Clinical Application Of Three-dimensional Technique In The Diagnosis And Surgical Treatment Of Hepatoma

BackgroundHepatoma is a frequent hepatic malignancy in China, and surgical intervention currently remains the primary choice of treatment. With the development of such imaging modalities as computed tomography (CT) and magnetic resonance imaging (MRI), many difficult and complicated surgical procedures become clinically possible for hepatoma treatment. Nevertheless, the conventional two-dimensional CT and MRI images can not provide sufficient information on the intricate anatomy of the liver and the intrahepatic vascular structure. The current MRI and multislice spiral CT equipment has powerful post-processing functions, which, however, have to be finished in expensive MRI or CT workstation, thus limiting its wide clinical application. In addition, the reconstruction result often fails to meet the requirements of surgeons, and the model reconstructed does not allow color-coding, zooming, or transparent rendering of the region of the interest (ROI), hence observation of the model from multiple angles is impossible, which is essential for virtual simulation of the surgical procedures and preoperative optimization of the surgical plan. In spite of the great advancement of the surgical techniques in recent years that enormously improved the success rate of hepatoma resection, the complexity of the hepatic anatomy, the frequent variations of the intrahepatic ductal system, and the peculiarities of the hepatic physiology all poses much clinical difficulty and great risks in liver surgery. In this context, a full understanding of the complex anatomy of the liver and the tumor-associated structures in an individual patient, especially the volume, shape, and spatial relations of the tumor as well as the caliber, course, branching, and anatomical variation the hepatic vessels and their adjacent relations with the tumor has been of great clinical significance in the precise preoperative diagnosis and clinical decision-making of the surgical plan. Also of great clinical value is the preoperative measurement of the residual liver volume for a comprehensive evaluation of the surgical risks of hepatolithectomy combined with patients’ other clinical data, so as to reduce the incidence of serious complications such as intraoperative bleeding, bile leakage, liver dysfunction and failure. Preoperative practice of the surgical procedures in the reconstructed hepatic model allows the surgeons to optimize the surgical plan, be familiar with the surgical procedure, and improve the surgical skills, thus improving the success rate of surgery. All these are the current hot research topics in hepatic surgery.In this study, we developed an abdominal medical image three-dimensional visualization system (MI-3DVS) based on the preoperative submillimeter64-slice spiral CT data from40patients with hepatoma. The reconstructed three-dimensional (3D) model clearly visualized the complex anatomy of the liver, the intrahepatic ductal system and the relationship between hepatoma and hepatic vessels, thereby providing valuable assistance in making the surgical plan and calculating the residual liver volume. Combined with patients’ other clinical data, this model allows continuous optimization of the surgical plan, virtual simulation of the surgery, accurate preoperative diagnosis, and precise surgical operation to ensure the precision in hepatoma resection and smooth and fast postoperative recovery.Objective1To study the clinical application of3D reconstruction based on the original submillimeter64-slice spiral CT data of hepatoma patients in the diagnoses and surgical treatment of hepatoma. 2To assess the value of residual liver volume measurement in the diagnoses and surgical treatment of hepatoma.3To study the clinical application of virtual simulation surgery in the diagnoses and surgical treatment of hepatoma.Methods1. Equipment (1)64-slice spiral CT-PHILIPS Brilliance64(Netherlands PHILIPS company) and image post-processing workstation Mxview;(2) HP blade servers;(3) abdominal three-dimensional medical image visualization system (MI-3DVS);(4) DICOM CT image viewer;(4) ACDSee10.0image processing and conversion software. FreeForm Modeling System (American SensAble Technologies company;(7) Force feedback equipment PHANToM (PHANToM Desktop).2. Study subjects Forty patients with hepatoma meeting the inclusion criteria of the study admitted in our hospital between January,2010and December,2011were enrolled, including31male and9female patients with a mean age of49.4±14.7years. The original preoperative submillimeter64-slice spiral CT data were collected. The patients with extra-hepatic metastases, liver function of Child grade C, poor general condition for hepatectomy, or the inoperable patients were excluded. All the cases were confirmed by pathological examination to have primary liver cancer.3Acquisition of the original CT data (1) Scanning parameters:Using abdominal CTA, the scan tube voltage was