Chi Liu, Ph.D. - US grants

DESCRIPTION (provided by applicant): Cancer chemotherapy often induces cardiotoxicity, which can have a significant impact on the overall prognosis and survival of cancer patients. Current guidelines to screen for cancer therapy-related cardiotoxicity are primarily based on serial assessment of left ventricular ejection fraction (EF), which is not a sensitive index of cardiotoxicity and may only decline at a time point that is too late to reerse the process. In addition to cardiac function, the microvasculature plays a critical rol in cardiotoxicity. There is a close bidirectional coupling of regional myocardial mechanics and microvascular perfusion. Many of the newer chemotherapy agents can directly cause microvascular injury, which may precede any EF drop. Due to an increasing aging population and rapid introduction of new therapy agents, more patients and cancer survivors are expected to suffer from cardiotoxicity. Therefore, there is an urgent need to develop novel non-invasive imaging techniques that might allow early detection of microvascular injury of patients with cardiotoxicity prior to a drop in EF. With this urgent clinical need, we propose to quantify Intramyocardial blood volume (IMBV) as a novel measurement of microcirculation function. 99mTc-labeled red blood cell (RBC) is a clinically available blood pool tracer for EF measurement and RBC imaging using Single Photon Emission Computed Tomography (SPECT) is a natural approach to estimate IMBV as the tracer stays in the intravascular circulation. However, accurate quantification of IMBV using SPECT is challenging, because 99mTc-RBC has ~5-6 fold higher activity in the blood pool than in myocardium, the spill-over counts from blood pool to the myocardium mainly due to poor resolution and respiratory/cardiac motion can cause substantial IMBV overestimation. We have developed various novel quantitative low-dose SPECT/CT methods including CT-based partial volume correction and motion corrections, and have demonstrated the feasibility of quantifying IMBV using SPECT/CT in large animal studies. We hypothesize that accurate measurement of IMBV can provide an early index of disruption of the microcirculation and vascular reserve and improve detection of cancer therapy induced cardiotoxicity. In this proposal, we will optimize, validate, and translate this low-dose (<2 mSv) quantitative SPECT/CT imaging approach into large animal and human studies. We will pursue the study through four Specific Aims. In Aim 1, we will optimize the low-dose SPECT/CT imaging approaches. In Aim 2, we will optimize the low-dose contrast CT data acquisition protocols. In Aim 3, we will quantify and validate the serial changes of IMBV in an established large animal model. In Aim 4, we will establish the feasibility of this SPECT/CT imaging approach in patient studies. This project is a stepping-stone to translate this imaging method to large clinical trials and clinical practice.

Project Abstract PET plays an important role in cancer management. However, image blurring and mismatched attenuation correction due to respiratory motion can substantially degrade detection efficacy and quantification accuracy for tumors located in the lung and abdomen. Existing motion correction methods might provide satisfactory results for patients with regular breathing patterns, which account for about 60% of patients. However, for the remaining 40% of patients with irregular breathing patterns, these methods neglect the major effects of intra-gate motion due to inter-cycle and intra-cycle motion variations. In addition, as dose reduction in PET imaging has become increasingly important, existing motion correction methods typically amplify image noise and degrade their performances on low-count data. Another important challenge is the mismatch between CT and PET that limits phase-matched attenuation correction for every gated PET image using a single helical CT. Therefore, to achieve accurate quantification for evaluation of response to cancer therapy and reliable detection of tumors using low-dose PET protocols, particularly for patients with breathing pattern changes including variable motion amplitude, baseline variation, and amplitude variation, it is critical to develop personalized motion correction strategies optimized for individual patient's breathing patterns and the imaging task to eliminate intra-gate motion and mismatched attenuation correction for low- dose PET. Extending our existing collaboration, Yale and Siemens form an ideal team to optimize a comprehensive solution to correct for breathing pattern variability with intrinsically phase-matched attenuation correction for both regular and irregular breathers in the first two Aims. We will then develop and translate a personalized strategy to automatically identify the most time-effective motion correction approach for each individual patient, considering task and breathing pattern. We will optimize our personalized motion correction methods and strategy particularly for low-count PET data, aiming to reduce radiation dose to 25%-50% of the dose in current PET protocols. The outcome of this research will be a comprehensive motion correction package including four correction approaches and a personalized strategy that is automatically optimized for each individual patient. This development will be ready to translate to commercial PET/CT scanners and clinical end-users. As existing motion correction methods only apply to ~60% regular breathers, but have substantial limitation for the remaining ~40% irregular breathers, our proposed development can provide a unified motion correction framework for all patients with both regular and irregular breathing. This fast translation with industrial partners can lead to a significant and timely clinical impact for cancer management.

