William D. Gibbs

Noise Identification and Removal in Positron Imaging Systems

The noise (scattered and random) generated in two positron emission imaging systems has been studied and mathematically modeled as an exponential convolving function. Compton scattering is presumed to be a dominant factor, as opposed to random events. Deconvolution of the sensed data has enabled us to eliminate noise and to obtain images which can be quantified. Further, these images have improved diagnostic quality.

Tumor Detection with 1-aminocyclopentane and 1-aminocyclobutane C-11-carboxylic Acid Using Positron Emission Computerized Tomography

Eight patients with proven or suspected neoplastic lesions were examined with 1-aminocyclobutane C-11-carboxylic acid and 1-aminocyclopentane C-11-carboxylic acid using positron emission computed tomography. The results of this comparative study have shown that both of these unnatural amino acids have a high affinity for malignant tumors. The potential of nonmetabolizable C-11-labeled amino acids as noninvasive indicators of metabolic tumor activity is demonstrated and discussed.

Positron emission computerized tomography: A potential tool for in vivo quantitation of the distribution of radiopharmaceuticals

The principles and some of the difficulties in quantitative positron emission computerized tomography have been discussed. We have shown that randoms and scattered events are a major cause of noise and counting errors in positron emission computerized tomography. The noise has been identified as a convoluting process and a mathematical solution has been presented. Examples of phantom studies and in vivo measurements have demonstrated that the distribution of positron emitting radiopharmaceuticals can be quantitated with much improved accuracy using the deconvolution equation to remove undesired noise.