set at120kV, tube current at300mAs,0.5s per lap, pitch at0.984, with a5mm slice thickness;(2) Scanning was divided into four phases:precontrast, arterial, portal and delay period. After scanning, the original5-mm-thick images obtained from the4phases were cut into1-mm-thick images. Before the scanning, the patients were fasted for food and water for more than4hours, and at20min before scanning, the patients were asked to drink300ml water for filling the gastrointestines. A red intravenous catheter needle tube was retained in the right elbow vein for the injection of the contrast agent (Ultravist370). In enhanced scanning, the contrast agent was injected at the rate of5ml/s with a scanning time of5s; Arterial phase scan was trigged automatically according to the peak of aortic agent contrast, generally at20-25s after the beginning of contrast agent injection. The portal phase scan was initiated generally at50-55s, followed immediately by the delayed phase scan.(4) Data transfer and storage:In Mxview image post-processing workstation, the original CT data were transmitted through a special-purpose communication link to HP blade servers at the Digital Medical Research Center, exported and saved.43D model reconstruction The image segmentation method employed in this study combined threshold segmentation method and regional growth method, and also combined3D and2D segmentation methods. DICOM viewer and ACDSee10.0software were used for format conversion and adjustment of the original CT data, after which the images were imported into MI-3DVS for fast image segmentation and3D reconstruction. The3D models of the liver, tumor tissue, hepatic arteries, portal system, and the inferior vena cava system were reconstructed separately, and each3D model was registered automatically.5Principles of individualized liver segmentation Differing from Couinaud liver segmentation, our approach to liver segmentation was based on the course and anatomical variation of the first and secondary branches of the portal vein and hepatic vein to ensure that the residual liver segment had a independent vascular system. This approach of individualized liver segmentation facilitated accurate localization of the tumor foci and subsequent surgical planning.6Surgical planning (1) Accurate localization of the hepatoma:Using a semi-manual marking point segmentation method and with the first and secondary branches of the intrahepatic vessels (portal vein and hepatic vein) as references, the liver was segmented according to the Couinaud segmentation principles, followed by transparent rendering of the liver. The liver segment that contained the hepatoma was hidden to allow independent observation of the distance between the tumor and the adjacent structures and their anatomical relationship to guide the establishment of the surgical plan;(2) Measurement of residual liver volume percentage:By zooming, transparent rendering and observation from different angles, the ROI in the3D model was thoroughly inspected to examine the spatial distribution, shape and blood supply of the tumor, the course, potential anatomical variation and branching of the intrahepatic ducts, and their spatial relationship with the tumor. The preliminary surgical plan was made and simulation surgery was carried out using MI-3DVS. Using the volume measurement tool, the residual liver volume, functioning tissue volume and liver cancer volume were measured. The residual liver volume percentage=(residual liver volume/functioning liver tissue volume)×100%. The hepatoma volume percentage=(hepatoma volume/function liver tissue volume)×100%. If the liver volume percentage can guarantee the postoperative compensatory proliferation of the liver, the actual operation was then performed according to the formulated surgical plan, otherwise the plan was modified or conservative treatment was administered.7Simulation surgery MI-3DVS operation simulation software was used to implement the virtual hepatectomy. By observing the anatomical relationship between the tumor foci and the vascular system and the spatial distribution, shape and size of the tumor visualized by the reconstructed models, the surgical plan was made. The simulation software allowed free three-dimensional control of the size and spatial position of the reconstructed models. With the simulation surgical instruments we developed previously, the entire surgical procedure of hepatectomy including ligation, suture and cutting was simulated.8Statistical methods This study involves only the average and standard deviation of the functioning liver tissue volume, tumor volume, resected functioning liver tissue volume, and residual liver volume.Results1Two-dimensional CT dataThe original two-dimensional enhanced CT data of the40patients clearly displayed the liver, tumor foci, hepatic artery, portal vein system, hepatic vein system and the inferior vena cava system. The effect of enhanced scanning was satisfactory, and the secondary intrahepatic branches of the hepatic artery, secondary intrahepatic