The Establishment Of A Three Dimensional Visualization Platform Fortomographic Image Data Of Abdominal Surgery

Background:During the preoperative planning of abdominal malignant tumors, surgicaltherapy was constantly abandoned due to the invasion or encasement of the majorvascular structures by primary lesion or enlarged lymph nodes; or, the tumor wasfound unresectable after risky explorative laparoscopy, confronting the patientswith injury instead of benefit.The present contrasted multi-slice computedtomographyscanning can provide axil thin layer two dimensional images of theabdomen with high resolution.The medical radiation technologistcan also offer3Dreconstructionservice if needed, but this program currently require specificapplication from the surgeons, and may not meet the clinical needs, either in timeconsumption, work efficiency or quality of images.Anyway, the surgeons are stillconfronted with two dimensional data without free manipulation and observationor qualitative and quantitative analysis at any level directly on the3D imagesaccording to their professional knowledge anytime, which is no good to theaccurate preoperative planning. Moreover the present process of VR（volumerendering）is based on the gray value of the originalvoxel under specific windowwidth and window center without any segmentation, which makes it difficult todisplay the comprehensive information of the operating zone. Even theinformation provided on2D images may get lost due to the mutual screening ofdifferent parts.Magneticresonance imaging of abdomen is able to depict softtissues like major vesselsand tumors and provide valuable information of lymphnode metastasis in in sequences like DWI (diffusion weighted imaging). Howeverits resolution, number of images in single sequence and time consuming is far from satisfactory. There are a few softwares for surgeons of hepatic surgerycovering a narrow range of application without being able to take a preciseanalysis for organs and structures all over the abdomen. A fewstudies on thecharacters of peripancreatic vessels invasion has been held, but there is no similarstudy on the involved blood vesselsof primary retroperitoneal tumor and advancedgastric cancer.Purpose:The purpose is to establish a software platform on which clinical data oftwo-dimensional images can go over direct decompression, segmentation and3Dreconstruction.3D visualizedanalysisfor major involved blood vesselsof primaryretroperitoneal tumor and advanced gastrointestinal cancer can be made with theutilization ofthe platform independently by surgeons in accordance with therequirements of surgery.To explore the application value of this method in thepreoperative planningof malignant abdominal tumor and lay the foundation for thedigital solution ofthe subsequent development of surgeonsdedicated accuratepreoperative evaluating systemfor abdominal tumors.Methods:1. Propose the clinical demand of surgeon dedicated3D visualization system ofpreoperative evaluationby in-depth exchange with software engineers.2. Confirm the compressionformat and algorithm of the available data, andthendecompress the DICOM data with the general code base.3. Make automatic segmentationfor the abdominal arteries using differentalgorithms and select the optimal one, make manual modificationof thesegmentation resultsimultaneously.4. The2Dimage segmentation results was processed with appropriatevisualization algorithm,the abdominal aorta three-dimensional visual imageswere then abtained.5. Make direct measurements in the3D visualization imageof abdominal aorta with VTK-Widgets method, the diameter, origin, angle and other quantitativedataof the related artery were then obtained.6. Collection of the DICOM data of primary retroperitoneal tumors andadvanced gastric cancer patients underwent preoperative abdominal thin layerCT scan, carry out the segmentation,3D visualization and measurement andmarkingaccording to the above process, and feedback to the engineer forimprovement.7. By comparing with the traditional image data, the accuracy of the marking ofinvolved arteries and the value of preoperative planning were evaluated.Results:1. Surgeons can make decompression, automatic segmentation,3D visualization,vessel marking and some simple measurementswith the use of this platformof3D visualization at any time.2. The segmentation results of major vessels were not ideal using automaticsegmentation algorithm alone.It needsto be combined with manualsegmentation for more accuracy.However to a large extent,it depends on thephysician’s ability to read, but still, requires a considerable amount of manualwork.3. Although the3D visualization platform established in this study presents theinvolved areas of major vessels related to abdominal malignant tumors in3Dspace, venous system is still not obvious due to thelow enhancement.Conclusion:The software platform establishedin this study can preliminary meet thesurgeons’ demandof observing the abdominal malignant tumor patients’largeblood vessels (arteries)at any time from different anglesin3D space,andquantitative/qualitative analysis based on the professional knowledge of bloodvessels involvedat any level.