Research On Osteosarcoma Medical Image Analysis System

Osteosarcoma is one of the most prevalent types of malignant bone tumors, which occurs mostoften in children and adolescents with high malignancy and poor prognosis. Early diagnosis plays avery significant role in treatment and prognosis of Osteosarcoma.At present, CT, MRI imaging equipment can only output sequence of2D (two-dimensional) slice,which is not conducive to clinical doctors to observe and analyze the characteristics of Osteosarcomafrom multiple angles visually. In order to improve the accuracy and scientificity of clinical diagnosisand treatment, this paper researches on key technologies in Osteosarcoma medical image analysissystem, and several improvements are made for the existing problems. The main works include:1. Osteosarcoma image preprocessing. Because of Osteosarcoma images vulnerable to noise, wepropose a method based on anisotropic diffusion filter for Osteosarcoma images. Theexperiments show that this method can remove the noise, and retain the image edge feature well.2. Osteosarcoma image segmentation. Considering to complex characteristics of Osteosarcoma, twoimproved segmentation methods are proposed: a) A hybrid segmentation method based on FastMarching Method and Narrow Band Method; b) An interactive hybrid segmentation methodcombining modified relative fuzzy connectedness and confidence connected algorithm. Theexperiments show that the two proposed methods have higher segmentation accuracy, and withhigher similarity index and lower error rate.3.3D (three-dimensional) visualization of Osteosarcoma images. Firstly, researching on theOsteosarcoma images surface rendering based on Marching Cubes method. To watch more realtissues, we focus on the Ray Casting volume rendering algorithm based on GPU, and takeadvantage of hardware acceleration of the programmable GPU to improve rendering speed andaccuracy. Additionally, a user friendly stroke-based interface for the transfer function isproposed to obtain realistic3D model.4.3D interactive measurement. According to Osteosarcoma clinical indicators,3D space linearmeasurement technology is used to measure the distance between tumor edge and joint surface,and the area superposition based on polygonal approximation method is proposed to calculatethe tumor volume. The experiments show that clinical indicators can be well measured usingthese methods.5. Development of Osteosarcoma medical image analysis system. The system is designed anddeveloped in the environment of VC2008++by using MFC, ITK and VTK, which has thefollowing main functions: image preprocessing, image segmentation,3D visualization,3Dinteractive measurement and2D distance and area measurement.The system is user-friendly, and functions better, which is convenient for doctors to observe theOsteosarcoma tissue characteristics in a multi-faceted, multi-angle way, and to achieve quantitativeanalysis of Osteosarcoma clinical indicators. So it provides basis for assisting clinical doctors to judgeOsteosarcoma biological behaviors, and adjust chemotherapy regimens and formulate surgery plans.