Cohort-level Gaussian Process Modeling for Handling Missing Data in Preclinical Longitudinal PET Imaging Studies

This dataset contains raw longitudinal PET standardized uptake value (SUV) measurements from the described phantom and tumor studies, along with Python scripts for cohort-level Gaussian Process Regression (cGPR) and missing scan prediction analysis. The data and code support probabilistic reconstruction of cohort-level PET trajectories under irregular sampling and missing observations, enabling evaluation of uncertainty-aware longitudinal modeling in preclinical PET imaging.