Accurate Localization And Intervention Methods In Ablation Of Liver Cancer Using Optical Surgical Navigation

Minimally invasive ablation for liver cancer is one of the most effective methods to treat liver cancer,and the accurate localization and puncture of tumor targets is crucial important for the therapeutic effect of the liver ablation.Surgical navigation system had obtained remarkable achievements in the surgical fields such as neurosurgery and orthopaedics.However,the application of surgical navigation system in liver ablation has restricted by the large deformation of liver in abdomen caused by the factors such as breathing and heartbeat.Currently,the localization and puncture of tumor targets in liver ablation mainly depends on the clinical experience of the surgeons and the guidance of medical imaging,which made the procedure complex and caused the problems of low puncture accuracy,low operation efficiency,and much operation wound for patients.This paper designed an optical surgical navigation system for liver ablation.We have studied the real-time automatic registration technology and the accurate localization and puncture technique for liver ablation in the surgical navigation system using the technology of binocular vision positioning and the technology of medical image processing and visualization,the major research contents are summarized as follows:(1)We constructed the optical surgical navigation system for liver ablation.To solve the problem of low puncture accuracy and low therapeutic effect in liver ablation using the surgical navigation system,which was mainly caused by respiratory movement,we constructed the optical surgical navigation system for liver ablation,and designed auxiliary tools such as the near-infrared calibration board,calibration block for surgical instruments,tracking clamp for surgical instruments,custom fiducial markers,and surgical instruments.Then,we proposed the calibration method based on the 3D space error for the near-infrared optical tracking system,high accurate calibration method for surgical instruments,automatic identification method based on model matching for fiducial markers,stereo matching method based on projection error for fiducial markers in the left and right images,automatic space registration method based on geometric feature clustering,accurate puncture method based on real-time automatic registration method in the liver ablation intervention.(2)We studied the calibration method for the near-infrared optical tracking system and surgical instruments.On the basis of the previous studies,this paper proposed a calibration method based on 3D space error for the near-infrared optical tracking system and a high accurate calibration method for surgical instruments.In previous studies we made a near-infrared calibration board and assumed it as a 2D graphic template.In this paper,the calibration method was not based on the hypothesis of 2D graphic template that the calibration board was treated as a 3D template and 3D coordinate reconstruction error term of the feature points was added to the cost function,which was solved to find the optimal solution.The proposed method obtained higher calibration accuracy,and the accuracy of calibration results was higher than that of the previous studies.We analyzed the defects of current calibration methods for surgical instruments and designed a simple calibration block and the corresponding calibration method achieved high accurate for surgical instruments.The calibration error of the surgical instrument’s tip could achieved less than 0.7 mm used this calibration method.A more general calibration tool for surgical instruments was designed in this paper,which had many advantages such as convenient using,strong applicability,and high calibration accuracy.(3)We studied the automatic registration method for the surgical navigation system.We designed a custom fiducial markers and realized the real-time automatic registration method based on custom fiducial markers in this paper.There were two kinds of custom fiducial marker which were applied to the equipment of CT or MR.The custom fiducial marker had a unique structure which made it easily separated from the human anatomy in the image space and recognized by the optical surgery tracking system in the operation space.The real-time automatic registration method was implemented as three steps: The first step was to identify the fiducial markers in the image space that we proposed an automatic identification method based on model matching to automatically recognize the fiducial markers in the image space.Comparing with the manual identification method,this method had higher localization accuracy and robustness;The second step was to track the fiducial markers in the operation space in which stereo matching of left and right images was the key for the tracking accuracy and real-time performance.This paper presented a stereo matching algorithm based on projection error which combined geometric principle and projection error to solve the stereo matching problem effectively.This stereo matching algorithm was robust and could ensure real-time tracking for custom fiducial markers;The third step was to register the operation space to the image space.This paper proposed an automatic space registration method based on geometric feature clustering that the geometric distances between each fiducial markers in two spaces were calculated to get the matching results of the fiducial markers fast in the two spaces,and then the transformation relationship between the two spaces was solving.To verify the feasibility of real-time automatic registration method,we implemented a series of phantom experiments and the experimental results showed that the method could realize rapid,automatic,accurate,and real-time registration.(4)We studied the accurate puncture method for liver ablation.Based on the real-time automatic registration method we proposed an accurate puncture method for liver ablation.This method could be summarized as: During the operation,the fiducial markers pasted in patient’s abdomen was tracking in real time,and the real-time automatic registration method was used to calculate the curve of the registration error(FRE)of the fiducial markers in real time.Then the surgeon found out the best matching time with the preoperative image according to the curve of FRE in real time,and determined the optimal time for puncture.We had implemented a series of phantom model,in vitro,and in vivo experiments.The experiment results had verified that this method had higher accuracy than other methods which had small wound to the patient,remarkable reductive radiation damage to the patient,and clinical feasibility.