Pancreatic Endoscopic Ultrasound Image Based Computer Aided Diagnosis And Treatment Methods In The Study

Pancreatic Cancer is one of the top ten cancers declared by WHO, which greatly threatens people’s life. Computer aided diagnosis (CAD) system based on the endoscopic ultrasonography (EUS) images is an objective and harmless way to help to diagnose pancreatic cancer, which is valuable and meaningful. But the existing CAD system can only classify cancer and non-cancer cases, which cannot deal with the pancreatitis cases. On the other hand, implanting radioactive particles (125I) into the tumors is an effective palliative treatment for the terminal cancer. And it will do a lot of help for operating this kind of surgery if the spatial radioactive dose under the current implanting plan can be provided.In this paper, the CAD system is expanded by adding the pancreatitis category into the classification system. And then, a kind of software to simulate the implantation of125I radioactive particles is developed using C++code. Thus, a new computer aided system of diagnosing pancreatic disease and generating the implantation plan of125I radioactive particles is proposed, which contains these four major features:First is that we can use this system to classify the given EUS image as pancreatic cancer, normal pancreas, chronic pancreatitis (CP) or auto immune pancreatitis (AIP). For example, we can classify the non-cancer cases got from the previous CAD system as normal pancreas or pancreatitis and if the result is pancreatitis, a further step to classify it as AIP or CP can be taken.Second is that we can use this system to do fuzzy classification for the given EUS image which has gone through the therapy of radioactive particles implantation, and then get its probability of cancer. In addition, an aided comment on the effect of the implantation treatment can be made. Before doing the classification, a method of2-level neighborhood median filtering will be applied to remove the light spot caused by the implanted radioactive particles.Third is that we can use this system to simulate the implantation of125I radioactive particles on a EUS image or in a spatial model. And then the spatial radioactive dose under the current simulating implantation plan can be got, which will help the doctors to the real implanting operation.Forth is that we can use this system to get the high risk region of pancreatic cancer in the region of interest of a given pancreatic cancer EUS image through the blocking fuzzy classification of pancreatic cancer and non cancer. And then the system can automatically generate the simulating implantation plan of125I radioactive particles according to this high risk region of pancreatic cancer. That will provide a reference for the doctors to do the real implanting operation.