Increased sensitivity and field of view for a rotating positron camera.

Large area positron cameras for three-dimensional imaging in medicine have been under development for several years at a number of centres (McKee er a1 1979, Bateman er a1 1980, Jeavons et a1 1981, Del Guerra er a1 1983), and some groups have reported clinical applications (Ott er a1 1983, Townsend et a1 1984). To reconstruct the data, an algorithm based on a backprojection and frequency space deconvolution approach (Chu and Tam 1977, Schorr and Townsend 1981, Schorr et a1 1983) has been adopted by each group. Such deconvolution methods have the advantage of being computationally efficient but require that the point response function be shift invariant (Chu and Tam 1977). In this context, the point response function is the backprojected image of a unit point source of activity and ‘shift invariance’ means that displacing the source causes no change in the point response function other than an identical displacement. For a stationary, dual detector camera, shift invariance is ensured by rejecting positron annihilation event-lines (figure 1 ) with directions that fall outside specific angular