Studies Of Three-dimensional Reconstruction Of The Liver And Virtual Liver Surgery Planning For Clinical Application

BackgroundThe liver is one of the major organs vulnerable to malignant tumors both primary and metastatic. During the last two decades possibilities of curative resections of these lesions have improved markedly. Only visualization of the size and location of lesions in parenchymal organs with computed tomography (CT) or magnetic resonance imaging (MRI) makes planning of operations possible. Development of imaging techniques has made the procedures safer and more aggressive resections possible, without knowledge of major blood vessels or other important structures related to the lesion, surgery can not be performed curatively and safely at the same time. However, there are anatomical variants in the vascular or biliary systems that complicate the operation, moreover, new surgical strategies such as anatomical resections of tumors created a demand for more careful preoperative planning of the resection, taking into account the volume of the resected portion of the liver. Although sliced CT and MRI images contain all the information about the tumor(s), major vessels and bile ducts, the surgeon may have difficulties to perceive the relations of these structures to each other during surgery.However, better knowledge of three-dimensional appearances of liver structures may further improve the results of curative liver surgery. Accurate three-dimensional images showing vessels, bile ducts and tumor(s) from different directions on a screen in the operation theatre may help the surgeon further in perceiving the optimal resection line. Tumors close to crucial structures and of higher number could be resected curatively without removing healthy liver tissue unnecessarily. Virtual liver will help surgeons to optimize the planning, to ensure the quality and safety, and to reduce the operation complications.At present, most of the three-dimensional reconstruction of the liver in clinic was accomplished on CT. The multislice helical CT possessed the powerful post-processing function and supplied the detailed image for the clinic with all kinds of the image recombination methods. But the reconstruction must be performed on the expensive CT workstation. Radiologists will not reconstruct the images if surgeons do not request, even if reconstruction program is available in CT devices, radiologists do not know what three-dimensional parameters clinicians need. It can not be satisfied with requirement for virtual liver surgery planning. Possibly the rational way to perfectly use is having three-dimensional liver model reconstructed by hepatobiliary surgeons themselves. Therefore, it is necessary to develop a three-dimensional reconstruction software for liver surgery planning which should be robust, flexible, cheap, based on personal computer and easy to operate, thus allowing its use in daily clinical routine.However, only a few softwares assistant for preoperative planning in liver sugery are developed and have a patent in the world, and only a few centers world-wide deal with image analysis for liver surgery planning. The related studies have not been reported of domestic yet. For this purpose, we developed the software Liv1.0 with independent intellectual property rights for three-dimensional reconstruction of the liver and for virtual liver surgery planning in clinic.PARTⅠStudies of three-dimensional reconstruction of the liver based on CT images on personal computerObjectiveTo investigate the methodology of the three-dimensional reconstruction for the liver and its interior vessel and tumor based on CT images on personal computer, providing three-dimensional image data for liver surgery planning system.MethodsThe region of interest (ROI) was identified and extracted, and the data were segmented semi-automatically, including segmentation of the liver, segmentation of hepatic vessels and segmentation of liver tumors. With generalpurpose segmentation methods, such as thresholding and region growing, manual intervention is usually needed at some stage. The three-dimensional visualized model of liver, hepatic vein, intrahepatic portal vein was reconstructed by volume rendering and surface rendering reconstruction. The measurements of liver volume were obtained automatically. The liver surgery planning system software Liv1.0 was applied to the segmentation, three-dimensional reconstruction and measurement of liver volume on personal computer by the user.Results1.The visualized model could be studied freely from any direction on the computer screen. The appearance fidelity was good. Segmented structures could be visualized with different colors and opacities. The visualized model of different structures could be displayed singly or as a whole. After the liver visualization, the intrahepatic vessels and tumor could be showed in the stereo and distinct. The model displayed the complicated adjacent relationship of the hepatic portal vein, hepatic vein and tumor realistically in three-dimensional space. The virtual resection territory was identified according to the tumor.2.The mean liver volume of patients measured by the software and CT was (1268±307) ml and (1287±297) ml respectively. Comparison by means of linear regression analysis between volume measurement on the software and CT showed a nearly ideal correlation coefficient ( r =0.960, P<0.001) . The mean error of this method was2.2%.ConclusionsThe three-dimensional liver model with intrahepatic vessel, which was three-dimensionally reconstructed in the method of software Liv1.0, always have smooth and glossy surface, clear vessel structure, bright-colored, strongly stereo felling and can provide the image data of the virtual liver surgery. There is no significant difference between the software and CT for the measurement of liver volume, the volumetric estimation could be applied to clinical practice. PARTⅡSkeletonization of the hepatic vessel system and vascular territories automatic segmentation based on liver three-dimensional imagesObjectiveTo analysis vascular structures and to segment vascular territories automatically for virtual liver surgery