Project Summary The goals of this project are to improve the quality and quantitative accuracy of time-of-flight (TOF) PET imaging by inclusion of temporal information in fast 4D dynamic reconstruction approaches. This will include development of 4D reconstruction approaches that are optimized for a range of count statistics, which are encountered in dynamic PET imaging on TOF PET systems, in order to provide more robust estimates of time- varying regional tracer concentration and physiologic parameters derived from dynamic PET imaging. These new approaches will be systematically tested using dynamic cardiac and oncology patient datasets representative of research and clinical applications containing temporal information in PET. We will build upon the very efficient TOF reconstruction principles we have developed and investigated in our prior work, specifically the DIRECT approach (Direct Image Reconstruction for TOF). As part of this work, we will develop a new dynamic reconstruction software “toolbox” that will include both analytic and iterative 4D TOF algorithms with flexible, application-adjustable temporal models. Advances in PET technology as well as PET radiotracers have created a need for better methods for dynamic and low-count PET imaging for both clinical and research applications in cardiology, oncology, neurology, and other fields. This leads to a critical need for a 4D approach that efficiently and rapidly reconstructs dynamic data with short time frames and low counts with quantitative accuracy and high precision. Specific Aim 1 will develop and test novel TOF reconstruction tools for dynamic list-mode data using the DIRECT framework. This flexible framework will include both iterative and analytic algorithms with appropriate spatial and temporal models, reconstruction filters, and regularization procedures. Specific Aim 2 is directed towards optimization and assessment of the performance and quantitative accuracy of the proposed TOF reconstruction methods using temporally-varying datasets, both simulated and measured, over a range of relevant activity distributions, count statistics, and temporal variations encountered in research and clinical PET applications. Specific Aim 3 will apply the new tools to dynamic datasets for cardiology (82Rb and acetate PET) and oncology (extended time FDG data) to measure the improvement in precision and bias for quantitative parameter estimates obtained from these studies. The end-product of this research will be a new reconstruction method and 4D quantitative imaging toolbox that can be directly applied to clinical and research PET studies involving dynamic imaging data; these reconstruction techniques may also lead to improved quantitative accuracy for low-count data that result from companion diagnostic studies with novel tracers.