Penn PET Explorer

Abstract

Objectives: The nicotine metabolite ratio (NMR), a stable measure of hepatic nicotine metabolism via the CYP2A6 pathway and total nicotine clearance, is a predictive biomarker of response to nicotine replacement therapy, with increased quit rates in slower metabolizers. Nicotine binds directly to nicotinic acetylcholine receptors (nAChRs) to exert its psychoactive effects. This study examined the relationship between NMR and nAChR (α4β2* subtype) availability using PET imaging of the radiotracer 2-18F-fluoro-3-(2(S)-azetidinylmethoxy)pyridine (2-18F-FA-85380, or 2-18F-FA).

Methods: Twenty-four smokers—12 slow metabolizers (NMR < 0.26) and 12 normal metabolizers (NMR ≥ 0.26)—underwent 2-18F-FA-PET brain imaging after overnight nicotine abstinence (18 h before scanning), using a validated bolus-plus-infusion protocol. Availability of nAChRs was compared between NMR groups in a priori volumes of interest, with total distribution volume (VT/fP) being the measure of nAChR availability. Cravings to smoke were assessed before and after the scans.

Results: Thalamic nAChR α4β2* availability was significantly reduced in slow nicotine metabolizers (P = 0.04). Slow metabolizers exhibited greater reductions in cravings after scanning than normal metabolizers; however, craving was unrelated to nAChR availability.

Conclusion: The rate of nicotine metabolism is associated with thalamic nAChR availability. Additional studies could examine whether altered nAChR availability underlies the differences in treatment response between slow and normal metabolizers of nicotine.

Explorer Application: The ability to visualize whole body 2-[18F]FA distribution and measure differences between clinically relevant phenotypes offers a unique opportunity to understand pharmacokinetics and identify potential targets for pharmacological manipulation.

Research Support Pharmacogenetics Research Network Grant (NIDA/NCI/NHGRI/NIGMS), Abramson Cancer Center at the University of Pennsylvania, ITMAT UPENN (NIH), McCabe Pilot Award (UPENN)

Close

Abstract

Introduction: Oncogene-dependent reliance on glutamine is a cancer vulnerability that can be exploited for therapeutic gain. Inhibitors of glutaminase, the enzyme that catalyzes the conversion of glutamine to glutamate, have been developed and have shown antitumor effects in several tumor models. The specifıc and potent glutaminase inhibitor CB-839 demonstrated marked antitumor effıcacy in a triple-negative breast cancer (TNBC) cell line with inherently high glutaminase activity (HCC-1806), but not in an estrogen receptor-positive (ER) cell line with inherently low glutaminase activity. An early phase 1 clinical trial of CB-839 in combination with paclitaxel has shown encouraging results in TNBC patients. Cell-line specifıc effıcacy supports the need for clinical biomarkers to predict and evaluate therapeutic effıcacy. As a minimally metabolized glutamine analog with similar transport properties, [18F](2S,4R)4-Fluoroglutamine ([18F]4F-Gln) is an ideal radiotracer to infer glutamine pool size through estimates of tracer distribution volume (Vd). In this study, we demonstrate differences in Vd of [18F]4F-Gln in two xenografts with different levels of glutaminolysis (TNBC vs. ER) using dynamic PET imaging, as well as show the effect of anti-glutaminase therapy on [18F]4F-Gln Vd.

Methods: TNBC (HCC-1806) and ER (MCF-7) xenografts were established in athymic nu/nu mice. Mice were scanned in a dedicated animal PET scanner at baseline and after CB-839 administration. Dynamic imaging was obtained for one hour upon injection of [18F]4F-Gln (300-350 μCi) via the tail vein.  The field-of-view of the scanner captured the entire mouse allowing an image-derived blood input over the heart to be captured. Kinetic analysis was performed with PMOD.

Results: [18F]4F-Gln uptake was largely reversible with Logan plots demonstrating late linearity and k3 in a two-compartment (trapping) model was low (< 0.01/min in most cases). Strong correlation was seen in Vd estimates by Logan plot and a single-compartmental model (R2 >0.9), but not with estimates from a two-compartment model. MCF-7 xenografts demonstrated greater than 60% larger Vd at baseline than TNBC xenografts indicative of increased cellular glutamine concentrations in MCF-7 xenografts. An increase in Vd was detected in TNBC xenografts post-glutaminase inhibition (>30% mean change), but not in MCF-7 xenografts.

