Research And Development Of Software For Biological Target Volume Delineation For Intelligent Radiotherapy Of Malignant Tumors

At present, cancer has become the first leading reason of human death, and more than 70% cancer patients need radiation therapy. Usually, clinical radiotherapeutic physicians manually delineate the gross tumor volume(GTV)slice by slice in 2D cross-sections of PET(positron emission tomography)、CT(computed tomography) images of tumors, they design and make the tumor radiotherapy plan based on this delineation result. Therefore, the accuracy of the delineation of GTV is the key factor to the success of radiotherapy.We design and develope the software for GTV delineation and evaluation for intelligently biological IMRT, which based on the framework of the molecular imaging guided intelligent radiation system. The subsystem is a computer software system for malignant tumor, which is a simple, friendly interface and high intelligent tools for GTV delineation and evaluation.The details of the software are as follows:(1)A module of patients for standard DICOM( Digital Imaging and Communications in Medicine)radiation therapy files such as CT, PET, MRI, and RS images, which is used for the pre-process of image information and radiation therapy structure set information, and worked as the basis for the GTV delineation and evaluation.(2)A PET based GTV delineation module includs the adaptive and two-stage region growing algorithm,which consists of the three-layer implementation, the user interface, the core algorithm layer and the system interface.(3)A DICOM file maintenance and evaluation of GTV module, which creates a DICOM RS file of GTV, and provides a quantitative and visual assessment with the "GTV truth standard" image.However, the manual GTV delineation is strongly subjective,and delineation artificially is time-consuming. To overcome the shortcomings, this paper realizes adaptive and two-stage region growing algorithms for GTV delineation,provides an automatic delineation platform.We design and develop a GTV delineation system for PET molecular biological image. By improving the traditional reading and writing DICOM files, the interface module between GTV delineation system and clinical radiotherapeutic RS files is achieved. The system utilizes three evaluation indexes of similarity, sensitivity and specificity and provides a quantitative and visual assessment with the "GTV truth standard" image. Furthermore, the delineation results and multimodality images are displayed simultaneously which provide a more intuitive visual assessment for users. The system developed in our work enhances the interactive experience for clinical radiotherapeutic physicians and has better applicability and scalability.