The Research On PET/CT Alignment Phantom And Testing Methods

Accurate anatomical information is a prerequisite for successful radiation therapy planning and delivery to the entire extent of tumor,while minimizing dose to normal tissues surrounding it.For this reason,images in radiation oncology is very important.In recent years, the improvement from imaging modalities like computed tomography (CT),magnetic resonance imaging (MRI) and positron emission tomography have resulted in many advances in radiation oncology. In principle,CT is a morphologically oriented imaging method. Unlike CT,PET is a functionally oriented method. PET with an18F-FDG tracer utilizes the mechanism of biochemical reactions inside the patient and uses coincidence detection of annihilation photons to determine the tumor position.Both the imaging methods have advantages as well as limitations. CT provides high resolution imaging but has a difficulty in distinguishing between tumors and healthy tissues.PET is sensitive to small differences in metabolism between tumors and healthy tissues,however spatial resolution provided by PET is low.These limitations of CT and PET can be addressed by using the combination of both modalities.The combination of PET and CT scanning offers unique opportunities for oncology.PET/CT has the ability to provide a synergistic combination of PET and CT images, which could potentially be more valuable than the two examinations performed separately.Directly combining functional and morphological information has been shown to improve the accuracy and efficiency of PET. And a major advantage of integrated PET/CT over dedicated PET is that it allows the use of CT data for attenuation correction and for anatomic localization. It not only shortens the scanning time,but also improves the accuracy of diagnosis for lesions.Therefore,radiotherapy treatment planning utilizing PET and CT dual modalities is rapidly gaining acceptance in oncology. A limiting factor of the dual modal is the PET/CT alignment. A small error in PET/CT alignment may result in giving large doses of radiation to healthy tissues as a result of poor treatment planning.To ensure PET and CT images meeting the quality for clinical diagnosis and guidance of treatment planning, quality control of PET/CT must be performed.The concept of quality control is that using a series of specific methods makes nuclear medicine equipment working in optimal condition.It is divided into two parts,the performance test and the routine maintenance and preventive maintenance.And PET/CT performance parameters are representing the quality of PET/CT system which includes three parts. One is PET part performance parameters,such as spatial resolution,sensitivity and so on.Second is CT part performance parameters,such as low contrast resolution,noise and so on.Last is PET/CT performance parameters,such as image alignment,which is this study highly focused on.Image alignment is the parameter reflecting the precision of PET and CT image fusion. As an integrated system,PET/CT device must have a good image fusion accuracy, because it not only affects the accuracy of spatial localization of lesions,but also affects the accuracy of PET data attenuation correction.However,the inconsistency in data acquisition speed and spatial resolution between PET and CT has resulted in a few errors for PET and CT fusion image.The image alignment measurement can effectively guide the acceptance test, calibration test, maintenance test and failure testing of PET/CT systems.The research focused on the image fusion accuracy,designed and produced PET/CT image fusion accuracy phantom which was tested.The shape of alignment phantom is cylinder,the outer diameter is190mm,the thickness is5mm and the material is polymethyl methacrylate (PMMA).And the phantom consists of four cylindrical rods as the cold area and three tubes as the hot area while correcting data.In this research the phantom was scanning by two different scanning conditions when testing.To get the image fusion errors,we need to measure the center coordinate of objects from both PET images and CT images.Firstly,this research uses MATLAB to process CT images. In order to measure the center coordinates of objects from CT images,it is needed to detect the center of circle for images.To improve the precision of results,it is needed to make image preprocess before performing the Hough transform.Preprocessing steps include image binarization and edge detection.Edge detection identifying the points of obvious brightness variations in digital images is the basic technique in image processing and computer vision. There are many edge detection operators,this paper adopted traditional Prewitt operator,which has the character of strong anti-noise performance,low detection failure rate and good detection effects.In this research,we usually adopted Hough Transform algorithm to detect the center of objects after the image preprocessing. Hough Transform is a method that connected the edge pixels of image into closed region boundary by using the global feature of image, and Hough Transform realized the mapping relationship from the image space to the parameter space. This paper described in details the line detection and circle detection based on traditional Hough Transform. The basic idea of traditional Hough Transform detection algorithm is to transform the image space domain into the parameter space, and described the curve boundary of image with the boundary points which met some form of parameters. Because of the parameter space of circle is three dimensional, circle detection needed to consume a amount of computation time and occupy a lot of storage space.Therefore, Random Hough Transform algorithm which reduced the invalid cumulative detection, the amount of computation and the need of resource for improving the speed of detection was adopted in this research.Secondly,this study used IDL to analyze PET images. A PET image was displayed and we draw a profile across the object using an in-built profiling tool,PROFILE,in IDL to measure the center coordinate of objects.Finally,the difference of two center coordinates is the alignment error.The self-made phantom should provide the most reliable measurement of alignment error with the analysis, and it can also be used as a routine PET/CT performance quality control testing.