2D-3D Registration In X-ray Image-guided Pulmonary Puncture Navigation Systems

Lung cancer is a disease with a high morbidity and mortality rate in human cancers.It is the leading cause of cancer death in men and the second leading cause of cancer death in women all over the world.With the continuous development of technologies such as medical imaging and computer science and the continuous innovation and breakthrough of clinical medical technology,the application of computer-aided diagnosis,surgical and threapy is more and more widely used in clinical practices.Puncture biopsy is a minimally invasive interventional surgery and is currently widely used in clinical examination and treatment of cancer.In this paper,we found that the traditional CT image-guided puncture biopsy is very difficult and the surgical procedures for doctors are cumbersome,what’s worse,multiple scans of patients are needed cause with huge radiation problems.All above these problems put forward the application of X-ray images guided lung biopsy procedures.This paper focuses on the main problem of intraoperative lung deformation caused by respiration.We propose to use the two-dimensional intraoperative X-ray images of the anterior-posterior and lateral sides to register the preoperative three-dimensional CT using Sparse Motion Composition to solve the problem of intraoperative lung deformation of patients,it help doctors to get the three-dimensional intraoperative lung anatomical changes and accurately locate the lesion location with intraoperative threedimensional deformed CT.The use of X-ray images to guide the operation can reduce the radiation dose compared to CT images,reduce time cost of intraoperative imaging time and reduce surgical payment of patients.It can achieve better clinical results and has good application prospects.However,the poor image information of the lung tissue in the X-ray image and the insufficient description of the actual three-dimensional space by the two-dimensional image make it difficult to register the 3D lung deformation.The accuracy of the registration method in this paper accords with the precision requirements of clinical surgery,which is of great significance.The main contents of this paper are as follows:1)In the light of the problem of large amount of radiation,long operation time and complicated operation procedures in lung biopsy,we propose that the registration of X-ray images and preoperative CT images in lung biopsy surgery to guide the program to reduce radiation and reduce operating costs.2)Propose sparse linear combination of intraoperative lung deformation vector fields(DVF)using sparse representation method,and solve the problem of solving sparse coefficient of three-dimensional DVFs to solve the movement of marker point,speed up the calculation speed and ensure the accuracy of the result,CT data to experiment,the verification method is accurate and reliable.3)For the problem of the registration between the two-dimensional intraoperative X-ray and preoperative three-dimensional CT,we project the preoperative CT and generate digital reconstruction radiography(DRR).Then register DRR images and both sides of the X-ray to get twodimensional DVF.Finally,use back-projection algorithm on both sides of the two-dimensional DVF to get the three-dimensional motion information of intra-operative landmarks,which is used for solving the sparse combination coefficients for combining three-dimensional DVF.