Year 3

In year 3 we mainly focused on the advancement of the hardware aspect of the research, namely the main structure the microscope (Specific Aim 1), and the ancillary equipments (Specific Aims 3). We also continued on characterizing the reporter stem cells (Specific aim 2).
On the Design & fabrication of tri-resolution instrument, we continued to improve the nuclear microscope performance from several angles.
1. Modeling
We seek to understand the tradeoffs inherent in the microscope design (sensitivity, resolution, field of view) and the effect of alignment errors by modeling the system using the industry-standard software package GEANT4 (GEometry ANd Tracking), a simulation toolkit that provides the infrastructure to visualize the detector geometry and particle interactions on the detector. We have established the Macintosh based GEANT package. We have send our researcher to software training then put to work, and produced a simple model of the collimator used in the Nuclear Microscope. We will devise a complete model of the nuclear microscope system during the extension period to inform our decisions about how to position the collimator, detector, and object.
2. Background rate and lead bricks for shielding
The nuclear microscope depends on identifying objects that are producing very low count rates so that even a small background rate from the environment can mask the signal from the object. Although selective “gates” such as an energy window is applied, some of the false signals due to background sources will still be indistinguishable from counts from the object being imaged. We purchased a set of lead bricks and built a lead housing around the system to reduce the background rate. We have managed to reduce the background so that the rate from the smallest source would still be at least 10x more than the rate from background sources. In addition, the system was placed onto an optical table (supported on air bearings) for reduction of vibration noise in the alignment of collimator and detector.
3. Large-area Sensor
We decided to establish a large-area sensor to increase the sensitivity and field of view of the nuclear microscope. A large-area collimator will be matched to the new sensor; the new collimator will have more holes than the existing collimator, and thus drive the sensitivity higher. In addition, the useful field of view of the system will be increased. After careful evaluation and testing, we have established the censoring system with a XRI QuatroSi (300 um thick silicon sensor with 2×2 array of Medipix II readout chips, 50 um pixel size, 512×512 pixel array).
4. Energy calibration
Last year we have reported hardship of the accurate and effective energy calibration, this year we improved the calibration by employing a thin indium sheets between the calibration source and sensor. This creates an input energy spectrum containing two known energies, allowing two-point energy calibration from one source.

Specific Aim 2: Stem cell experiments
In year 3 we planned to adapt microMRI with microSPECT to further characterize the stem cell activity in vivo. To better take advantage of the MR technology, we have extended our therapeutic target and established rat model where injuries were inflicted on the tendon instead of bone of the hind legs. We have implanted the cells and subjected to a few imaging experiments.
Although stem cells were proved to be active prior to implantation and tendon area showed inflammation in MRI, significant stem cell activity could not be established with SPECT. Currently we are trying to investigate different ways of the stem cell implantation as well as more flexible delivery of the Tc99m.

Specific Aim 3: Layered collimator, source, and reconstruction
1. Radioactive Phantom
The nuclear microscope has very high spatial resolution capabilities, far exceeding the ability of standard radioactive phantoms to measure. A point source phantom required to demonstrate such resolution hence has a very high precision and size criteria. Through several design and trills, we have determined on a 44 um-sized metal granules first, then make them radioactive by exposure to neutrons in a reactor, forming Cd-109. The total activity of the granules will be very small, but each granule will be radioactive and could be positioned on a glass slide as a spatial resolution and sensitivity phantom.
2. Software to extract each single gamma-ray interaction from sequence of images
We have devised and tested dedicated software programs to extract and identify single event gamma-ray interactions on the detector. Energy of event and shape of energy distribution for each event is used to partially filter out events due to cosmic rays and other background sources from the desired events from radiotracer-labeled cells. Undesired low-energy events are filtered out by use of a low-energy threshold which does not register an event for interactions that deposit energies below a set level.