Project summary Quantitative PET has become increasingly important in clinical management and research, in particular for predicting and assessing response to therapy for cancer patients. Current PET protocols involve injection of PET tracers that typically result in ~6-7 mSv radiation dose to patients. For patients who require multiple repeated PET scans to monitor the response to therapy, and for patients who need PET scans with two or more tracers (e.g., FDG + FLT) to optimally predict response to therapy, it is critical to reduce the radiation dose from the PET tracer injection, while still maintaining the quantitative accuracy and image quality for cancer management. When reducing injection dose, the PET images will have higher noise due to fewer detected counts, which will subsequently introduce errors in quantitative measurements. For moving organs and tumors such as those in the lung and abdomen, respiratory motion can substantially degrade quantitative accuracy, so motion correction is required. Conventional motion correction uses a gating strategy that rebins the PET data, resulting in substantially higher noise in each gate. More advanced methods incorporate motion vector estimation in the image domain for post-registration or motion compensated image reconstruction using all detected events without increasing noise. The motion vectors need to be derived from gated PET, which are even noisier when using a reduced tracer injection in low-dose studies, imposing substantial challenges for accurate and reliable voxel-by-voxel motion vector estimation. In dynamic PET studies with clinical cardiac tracers and other novel oncology and neurology tracers, quantification is even more challenging for low-dose PET as each dynamic frame only contains a small fraction of detected events so the high image noise will affect the determination of image-derived input functions and can lead to bias and high noise in parametric images. In this project, to reduce image noise and maintain quantitative accuracy in PET, we propose to develop, optimize, and evaluate multiple innovative imaging methods for low-dose PET data to achieve comparable quantitative accuracy as full-dose PET. While the imaging developments are generally applicable to all PET tracers in oncology, neurology, and cardiology, since cancer is the primary clinical application of PET, we will focus our investigation and optimization in this project on three lung cancer imaging tracers at different clinical adoption stages as examples: 1) 18F-FDG as a routine clinical tracer, 2) 18F-FMISO for hypoxia studies as a tracer for human research, and 3) 18F-PD-L1 that specifically binds to human PD-L1 in tumors and other organs as a recent first-in-human tracer. For each tracer, we will investigate 1) static PET, 2) gated and respiratory motion corrected PET, and 3) dynamic PET.

In this proposed project, we will develop a so-called Dynamic Extremity SPECT (DE-SPECT) system that utilizes the 3-D HEXIETC CZT detector technology and a synthetic compound-eye camera design for dynamic and multi-tracer SPECT imaging of PVD in lower extremities. Peripheral Vascular Disease (PVD) affects approximately 8 million Americans and an estimated 10% of the worldwide population, with increasing prevalence in older individuals. PVD has significant health implications, resulting in progressive limb ischemia that can lead to life-altering claudication, non-healing ulcers, limb amputation, and, in severe cases, death. Despite these numbers and a close association with coronary artery disease, PVD remains a relatively under-diagnosed disease. Diabetic patients often require revascularization along with medical and life style interventions to achieve symptom relief, wound healing and limb salvage and have higher rates of restenosis, and higher mortality rates following revascularization when compared to non- DM patients. Although treating large vessel disease and restoration of skeletal muscle perfusion is important for critical limb ischemia and a critical contributor to intermittent claudication, there are other important metabolic and molecular abnormalities involved that need to be evaluated in conjunction with changes in muscle perfusion in order to optimize therapy. Revascularization of obstructed vessels often results in an incomplete or heterogeneous restoration of flow and does not fully restore the functional limitations in PAD patients with intermittent claudication. The proposed DE-SPECT system could provide a unique non-invasive approach for the comprehensive assessment of molecular and physiological changes of the lower extremities in patients with peripheral vascular disease (PVD) in response to therapeutic interventions and assessing these changes will be critical for optimizing and following PVD therapy.

