Innovative Scan Protocols With Combined Long Axial FOV PET and Spectral CT for Improved Quantification in Oncology

University Of Pennsylvania, Philadelphia PA

Investigators

Abstract

Innovative scan protocols with combined long axial FOV PET and spectral CT for improved quantification in oncology: The goal of this project is to i) develop methods and protocols to obtain high-quality dynamic information from both PET and CT modalities and, ii) use dynamic Spectral CT for PET kinetic modeling to reduce overall scan time to match existing clinical PET protocols (~20 minutes). This technological advancement will enhance diagnostic capabilities in oncology, with broad applications in other diseases, while also reducing the burden on patients with shorter imaging protocols. We will achieve these goals through 3 specific aims: (1) Develop spectral CT protocol and processing pipeline for IDIF and perfusion maps; (2) Develop advanced modeling to combine Spectral CT and PET data for kinetic modeling; (3) Evaluate PET-Spectral CT to quantify blood flow and glucose metabolism in breast cancer. In Aim 1 we will implement an ultra-high-pitch spectral CT protocol to generate dynamic iodine maps with extensive axial coverage, including the blood pool and targeted tissue/organ. Our deep learning (DL) based approach utilizes spectral CT to differentiate anatomical background from iodine enhancement, preserving image quality and minimizing radiation exposure. We aim to improve iodine contrast bolus delivery and timing, assessing concentration and injection/acquisition timing. In Aim 2 we will utilize a custom flow phantom for testing and validation of the CT methods in Aim 1 methods as well as pre- and post-reconstruction data corrections in PET. Initial assessments of the relationship between CT, PET, and CT-informed PET IDIFs and resultant kinetic perfusion parameters will be made with the flow phantom. In Aim 3 we will test the CT-dose reduction, bolus optimizations and PET data correction methods developed in Aims 1 and 2 in a translational porcine model (N=4) followed by a cohort of breast cancer patients (N=10). Patlak graphical analysis will also be implemented to compare the net influx rate of FDG when the IF is estimated with and without CT information. These metrics will be compared across modalities and estimation methods with the goal of developing a clinically translatable hybrid Spectral CT-PET protocol for dynamic imaging. Our overall outcomes will be solutions that facilitate the acquisition of high-quality dynamic information from both PET and CT, to be leveraged for PET kinetic modeling purposes, and used for clinically practical protocols to enhance the accuracy of oncological diagnostics.