PET

The Waisman Lab houses a complete PET research facility, complementing the previously established PET research labs under the Medical Physics department.

Radiochemistry: Radioisotope production is carried out with a National Electrostatics Corporation 9SDH-2 electrostatic tandem accelerator capable of accelerating >100uA of protons or deuterons to 7 MeV with fine beam shaping and positioning capabilities, with high yield targetry systems for 17F, 15O, 13N, 11C, and 18F. A research chemistry lab with dual fume hoods and a class 10,000 clean room with class 100 laminar flow hood are available for radiotracer production and radiopharmaceutical preparation. The lab is well equipped for in-line production and delivery of short-lived tracers. 11C and 18F radiotracers are also available from the UW Medical Physics facility.

Human scanning: The scanning facilities consist of a Siemens ECAT EXACT HR+ PET scanner and its adjacent control room, subject prep room, and blood metabolite analysis lab. A dedicated particle accelerator (see below) and radiochemistry lab support the imaging experiments. The PET scanner room is also equipped with a positional computer monitor for performing computer tasks during the acquisition of the scans.

The EXACT HR+ scanner consists of 4 detector rings of 72 BGO blocks per ring. Each block contains an 8 x 8 array of discrete detector elements with dimensions of 4.39 x 4.05 x 30 mm3, providing 63 contiguous 2-D image planes through an axial field of view (FOV) of 15.5 cm and a patient port diameter of 56.2cm. The rings are separated by extendable tungsten septa for acquisition in both 2D and 3D modes. The transaxial intrinsic resolution of this scanner is 4.3 mm FWHM, and axial intrinsic resolution is 4.7 mm FWHM in the center of the field of view. Reconstructed spatial resolution for a head-shaped object is in a nearly isotropic resolution of 6mm FWHM throughout the entire region of the brain.

Animal scanning: Positron emisssion tomography (PET) is used for in vivo longitudinal studies of potential therapies in animal models of human disease.  Small animal PET moreover permits studies of fundamental biological processes that are not possible with other imaging modalities.  With the advent of high resolution dedicated small animal scanners the field has grown rapidly in the last few years. We have used the microPET P4 small animal scanner (Concorde/Siemens) routinely for in vivo PET brain imaging [1].  From installation (August 2002) to date,  2066 animal scans have been performed using 30 tracers, including 1720 rhesus brain scans. Elements of a typical rhesus scan protocol include the following:

  • Tracers:  All tracers are synthesized by UW Medical Physics and Waisman Center radiochemistry personnel, except for FDG, which is obtained commercially (IBA).  In the case of activation paradigms, FDG is first injected i.v. immediately prior to a 30 minute behavioral session.
  • Anesthesia and monitoring: Either in their home cage or following a behavioral session, rhesus are first anesthetized with ketamine (<15 mg/kg i.m.), transported to the scanner, intubated, and maintained on isoflurane (<2% in 1 L/min O2).  Heart rate, SpO2, respiration, and rectal temperature are measured electronically and recorded every 15 minutes.
  • Positioning:  Rhesus are positioned in a head holder which prevents accidental head motion by means of ear bars, tooth bar, and a pad that presses on the top of the skull.
  • Emission Scan: Data are collected in event mode with a 350-650 keV energy window and 6 ns coincidence window.  In FDG activation studies, the emission scan lasts 45 minutes, and for other tracers, in which the full time course of the pharmacokinetics is studied, scans range from 60-180 min.
  • Transmission Scan: Prior to pharmacokinetic scans and following FDG activation scans, a transmission scan is performed using a 57Co point source.
  • Image Reconstruction: The transmission scan list is sorted into a single sinogram and reconstructed to form an attenuation coefficient (µ) map, which is then smoothed, calibrated, and segmented to values of µ = 0.095 cm-1 in tissue and plastic and 0 in air.  The µ map is then forward projected to form a calibrated attenuation sinogram.  Emission scan events are statically or dynamically framed, e.g. 5x1+ 5x2 + 3x5 + nx10, binned into 3D sinograms (168 projection angles x 192 bins, span 3 ring difference 31, 11 segments), and corrected for detector sensitivity and deadtime.  Emission Images are then reconstructed using filtered back projection (fourier 2D rebinning, 128x128 pixel size 0.94 mm x 0.94 mm in-plane x 1.21mm slice thickness, ramp filter) with the forward projected (µ) map used for attenuation and scatter corrections.

Quality assurance procedures are performed by a full-time technician and the Director of MicroPET according to the following schedule. Daily: Line source scan to confirm detector function; Backup from host to server and external hard drives; Archive to tape and DVDs. Weekly: Line source scan to check sensitivity and resolution; Host computer disk defragment and check; Download and check scanner system report. Monthly: Cylinder phantom scan to check uniformity. Quarterly: Check detector position profiles and energy spectra; Normalization and activity concentration calibration scan; Blank transmission scan. Annual: Lubricate bed and point source lead screws; Clean air intake filter. Ad Hoc: Dr. Converse has repaired the microPET P4 as needed including diagnosis and replacement of a detector block, low and high voltage power supplies, and position sensitive photomultiplier tube signal processing electronics.

We recently acquired a Siemens microPET Focus 220, which provides improved sensitivity and resolution (Fig. 1) [2]. The scanner was purchased with an NIH Shared Instrumentation Grant (1S10RR029358-01A1) and additional funds from the School of Medicine and Public Health and the Graduate School. The scanner was delivered in December 2011 and acceptance testing is complete: Crystal energy resolution = 16.2 +/- 1.7%. Sensitivity at 250-750 keV and 10 ns = 12%. Image resolution = 1.37 mm FWHM. System linearity is within +/- 3% from 1 kHz to 90 kHz . Attenuation and scatter corrections are accurate to 1%. Spatial Uniformity = 3.8% (0.23 mL ROIs in 10 cm phantom). Temporal Noise = 6.6% (0.16 mL ROIs, 100M counts frames, 10 cm phantom). Research rhesus brain scans are expected to begin in June 2012.