planning with mathematical methods based on liver three-dimensional images.MethodsThe voxel-based shape representation of the vascular tree was transformed into an abstract graph representation, utilizing skeletonization algorithm, then the hierarchical information was stored in the vascular tree. The nearest neighbor segment approximation(NNSA) method was utilized to segment vascular territories automatically. With these mathematical methods, the portal vein tree was labeled, where segment-feeding branches are marked and liver segments are generated based on the labeled portal vein tree, so the simulation of an anatomical resection can be carried out.Results1.The erosion of the voxels did not change the topology of the original structure. The skeleton line was single voxel wide and smooth, the node points were marked clearly and accurately, the creation of"irrelevant"skeleton lines due to boundary noise of the vessel surface was adjusted and controlled. The resulting skeleton will reliably reflect the original shape of the vessel system. If a node point of the skeleton was selected, all child branches are selected recursively and the vascular territory was annotated by the corresponding color .2. According to the portal segment branches, each individual liver segment can be divided automatically. The liver 3D model was reconstructed based on liver segments, the margin of liver segments was displayed clearly. The anatomical resection simulation and the volumetric estimation can be applied. ConclusionsThe 3D liver model with intrahepatic vessel was reconstructed clearly, the vascular territories segmentation, resection simulation and volumetric estimation can be applied. The planning can be accomplished with software Liv1.0 on personal computer by the user. The software allowed the surgeon to utilize virtual liver techniques for clinical application.PARTⅢStudies of virtual liver surgery planning applied to hepatic resectionObjectiveTo evaluate the impact of preoperative three-dimensional visualization and virtual liver surgery planning on hepatic resection.MethodsReconstruction and image analysis of liver tumors was performed using the research software Liv1.0. Livers, tumors, and the vascular system were extracted from multislice computed tomography scans. After hierarchical analysis of the vascular system, territories supplied or drained by the major vascular branches were calculated. Results were explored and virtual resections of organs were carried out using the research software Liv1.0. Data were correlated to intraoperative findings.Results1. Three-dimensional visualization revealed the spatial relationship of tumors and organs to crucial anatomical structures, thus giving impressions how the neoplasms were situated. Intrhepatic structrues could be identified and quantified more precisely then compared with the original data. Virtual tumor resections corresponded to the intraoperative findings. Virtual simulations of tumor resection were used successfully for planning of surgical procedures in the hepatic tumors.2. With the planning, an intended resection could be performed virtually and optimal identification of resection margins could be achieved. The ischemia and congestion territory within the remaining liver parenchyma can be calculated. Simulation resections could avoid liver parenchyma over resection and maintain a sufficient amount of liver tissue to sustain hepatic function. Virtual simulations of tumor resection were used successfully for planning of surgical procedures in the hepatic tumors.ConclusionsThe technique of three-dimensional tumor visualization and virtual simulation of tumor resections of the software Liv1.0 provides the important reference for a valuable planning of complex hepatic resections.PARTⅣResidual liver volume measured by virtual resection as a predictor of hepatic dysfuntion after hepatic resection for the patients with primary hepatocellular carcinomaObjective Using three-dimensional hepatic volumetry and virtual resection to calculate the residual liver volume(RLV) and standardized estimated liver remnant ratio (STELR).To explore the relationship of STELR to the incidence of severe hepaticdysfunction following hepatic resection in patients with primary hepatocellular carcinoma (PHC).MethodsThe study population consisted of 76 patients with PHC undergoing hepatic resection.The standardized estimated liver remnant ratio (STELR) was determined as the ratio of the residual liver volume (based on CT images and estimated by software Liv1.0 on personal computer) to the standardized total liver volume (STLV), estimated from the body surface area using the equation: liver volume [ml]=706×body surface area [m2]+2.4, body surface area [m2]= 0.0061×body height(cm)+0.0124×body mass (kg)-0.0099. STELR was related to postoperative hepatic dysfunction. Receiver operator characteristic curve analysis was undertaken to determine the critical STELR predicting severe hepatic dysfunction. Univariate analysis and multivariate logistic regression analysis were performed to delineate perioperative predictors of severe hepatic dysfunction.Results1.The mean liver volume of patients measured by the software and the real resected was (489±206) ml and (459±199) ml respectively. Comparison by means of linear regression analysis between volume measurement on the software and the real resected showed a nearly ideal correlation coefficient ( r =0.916, P<0.001) . The mean error was 6.4%.2.The incidence of severe hepatic dysfunction following hepatic resection increased significantly with smaller STELR. A critical STELR of 53% was identified as associated with severe hepatic dysfunction (P<0.001). Additionally, Univariate analysis revealed that operating time, and intraoperative blood loss were significant prognostic indicators for severe hepatic dysfunction.ConclusionsThe likelihood of severe hepatic dysfunction following hepatic resection can be predicted by a small STELR. Shorten the operation time and control intraoperative blood loss can decreace the risk of operation. Residual liver volume measured by virtual resection could be a predictor of hepatic dysfuntion after hepatic resection.