three-dimensional transverse section positron imaging: the cone-beam reconstruction algorithm

Around the nuclear world 287

Results will be presented using these techniquesboth for simulation and real data from a high resolu-tion, multiwire proportional chamber positroncamera.

9. Three-dimensional Transverse Section Posi-tron Imaging: The Cone-beam ReconstructionAlgorithm by JAMES G. COLSHER, Searle Diag-nostics. Inc., Des Plaines, IL 60018, U.S.A.

A MATHEMATICAL algorithm for performing fullythree-dimensional transverse section positron imagingis described. The method consists of filtering a three-dimensional back-projected image and is useful foranalyzing data collected when two opposed areadetectors are rotated about the object to be imaged.The principal contribution of this work is the deri-vation of the appropriate three-dimensional filter. Itis noted that the Fourier transform of the filter isnonsingular, allowing the object distribution to be un-iquely determined. The algorithm is tested on phan-tom and clinical data and the results presented.

10. Transverse Section Imaging with the MGHPositron Camera PC11 by G. L. BRO~NELL, C.A. BURNHAM, D. A. CHESLER , B. AHLUWALIAand S. COCHAVI, Physics Research Laboratory,Massachusetts General Hospital, Boston, MA,U S.A.

TI-IE MGH Positron Camera consists of two detectorbanks, each containing 140 detectors. Coincidencesbetween detectors in each of the two banks arerecorded on a time scale of less than 20nsec andthe number of events in each data channel recordedin computer memory. An image of the activity distri-bution between the two detector heads may beobtained by focusing the data channels on planesparallel to the two detector banks.

Transverse section images are obtained by rotatingand translating the two detector banks about theobject. Typically 180 image matrices are obtainedautomatically in one imaging procedure. The time forcollection ranges from 5 to 25 min. Transmission datais obtained by placing a plane source in front of oneof the detectors and repeating the procedure. Fromthis information, 22 emission and 22 transmission CTimages may be obtained spaced at 1.4cm.

Studies have been carried out on sensitivity, maxi-mum counting rate, resolution and linearity of re-sponse for both emission (ECT) and transmission(TCT) images. Phantom studies have been carried outto determine other parameters such as plane thick-ness, variation of resolution with position and effectof prompt and delayed random coincidences.

An intensive clinical program combining conven-tional and transverse section imaging is being carriedout in areas of heart, lung and brain. The high resolu-tion capability of PC11 has been found to be particu-larly useful in imaging the myocardium of the imag-ing CO-labelled hemoglobin distributions in thebody. The measurement such as the determinationof regional cerebral blood flow and glucose metabo-lism.

11. Design and Performance Characteristics ofthe ECAT Positron Tomograph by M. F.PHELPS, E. J. H O F F M A N, S. C. HUANG andD. E. KUHL, UCLA School of Medicine, LosAngeles, CA, U.S.A.

THE ECAT was designed and developed as a positronimaging system capable of providing high contrast,high resolution, quantitative images in two-dimen-sional (2-D) and tomographic formats. The flexibility,in its various image mode options make it useful fora wide variety of imaging problems. High (HR),medium (MR) and low (LR) tomographic resolutionsare 0.95 + 0.1, 1.3 + 0.1 and 1.7 + 0.1 cm FWHM;high, medium and low resolution in 2-D images are0.85 _+ 0.1, 1.3 + 0.1 and 1.7 + 0.1 depending onresolution mode employed. ECT system efficiency is30,100, 15,900 and 9,200 c s-’ #Zi cme3 with a 20 cmdiameter phantom at LR, MR and HR. Because ofthe geometric, detector, electronic and shieldingdesign of the system, countrate capability and linear-ity are high, with minimum detection of scatteredradiation and random coincidence. Measured erroragrees well with theoretical statistical predictionsdown to a level of 1.4% standard deviation. Theredundant sampling scheme of this system signifi-cantly reduces errors caused by motion and detectorinstability. Scan times are variable from 10s to mul-tiple min/slice in which multiple levels are automati-cally scanned by computer control of patient bed. Avariety of human studies illustrate image quality,resolution and efficiency of both ECT and 2-D imag-ing mode. These studies also provide examples of thenon-invasive study method that has been made poss-ible through development of EC’T.

12. Evaluation of a High Resolution Ring Detec-tor Positron Camera System by L. ERIKSSON ,L. WIDEN, M. BERGSTROM, CHR. BOHM, G.FL O O M, T. GREITZ , B. HOVANDER and J. LIT-TON, Institute of Physics, University of Stock-holm; Departments of Neuroradiology andClinical Neurophysiology, Karolinska sjukhu-set, Stockholm, Sweden.

A RING detector positron camera system for highresolution computed tomography of the brain will bedescribed. The system utilizes 9.5 8 x 20 x 50mmNaI(Tl) detectors and records simultaneously coinci-dences from 1900 detector combinations. This corre-sponds to a simultaneous sensing of the activity distri-bution within a 30cm diameter cross-section along1900 detection channels.

When operated stationary, the spatial resolution is11 mm, defined as the FWHM of a reconstructed linesource image. However, the circular geometry of thesystem provides the additional capability to reducethe linear sampling distance by simply observing theactivity distribution in several positions, leaving asa net result a system resolution approaching theFWHM of the detector responses, i.e. 5 mm.

Long term stability tests of the electronic systemhave shown a reliable performance. The overall func-tion has been tested with phantom studies. Unifor-