branches of the portal vein, tertiary branches of the hepatic and portal veins were all clearly shown. The blood supply of the tumor and the relationship between the lesions and the hepatic vascular system were clearly displayed.23D reconstruction of the liver and intrahepatic and extrahepatic vascular systemThe reconstructed3D models clear displayed the liver, hepatoma and anatomy of the intrahepatic and extrahepatic vascular system. The presence of liver hyperplasia and hypertrophy, and the spatial distribution and blood supply of the hepatoma were all clearly observed. The models allowed zooming, transparent rendering and rotation of the ROI for observation of the spatial distribution, anatomical variation, and course of the hepatic arteries, veins, and the portal vein as well as the relationship between the lesion and hepatic vascular system from different angles.3Individualized liver segmentationIn individualized liver segmentation,3cases had no definite classification of the left lobe, in which the hepatoma invaded the root of the hepatic vein so that the left lobe did not contain a ductal system as the basis for segmentation. Four patients had two branches of the middle hepatic vein in the liver segment, and the classification principle was the same with the Couinaud segmentation, but the separation of Ⅳand Ⅷ, or Ⅳ and Ⅴ liver segment was in the left branch of the middle hepatic vein. Four patients had inferior right hepatic vein in the liver segment, and in this type, the right hepatic and the inferior hepatic veins divided the right liver into the upper, middle, and lower sectors, and the right branch of the portal vein further divided each sector into two segments. The segments was named, in a clockwise sequence, as the V-X segments. The remaining29patients were all conventional type and had the same liver segmentation as Couinaud segmentation.4Rsidual liver volume percent and tumor volume percent calculationThe40patients had a mean functioning liver volume of1287.31±263.32ml, a mean hepatoma volume of235.73±224.53ml, a mean resected functioning liver volume of432.57±234.21ml, and a mean residual liver volume of854.73±241.70ml. The tumor volume percentage was<10%in18cases, between10%and20%in9cases, between20%and30%in7cases, between30%and40%in3cases, and over 40%in3cases. Evaluation of the residual liver volume percentage measured preoperatively in combination with the patients’ clinical data for surgical risk assessment showed that the preoperatively calculated residual liver volume percentage met the requirements for postoperative liver compensatory regeneration in all the40cases, and the evaluation results was consistent the actual risk. The operation was implemented according to the preoperative surgical plan, and all the patients showed smooth postoperative recovery without such serious complications liver dysfunction or failure.5Simulation surgery and actual surgeryThe simulation operation had a vivid "force" feedback with a mechanical feeling and realistic visions. Repeated practice of the preliminary surgical plans using different methods and approaches provided valuable assistance in optimizing the surgical plan and successfully guided the actual hepatoma resection with precision in the40cases.6PrognosisAll the40patients had smooth postoperative recovery. Postoperative pathological examination confirmed hepatoma in all the cases. The mean hospital stay of the patients was11days. No tumor recurrence was found in these cases one month after the operation, and no deaths occurred during hospitalization. After discharge, the patients were asked to return to the hospital every month for follow-up examination. The3-month follow-up of the patients showed obviously improved quality of life and good tumor-bearing or tumor-free survival. After the operation, all the patients showed good compensatory function of the residual liver, and none of them were found to have serious complications such as liver dysfunction or failure. These results confirmed the reliability of the preoperative surgical safety and risk assessment using MI-3DVS.Conclusion13D reconstruction models based on the submillimeter64-slice spiral CT can clearly and vividly display the complex anatomies of the liver, hepatoma and the intrahepatic and extrahepatic vascular system to help in accurate preoperative diagnosis.2. Individualized liver segmentation well suits the anatomical characteristics of liver cancer, facilitates precise hepatoma localization, and helps in the formulation of the surgical plan.3. Preoperative calculation of the residual liver volume percentage helps in the assessment of the feasibility and risks of the surgery and in the formulation of the surgical plan to reduce the surgical risks and complications.4. Preoperative simulation surgery can optimize the surgical plan, help the surgeons be familiar with the surgical procedures, facilitate the intraoperative decision, reduce intraoperative bleeding and surgical trauma, and ensure accurate intraoperative operation.