Multi-Tracer PET Assessment of Primary Brain Tumors

Positron emission tomography (PET) is a molecular imaging modality that can probe various aspects of tumor function using a variety of radio-labeled imaging agents ("tracers"). Oncologic PET imaging has seen a dramatic rise in clinical utilization over the past decade for cancer detection, staging, and evaluating residual or recurrent disease following therapy. These clinical scans use the tracer [18F]fluoro-2-deoxy-D-glucose (FDG), which accumulates in cells in proportion to glucose (GLUT) transporter and hexokinase activity. FDG thus provides a measure of tissue glucose metabolism. Concurrent with this clinical growth, a number of other PET tracers have received significant attention in research for a variety of imaging targets. Of special interest are the tracers 3'-deoxy-3'-[18F]fluorothymidine (FLT), 1-[11C]-acetate (ACE), and [15O]water (H2O). The uptake, retention/washout, and ultimate biodistribution of these tracers are each related to different functional or molecular processes. As such, each can be used to probe a different aspect of tumor function: FLT directly assesses tumor proliferation, ACE provides a measure of tumor growth related to fatty acid and membrane synthesis, and H2O quantifies tumor perfusion.

OBJECTIVES:

This study has two primary objectives: a translational objective in which a new PET imaging technology will be translated from experimental development (with simulations and in animals) to the first use in human subjects; and an exploratory objective in which the complementary value of multiple PET tracers will be investigated. Each of these objective is described below, where the study design has been carefully setup to fulfill both objectives in the same study population.

The translational objective of this study is to implement and evaluate a new imaging technology for rapid, single-scan multi-tracer PET imaging of these tracers. Current PET technology prohibits imaging of more than one tracer in a single scan since the imaging signals from each tracer cannot be distinguished by normal techniques; as such, separate scans with each tracer currently need to be acquired hours or days apart. Our group has developed techniques and algorithms for recovering individual-tracer images from rapidly-acquired multi-tracer PET data using dynamic imaging techniques. These methods have been tested through extensive simulations and verified experimentally in a canine model with spontaneously-occurring tumors. Refinement of the methods with more advanced algorithms is ongoing. The patient imaging studies of this protocol will be implemented in two phases. In Phase A, separate single-tracer imaging of each tracer will be performed. The data from these scans will be co-registered and combined to "emulate" multi-tracer scans, which will then be processed by the multi-tracer signal-separation algorithms. This will permit a direct comparison of imaging biomarkers from multi-tracer vs. single-tracer scans for each tracer. Such comparison techniques have been established by the investigators and have been accepted by peer review for testing multi-tracer signal-separation algorithms. Once statistically-significant evidence is obtained that multi-tracer scans can accurately provide the same imaging biomarkers as separate single-tracer scans, the imaging will transition to Phase B-in which actual multi-tracer scans will be performed.

The objectives of this exploratory study is to preliminarily evaluate the complementary value of FDG, FLT, ACE, and H2O PET in patients with primary glial neoplasms. Multi-tracer PET profiles with these four tracers will be obtained in 20 patients with primary glial neoplasms at up to three timepoints: (1) at "baseline" prior to surgery or immediately after surgery providing a complete surgical resection was not possible and confirmed by a post-operative contrast MRI scan where residual tumor greater than 1.0 cm in diameter was present and prior to any tumor-directed therapy; (2) at the conclusions of the initial (~6-8 weeks) chemoradiotherapy; and (3) at the time of MRI-documented recurrence within 2 years. In addition, patients with a known primary brain tumor who have previously undergone treatment and have recurred based on standard clinical and imaging criteria will be eligible for the study. A number of quantitative and pseudo-quantitative imaging biomarkers for each tracer will be computed at each imaging timepoint, and the change in each biomarker between timepoints will also be computed. These data will be compared with clinical endpoints (survival, time to progression), and with tumor biologic information (histology, WHO grade, vascularity, Ki-67, vascular endothelial growth factor (VEGF), epidermal growth factor receptor (EGFR), p53) in cases when tumor tissue becomes available from standard care. These data will provide pilot information into the potential value of concurrent multiple PET biomarkers for predicting tumor behavior prior to the start of therapy, for improved prognostication, for more efficient and effective tumor surveillance, and/or for more appropriate assignment of patients to conventional, aggressive, or investigational therapies early in their clinical courses.

The driving hypothesis for the overall line of research is that multiple PET imaging biomarkers obtained in conjunction can provide improved image-guided personalized care of patients with primary glial neoplasms. The term "personalized care" is used here to broadly include the prediction of tumor behavior prior to the start of therapy, tumor surveillance, prognostication, and individualized assignment of patients to conventional, aggressive, or investigational therapies early in their clinical courses. This pilot project will obtain initial data on the value of these PET biomarkers for such image-guided personalized care.

Specific hypotheses to be tested include:

• HYPOTHESIS I a: Rapid, single-scan multi-tracer PET imaging can recover PET imaging biomarker information of each tracer that are not significantly different from those obtained from conventional, single-tracer scans of each tracer.
• HYPOTHESIS II b: Multi-tracer PET biomarkers, obtained in conjunction, are better able to predict tumor aggressiveness than individual-tracer biomarkers or conventional radiographic imaging.
• HYPOTHESIS III b: Multi-tracer PET biomarkers, obtained in conjunction, are better able to detect functional changes in tumor state that occur in response to therapy than individual-tracer biomarkers or conventional radiographic imaging.
• HYPOTHESIS IV b: Characterization of multiple aspects of tumor function (glucose metabolism, proliferation, membrane growth, and perfusion) provides new insight into tumor status that can guide selection of the most appropriate therapy. a Sufficient statistical power is expected to be obtained under this protocol to validate the extensive simulations and experimental evaluations performed previously and concurrently with these patient imaging studies. b Pilot data regarding these three hypotheses will be obtained in this work by studying the correlation of PET imaging biomarkers with clinical outcomes and tumor biologic information. Though high statistical power cannot be expected from the limited number of patients in this pilot study, underlying trends in the data will be identified, permitting the formulation of formal hypotheses to be tested in future rigorous trials.