Conclusion: Estimates of [18F]4F-Gln Vd obtained through kinetic analysis offer the ability to non-invasively infer tumor glutaminolysis. Low [18F]4F-Gln uptake in highly glutaminolytic tumors and an increase in [18F]4F-Gln uptake with glutaminase inhibition is concordant with changes in cellular glutamine concentration as measured by MR spectroscopy of tumor extracts.  These results underscore the utility of dynamic imaging and kinetic analysis in the analysis of this radiotracer.  Translation to humans would benefit from a long axial field-of-view scanner whereby dynamic imaging of the entire patient could be obtained.  This would enable the construction of time activity curves for all sites of disease, as well as an image-derived blood input function.  Acquisition of such data would enable kinetic modeling and subsequent characterization of each of the patient’s tumors, capturing biologic heterogeneity and possibly informing subsequent treatment.

Supported by: Komen SAC130060, DE-SE0012476, R21-CA198563 We thank Calithera Inc. for providing CB-839 and Dr. Susan Demo for discussions.

Close

Abstract

Introduction: Poly(ADP-ribose) polymerase (PARPi) inhibitors are emerging targeted therapeutics for the treatment of ovarian cancer and were initially recognized to work in BRCA1/2 mutation carriers. However, a recent clinical trial showed all platinum sensitive ovarian cancer patients benefited from the second generation PARPi, niraparib, suggesting newer biomarkers would be useful for patient selection. Second generation PARPi are more effective at trapping PARP-1 on DNA lesions and have improved cellular lethality compared to first generation PARPi. Due to the critical role of PARP-1 in the PARP trapping hypothesis, PARP-1 expression has the potential to serve as a biomarker for the appropriate selection of ovarian cancer patients for PARPi therapy. In this work we present the non-invasive real time imaging of PARP-1 through positron emission tomography with a radiolabelled PARPi, [18F]FluorThanatrace ([18F]FTT) in primary and disseminated ovarian cancer as a novel approach to quantify PARP-1 expression and monitor patients on PARP inhibitor therapy.

Methods: Patients underwent [18F]FTT PET/CT imaging on a whole-body PET/CT scanner (Phillips Medical System, Netherlands). Image reconstruction was carried out using standard procedures. Images were interpreted by trained nuclear medicine physicians who first located lesions through the patients most recent clinical [18F]FDG study and then found corresponding lesions on the [18F]FTT image and quantified standard-uptake-values (SUV’s). Immunohistochemistry using clinical adjacent tumor sections was performed for hematoxylin and eosin (HE) and biomarkers including, PARP-1, γH2AX, p-53, LCA, and AE1-3. Furthermore, we evaluated PARP-1 by both fluorescent IHC and [125I]KX1 autoradiography on adjacent sections and was correlated with [18F]FTT PET SUV’s for known sites of disease.

Results: We enrolled 10 EOC patients, 7 of which underwent [18F]FTT imaging in addition to standard clinical management including surgical debulking or biopsy. A total of 14 tissue specimens were available from the 10 patients and were used for in vitro assays including PARP-1 fluorescent IHC and [125I]KX1 autoradiography. We observed a spectrum of (SUV) of [18F]FTT. Tumor SUV’s were observed as low as 2 (background) and as high as 12.1. In addition, we found a linear correlation (r2 = 0.7368, 0.7906) between PARP-1 fluorescent IHC and [18F]FTT PET imaging, as well as [125I]KX1 autoradiography. No correlations were found between PARP-1 fluorescent IHC or [18F]FTT and [18F]FDG.

Conclusion: We found that [18F]FTT SUVs positively correlate with PARP-1 expression by fluorescent IHC. This provides early proof of concept of [18F]FTT as a clinical biomarker of PARP-1 expression in primary and disseminated disease. The spectrum of PARP-1 expression measured by [18F]FTT ranged from 2 to 12 SUV’s and provides preliminary evidence for differences in PARP-1 expression effecting PARPi tumor targeting. Future studies are now underway evaluating [18F]FTT as a biomarker to assess PARPi drug target engagement with the goal of entering multi-center clinical trials in ovarian, breast, and prostate cancer.  

Close

Abstract

Dopamine D3 receptors are thought to represent principal neural substrates in the dopaminergic mesolimbic circuit, the primary reward pathway in the mammalian brain. D3 receptor dysfunction has been implicated in many diseases including drug addiction, Parkinson’s disease, and schizophrenia. The purpose of this study was to assess the biodistribution and radiation dose resulting from administration of the dopamine D3 receptor selective PET radiotracer [18F]fluortriopride ([18F]FTP) in healthy humans. All dose estimates were calculated using the Standard Adult Male phantom in the OLINDA | EXM v.1.1 software. The highest individual organ dose was to the gallbladder. Simulation of the impact of having all patients ingest a serving of a fatty-meal equivalent (Ensure Plus) following completion of their [18F]FTP PET imaging session predicted their gallbladder dose would be reduced by 49%. An injected dose of 240.5 MBq yielded an effective dose of 5.7 mSv and confirmed expectations that biodistribution and levels of absorbed doses were adequately low to continue researching [18F]FTP radiotracer uptake in humans. Patients should consume a fatty-meal equivalent following an [18F]FTP PET scan to reduce radiation dose to their gallbladders in half.

