During this period, emphasis of the work shifted from head restraint methods to methods of patient repositioning and to methods for reducing head motion not entirely eliminated by the thermoplastic mask restraint system now adopted as """"""""standard"""""""" across the PET program. Testing revealed that the thermoplastic mask cannot, itself, be reliably used to reposition a subject from one imaging session to the next even though the mask is custom-made to each individual and can be attached to the imaging table in only one way. As a result, we used the spatial tracking system described in previous reports to measure (in normal volunteers) the repositioning accuracy of a method based on wall-mounted laser line projectors. Testing of this widely used technique revealed that although portions of the same tomographic plane can usually be replaced in the same imaging slice and within a spatial resolution element, exact repositioning of the entire slice is an uncommon event. Given this pair of results, work on the interactive repositioning system scheme using the spatial tracking device was resumed with testing ex- in the near future. Measurements have revealed that even with aggressive head restraint, residual motions within the restraint continue to compromise image quality. Independent knowledge of head position during the scan procedure can be used to correct these image data for patient motion. Feasibility studies of two such correction schemes have been carried out for both PET and conventional scintigraphic imaging studies with favorable results in both cases. Methods for efficient implementation of these computationally intensive techniques are now under development.
