Testing Technique For Image Performance Parameters Of Spiral Computed Tomography Scanners

With the far and wide application of spiral computed tomography (SCT) scanners,the emphases of quality assurance (QA) study have turned to performance evaluation of spiral scanning,which characterize with changes on such image parameters as longitudinal resolution,noise and artifacts. It brought forward new challenges on testing techniques especially after the invention of multi-slice CT (MSCT). Some relative studies have been done but it is far from application. The purpose of this study is to solve the problems in detail,which is a significant work for the QA system.Starting with analysis on the fundamentals of spiral scanning and conventional scanning,the main performance parameters has been studied deeply,and the testing and evaluating methods were proposed as following:First,testing of slice sensitivity profile (SSP). SSP is a key character of spiral scanning for it reflects the effect of interpolation algorithms on image reconstruction. The common mathematical mode of SSP of spiral scanning modes was induced. A simple but useful phantom was devised and satisfactory results were achieved both on single-slice and multi-slice spiral scanners. It was more accurate to measure slice thickness by full width at half-maximum (FWHM) of SSP than by the method of national standard especially for thin slices less than 5mm. The changes of SSP due to its position in the whole imaging field were discussed too. It was pointed out that the thickness of phantom objective material should be modified according to slice thickness.Second,testing of modulation transfer function (MTF). It is simple and immediate to test spatial resolution by periodic structure phantom. But with improvement of CT technique,more and more problems come out. In contrast,it is more impersonal,accurate and repetitive by measuring MTF. A simple but useful phantom was devised and the results were consistent withthe results by CATPHAN 500 considerably.Third,computer-aided testing of noise and field uniformity. It is to pick up some regions of the whole field to reflect the noise and uniformity of the whole field by the method of the national standard. In view of statistics,it is not the same when there are some local strong noises or CT number fluctuations. A new method aimed at the overall imaging field was given,which could reflect small differences between slices under multi-slice scanning or different scanners.The last,quantitative evaluation of CT number linearity. CT number linearity reflects the authenticity of images. The problem of conventional testing method is short of a quantitative evaluation parameter. A new definition was brought forward based on regression analysis of testing data. Its feasibility was validated through a mass of experiments. It is pointed out improving method of the phantom,which demonstrates the directions of farther studies.The main innovations in the study are below:1. SSP was brought into testing items of spiral scanning for the first time,and its testing technique was given in detail. It was far more accurate to reflect slice thickness by full width at half-maximum (FWHM) of SSP than national standard especially for thin slice less than 5mm.2. An improved testing method of noise and field uniformity was put forward whose characteristic was numerical comparability. It could distinguish the small differences between images under multi-slice axis scanning.3. A quantitative evaluation of CT number linearity was brought forward.