Explorer Application: The Penn PET Explorer scanner will allow us to simultaneously image the whole body, which will dramatically improve the temporal resolution of time activity curves in organs from this study’s 7 static PET measures over the course of 4 hours to a PET measure at least every 5 minutes to enhance patient safety by greatly increasing the accuracy of our radiation dose estimates.

Acknowledgements: NIH/NIDA: K01 DA040023 (Doot), K23 DA038726 (Dubroff), & R01 DA029840 (Mach). The authors also appreciate the support of the University of Pennsylvania NM-AIA lab and the University of Pennsylvania McCabe Pilot Award (Dubroff).

Close

Abstract

Clinically, PET/CT is a widely-used technology in oncology and cardiology. In the realm of research, dynamic PET imaging provides valuable information about ongoing biologic processes in the body. Currently, all commercial PET scanners have an axial field of view (AFOV) less than 25 cm, and whole-body scans are taken in sequential, step-and-shoot bed positions, or possibly continuous bed motion. Recently, however, there has been interest in developing PET systems with significantly longer AFOVs in order to image the whole body simultaneously. The increased sensitivity from extending the AFOV will allow for lower dose imaging and improved count statistics, but the most significant advantage of a large axial FOV scanner for research applications may be that it will allow simultaneous whole-body static and dynamic imaging over a larger axial extent of the body. To quantify the performance of a PET scanner with a large axial FOV, we have extended the GATE simulations of the 16.4-cm AFOV Philips Vereos PET/CT scanner to 23-cm and 70-cm AFOV PET scanners. The simulation model is based on the digital detectors used in the Vereos scanner that has a timing resolution of 320 ps. NEMA NU2-2012 sensitivity, spatial resolution, count rate, and image quality standards were run on both scanner geometries. Using software for data correction and image reconstruction that we have adapted to these new geometries we find little variation in contrast recovery coefficient (CRC) values, although there were notable differences in sensitivity, spatial resolution, and count rate as expected. To further characterize the performance of the larger axial FOV scanner under realistic conditions, an anthropomorphic female was simulated using the XCAT phantom. Four lesions (2 10-cm and 2 7.5-cm) were embedded in the lung, liver, and breast each. To quantify the efficacy of dose reduction, precision and CRC were calculated. Moving forward, the XCAT phantom with realistic activity distributions will assist in planning dynamic patient studies on the 70-cm PennPET Explorer prototype that we expect to be operational in late 2018. We also plan to expand the simulations to test the merits of scanners with even longer AFOVs up to 2 meters.

Close

Abstract

A long axial field-of-view PET scanner offers new clinical and research imaging opportunities not available in commercial whole-body scanners. However the long axial FOV data brings considerable challenges, such as strong spatially and view dependent resolution variations in data and image spaces and dramatically increased volumes of acquired events and data sizes, significantly challenging both the acquisition hardware and processing and reconstruction tools. The DIRECT (Direct Image Reconstruction for TOF) reconstruction framework uses efficient histo-image partitioning and is well suited to address these challenges. In this presentation we outline the DIRECT framework and its application to data with large axial acceptance and long axial extent - the maximum obliqueness is +/-67˚ for a 2-meter long scanner. We further demonstrate data resolution characterization within histo-image partitioning and illustrate utilization of efficient view (tilt) dependent resolution models. Finally we demonstrate the application of these DIRECT reconstruction tools to simulated data of the long axial FOV PET scanner geometry.

Close

Abstract

We are developing a prototype Explorer scanner that will have advanced time-of-flight (TOF) capabilities and 70-cm axial field-of-view (FOV). This scanner will be capable of total-body imaging for very young patients, which will enable a wide range of new imaging applications while also allowing very low-dose or short-time acquisitions. In this work we investigate the improvement from TOF-PET in pediatric patients using a combination of computer simulations of a small diameter phantom and by assessing clinical images acquired on a TOF-PET/MRI scanner with 375 ps TOF resolution. TOF-PET images appear visually improved for TOF resolutions better than 600 ps and have a lower bias and variability, allowing for either a reduction in the injected activity or a reduction in the scan time. We also investigated how extending the axial FOV and improving TOF, with 72-cm length and 200 ps, combine to improve image quality using our small diameter phantom. Significant gains are expected for pediatric imaging applications with advanced TOF performance and a long-axial FOV of ~70 cm.