Project Abstract Single Photon Emission Computed Tomography (SPECT) continues to play a critical role in the diagnosis and management of coronary artery disease (CAD). While conventional SPECT scanners using parallel-hole collimators are still the foundation of cardiac SPECT, recently our field observed an exciting growth of new developments of dedicated cardiac scanners. Such dedicated scanners, such as the GE Alcyone 530/570c systems and the D-SPECT systems both with CZT detectors, typically have multiple detectors collecting photons emitted from the heart simultaneously, leading to dramatically improved sensitivity (2-5 X). In addition, the GE systems use pinhole collimators and can achieve much higher resolution. These dedicated scanners opened doors to new applications with significant clinical impact, including ultra-low-dose imaging, absolute quantification of myocardial blood flow (MBF) and coronary flow reserve (CFR), high resolution molecular imaging, multi-isotope imaging, motion correction, and many more. Most of these new applications are uniquely achievable only using dedicated scanners. While the dedicated cardiac SPECT systems can improve clinical practice and lead to numerous new clinical applications, such systems are far from being optimized to fully realize their great potentials. In this grant, we propose to systematically develop and optimize innovative imaging technologies for the GE 530/570c systems to further improve its clinical efficacy in a variety of significant ways. In Aim 1, we will develop and optimize methods for static cardiac SPECT imaging. We will develop various deep learning methods and investigate approaches to increase angular sampling to reduce noise, and improve resolution and quantitative accuracy. In Aim 2, we will develop and validate methods for dynamic SPECT imaging, particularly involving direct parametric image reconstruction. In Aim 3, we will develop and validate methods for dual-isotope SPECT. Monte Carso simulation and deep learning based methods will be developed for tracers with different spatial distributions and fast kinetics. In all three aims, large animal studies and human subject data will be used for optimization and validation.

Project Summary/Abstract Positron emission tomography combined with computed tomography (PET/CT) using the radiolabeled tracer 2- deoxy-2-(18F)fluoro-D-glucose (FDG) has become a standard imaging tool for cancer patient management. The semi-quantitative parameter standardized uptake value (SUV) is routinely used in clinical for tumor uptake quantification, which is computed on the static PET image acquired at a certain time (typically 60 min) post tracer injection for a short interval (typically 5-15 min). However, the quantification accuracy of SUV from a single PET scan suffers from the variabilities of tracer plasma clearance and acquisition start time. The dual- time-point FDG PET imaging has been intensively investigated and used in both clinical and research studies, typically one scan at 60 min and the other at 120 min, showing the potential to enhance the diagnostic accuracy of FDG PET by differentiating malignancy from inflammation and normal tissue. However, the current clinical dual-time-point FDG PET studies use the relative SUV change between two scans as the quantification index, which cannot eliminate the variations in tracer plasma clearance. Meanwhile, the dual-time-point protocol has not been optimized and standardized currently, leading to conflicting results. The fully-quantitative parameter, tracer net uptake rate constant Ki, is the most accurate parameter to quantify FDG PET, which is calculated using dynamic imaging with compartmental modeling. Ki is independent on the plasma clearance or acquisition start time. However, the long and complex acquisition protocol (typically at least 60 min), which requires dynamic scanning and sequential arterial blood sampling (or image-derived blood activity) used as input function from the time of injection, limits its application in clinical practice. Meanwhile, generation of the parametric Ki image, which can provide additional heterogeneity information for FDG PET, is challenging clinically using voxel-by-voxel compartmental modeling due to the computational cost and being sensitive to noise using non-linear least squares. The graphical Patlak plot, can be used for simplified Ki calculation and Ki image generation by voxel-by-voxel fitting. However, it still needs dynamic scanning starting from 15-30 min after injection and input function from the time of injection. The aims of this proposal are 1) to optimize the dual-time-point protocol for accurate Ki quantification using Patlak plot without the need for individual patient's input function, and 2) to generate high-quality low-noise dual-time-point Ki images using novel techniques based on deep learning. Upon the success of this project, our proposed approach can obtain reliable tumor Ki quantification and parametric Ki image for free without adding any additional complexity on the existing dual- time-point protocol currently used in clinical practice, with great potential of improving diagnosis and therapy assessment in oncology. We expect the translation of this approach to clinical investigation to be fast, as this is a post-processing approach and is based on data already acquired using clinically used protocol without imposing additional burden to technologists.