Funds are requested for a high speed image processing and display system to be used for the analysis of positron emission tomographic data and for correlation with other imaging modalities such as MRI and SPECT. The proposed system will replace a single user stand alone unit which is no longer supported by the manufacturer. This system is to be located within the PET Facility and available to all investigators for simultaneous display and processing by three users with no apparent degradation in system throughout. Analysis of PET data is complex and computationally bound. The transformation from raw PET projection data into a final parametric image requires numerous steps. Studies are proposed to investigate the effect of these intermediate steps on the resultant images. Current limitation of the existing reconstruction techniques and assemblage of parametric images require several hours of CPU processing time. Techniques that involve rapid reconstruction methods to overcome this problem require special configurations of display/processing memory. The proposed system is ideally suited for examination of these techniques. Concurent display of multimodality images of tomographic data are proposed. Image processing schemes for standardization of region-of-interest analysis are necessary conditions to be able to share PET data with other laboratories. Manipulation of the images to accommodate differences in shape, distortion and artifacts will be studied. Mechanisms for the elucidation of acquisition hardware dependencies such as scanner resolution, dead time and scatter fraction will be investigated. Physiological modeling schemes will be studied that involve estimation of local cerebrallar blood flow (LCBF), receptor density and binding, local cerebellar metabolic rate of glucose utilization (LCMRG) and local protein synthesis for neuro-PET studies. These proposals require enormous computational and image processing capabilities. Numerous proposals involve PET scans for determination of LCBF and LCMRG. Conversion of the raw data to final parametric images occurs several days after the studies are performed. Complete data analysis and reduction of images to specific and meaningful ROI values takes days of the investigator's time. The rate limiting step of data analysis is overcome by simultaneous access to the data by several users and by the greatly enhanced speed of image display, manipulation and calculation.
