first step, basic PET scans were obtained in 4 different preclinical imaging facilities to compare their standard imaging protocol for 18F-FDG. In the second step, the influence of the personnel performing the experiments and the experimental equipment used in the experiment were compared. In the third step, the influence of the image analysis on the reproducibility and comparability of the acquired data was determined. Distinct differences in the uptake behavior of the 4 standard imaging protocols were determined for the investigated organs (brain, left ventricle, liver, and muscle) due to different animal handling procedures before and during the scans (e.g., fasting vs. nonfasting, glucose levels, temperature regulation vs. constant temperature warming). Significant differences in the uptake behavior in the brain were detected when the same imaging protocol was used but executed by different personnel and using different experimental animal handling equipment. An influence of the person analyzing the data was detected for most of the organs, when the volumes of interest were manually drawn by the investigators. Coregistration of the PET to an MR image and drawing the volume of interest based on anatomic information yielded reproducible results among investigators. It has been demonstrated that there is a huge demand for standardization among multiple institutions.