Close

Abstract

Dynamic PET image reconstruction is a challenging problem because of the ill-conditioned nature of PET and the low-counting statistics resulted from short time-frames in dynamic imaging. The kernel method for image reconstruction has been developed to improve image reconstruction of low-count PET data by incorporating prior information derived from high-count composite data. In contrast to most of the existing regularization-based methods, the kernel method embeds image prior information in the forward projection model and does not require an explicit regularization term in the reconstruction formula. Inspired by the existing highly constrained back-projection (HYPR) algorithm for dynamic PET image denoising, we propose in this work a new type of kernel that is simpler to implement and further improves the kernel-based dynamic PET image reconstruction. Our evaluation study using a physical phantom scan with synthetic FDG tracer kinetics has demonstrated that the new HYPR kernel-based reconstruction can achieve a better bias versus standard deviation trade-off for dynamic PET parametric imaging than the post-reconstruction HYPR denoising method and the previously used nonlocal-means kernel.

Close

Abstract

This poster describes the mini-EXPLORER PET scanner developed at UC Davis in collaboration with Siemens Medical Solutions, and currently installed at the California National Primate Research Center. The mini-EXPLORER system was built using the detectors and electronics from a clinical PET system, and reconfigured into a long axial field-of-view scanner for non-human primate imaging. Here we present the scanner’s physical performance, along with an initial assessment of the in vivo performance of the mini-EXPLORER scanner with a 60 minute dynamic ¹⁸F-FDG rhesus monkey imaging study as well as a late time-point image at 18 hours post-injection.

Close

Abstract

The UC Davis EXPLORER project aims to build the first 2m long total body human PET imager. For this endeavor we have partnered with United Health Imaging to develop the PET system. In a first phase a 50 cm long and 55 cm diameter prototype (Mini EXPLORER II) was built. This prototype, aimed for PET imaging of companion animals (e.g. dogs and cats) and non-human primates, was recently completed and its performance was evaluated. The PET scanner uses the exact same PET detector modules as the human EXPLORER. Each module has 70 detector blocks which consist of 6x7 arrays of 2.76x2.76x18 mm3 LYSO scintillator blocks (96.8% packing fraction) coupled to 2x2 arrays of 6x6 mm2 Silicon photomultipliers (SensL). The complete system consists of two axial sections with 16 modules each and will allow us to evaluate performance and reliability of the PET components as well as new aspects like real-time cross section coincidence formation. The PET system is combined with a clinical 16-slice CT scanner that is used for attenuation correction Our performance evaluation showed a global coincidence timing resolution of 410 ps and an average energy resolution 11.7% (FWHM @ 511 keV). The spatial resolution was 2.6 mm and the (NEMA) sensitivity was approximately 58 kcps/MBq.  A dynamic PET scan with [18F]-FDG of a rabbit was performed to showcase the advantage of high detection sensitivity and total body coverage.

Based on these promising results the construction of the 8 section human EXPLORER system is now on its way.

Close

Close

Abstract

Respiratory and body motions can cause significant image blurring and tracer concentration misestimation. In addition, any motion between CT and PET can cause attenuation mismatch artifacts. For dynamic PET, the additional requirement of continuously correcting each dynamic frame needs sophisticated algorithms. The Yale PET Center has developed a non-rigid event-by-event motion compensated reconstruction platform – non-rigid MOLAR (NR-MOLAR), which has the potential for the greatest accuracy for both static and dynamic PET. For respiratory motion, we developed a fully automated framework, AIM (APN-INTEX-MOLAR, where APN stands for automated phase-matched non-rigid, that 1) automatically determines the reference respiratory phase matching that of the CT scan; 2) builds non-rigid internal organ motion and external respiratory signals correlation (INTEX) based on phase-matched gated-PET reconstructions and 3) non-rigidly corrects both intra- and inter-cycle breathing variations in an event-by-event fashion. For body motion, we investigated an automated data-driven body motion detection method based on a centroid-of-distribution (COD) trace, followed by an event-by-event non-rigid correction. Body motion between CT-PET scans was corrected by registering the CT image to MLAA generated attenuation map. For respiratory motion correction, the new framework reduced artifacts in the lower-lung regions, maintained the accuracy of bone marrow activity estimation, and achieved higher tumor contrast recovery. For the body motion compensation, COD can sensitively and reliably detect body motion. With intra-frame motion correction, the event-by-event correction showed a substantial improvement over conventional frame-based registration approach in static and dynamic studies. For dynamic PET, both respiratory and body motion corrections substantially improved the accuracy of parameter estimation. In conclusion, our non-rigid event-by-event respiratory and body motion corrections are promising for improving both static and dynamic PET.

Close