Abstract

Removal of interplane septa in a PET scanner en- ables acquisition of all possible lines of response (3-D mode) in an effort to maximize the available number of detected events. One problem with this method at high count rates, however, is a markedly increased deadtime and randoms rate, which has a deleterious effect on data quality. The noise-equivalent count- rate (NEC) performance of a neuro-PET scanner has been de- termined with and without interplane septa on uniform cylin- drical phantoms of differing radii and in human studies to assess the optimum count rate conditions that realise the maximum gain. In the brain, the effective gain in NEC performance for 3-D ranges from >5 at low count rates to ~3.3 at 200 kcps (equivalent to 37 kcps in 2-D). The gains of the 3-D method assessed by this analysis are significant, and are shown to be highly dependent on count rate and object dimensions. I. INTRODUCTION OMPARISONS of count rate performance between C different positron emission tomography (PET) scan- ners, or a single scanner operating under different condi- tions, are difficult to make because of the vastly different strategies for handling physical effects such as scattered events and random (accidental) coincidences. A naive as- sessment of acquired coincidences will fail to examine the numbers of scattered or random events included in this figure. Differences in design, composition of detectors and septa together with methods for handling corrections ne- cessitates a method for analysing such data that includes these effects and their contribution in degrading image quality. The noise-equivalent count rate (NEC) method proposed by Strother et a/. (I) provides a means for mak- ing meaningful intercomparisons that incorporate these ef- fects. This analysis technique has been applied to a neuro- PET scanner (953B, Siemens-CTI, Knoxville, TN) which can operate with or without interplane septa. The scanner utilizes block technology detectors and contains 16 indi- vidual rings each 6.5 mm thick axially, which results in

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call