Research On Key Technologies In Computer-Assisted Navigation System For Thermal Ablation Of Hepatic Tumors

As a minimally invasive treatment method,tumor thermal ablation has been identified as one of the conventional methods for the treatment of liver cancer.The principle of thermal ablation involves inserting the ablation needle into the tumor as the heating source under the guidance of medical image.Heating the tumor to a high temperature will lead to complete coagulative necrosis of tumor tissues.Howerver,there still have many problems to be solved for thermal ablation therapy.It is of great significance to study the objective,scientific and quantitative computer-assisted thermal ablation therapy.The dissertation focused on key technologies in computer-assisted navigation system for thermal ablation of hepatic tumors.The main research contents are as follows.(1)Research on automatic liver segmentation method in computed tomography(CT)images.A new method was proposed for fast automatic liver segmentation using the simple linear iterative clustering(SLIC)and graph cuts algorithms.The patient-specific liver volume of interest(VOI)extraction procedure was performed firstly using maximum intensity projection(MIP)and a histogram-based adaptive thresholding methods.Then,SLIC supervoxels were gererated for the liver VOI.For graph cuts segmentation,graph was constructed over the supervoxels,liver and background seed points were selected automatically,and gaussian mixture models(GMMs)were used to estimate the intensity probabilistic models.The experimental results indicate that the proposed method may be useful for efficient liver segmentation.(2)Research on automatic liver segmentation method in CT images.A new method was proposed for fast automatic liver segmentation using the kernelized fuzzy c-meas with spatial constraints and graph cuts algorithms.With a single seed point,the tumor VOI was extracted using confidence connected region growing algorithm to reduce computational cost.Then,initial fore/background regions were labeled automatically,and a kernelized fuzzy C-means(FCM)with spatial information was incorporated in graph cuts segmentation to increase segmentation accuracy.The experimental results showed that the proposed method was accurate for 3D liver tumor segmentation with a reduction of processing time.(3)Research on automatic liver vessel segmentation method in CT images.A new method was proposed for fast automatic liver vessel segmentation using the Hessian based vesselness enhancement filtering and graph cuts algorithms.With the input liver mask,candidate liver vessel regions were extracted firstly using non-linear intensity conversion and adaptive thresholding algorithms.A vesselness filtering algorithm based on Hessian matrix was incorporated when estimating the probabilistic models of tumors and liver tissues.For graph cuts segmentation,initial vessel and background regions were labeled automatically,gradient information and vessel shape constraints was incorporated into the energy function to increase segmentation accuracy.The experimental results indicate that the proposed method may be useful for efficient liver vessel segmentation.(4)Development of computer aided treatment planning system for liver tumor thermal ablation.Based on the proposed liver,tumor and vessel segmentation methods,the 3D reconstruction and visualization of patient's structural models in abdominal region were implemented,with the used of open source image processing toolkits(ITK/VTK/OpenCV/MITK).The intervention trajectory(puncture path)planning of the ablation needle was realized through human-computer interaction.The system can effectively improve the efficiency of thermal ablation needle puncture path planning.