Abstract
BACKGROUND: Positron Emission Tomography (PET) imaging using 3’-deoxy-3’-[F-18] fluorothymidine (FLT) is a valuable non-invasive marker of tumor proliferation. However, variability in imaging protocols and quantitative analysis techniques hampers comparability across preclinical chemotherapy response studies. METHODS: To address this issue in preclinical studies, we evaluated two key areas: (1) imaging standardization (2) semi-quantitative analyses, including Standard Uptake Value (SUV) SUV(max) and SUV(mean), and percent injected (%ID)/g. Using standardized methods of animal handling, image acquisition, image reconstruction and analyses, we assessed multiple FLT uptake metrics in three responsive Patient-Derived Xenograft (PDX) models treated with temozolomide/berzosertib or cisplatin/berzosertib. Metrics included: %ID, SUVbw(max), SUVbw(mean) with a Region of Interest (ROI) threshold of 50% of the maximum value [SUVbw(mean-50%)], and the ratios tumor-to-tissue and tumor-to-liver. Statistical comparisons were made using Brown-Forsythe and Kruskal-Wallis ANOVA and unpaired two tailed t-test to assess cohort differences. RESULTS: Across models and timepoints, no statistical difference among analysis techniques were observed except for one outlier at day 3 in one vehicle cohort, likely due to variability in the normalization parameter. SUVbw(max) and SUVbw(mean-50%) showed the strongest ability to differentiate between vehicle and treated cohorts. CONCLUSION: Semi-quantitative FLT PET metrics can be harmonized across preclinical studies, with caveats in understanding the normalization parameters, with SUV-based metrics preferred and that larger numbers of animals per cohort will be required to power (≥0.8) robust conclusions. These findings support FLT PET standardization in translational imaging. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12967-025-07475-2.