A Unified Framework for Plasma Data Modeling in Dynamic Positron Emission Tomography Studies.

Full quantification of dynamic positron emission tomography (PET) data requires the knowledge of tracer concentration in the arterial plasma. However, its accurate measurement is challenging due to the presence of radiolabeled metabolites and measurement noise. Mathematical models are fitted to the plasma data for both radiometabolite correction and data denoising. However, the models used are generally not physiologically informed and not consistently applied across studies even when quantifying the kinetics of the same radiotracer, introducing methodological variability affecting the results interpretation. The aim of this study was to develop and validate a unified framework for the arterial data modeling to achieve an accurate and fully automated description of the plasma tracer kinetics. The proposed pipeline employs basis pursuit techniques for estimating both radiometabolites and parent concentration models from the raw plasma measurements, allowing the resulting algorithm to be both robust and flexible to the different quality of data available. The pipeline was tested on four PET tracers ([ Compared to the standard procedure, the pipeline provided similar fit of the parent fraction but yielded a better description of the total plasma radioactivity, which in turn allowed a more accurate fit of the tissue PET data. The new method showed superior fits compared to the standard pipeline, for both continuous and discrete arterial sampling protocol, yielding to better description of PET data. The proposed pipeline has the potential to standardize the blood data modeling in dynamic PET studies given its robustness, flexibility and easiness of use.