Three-dimensional Reconstruction Of Individual Hepatic And Portal Veins System In Liver Surgery

BackgroundMajority of liver tumors are malignant.Primary liver cancer, which consists predominantly of hepatocellular carcinoma (HCC), is the fifth most common cancer worldwide and the third most common cause of cancer mortality. Surgical resection is the most effective treatment at present. However, most patients worldwide with HCC have underlying cirrhosis, the poor functional reserve and poor regeneration capacity of liver in patients with cirrhosis will greatly increase the risk of liver failure after hepatectomy. So, how to preserve maximally the postoperative residual functional hepatic volumes becomes the current focus of attention.The distribution of the hepatic and portal veins is an important factor for the choice of surgical plan but their route and distribution have markedly individual variation. Inadequate drainage of veins plays an important role for liver dysfunction. In addition, HCC has a propensity to disseminate into the portal venous territory, which is the most commonly reported risk factor for intrahepatic recurrence. Therefore, accurate assessment of the preoperative route and distribution of venous system is closely related to retain complete inflow and outflow tract or remove the involved portal venous system completely, and it is also the premise of maximum retention of postoperative functional liver volume.CT, MRI and other imaging examination can provide useful information regarding tumor infiltration and adjacent main hepatic vasculature. Recently it has been reported that to select the appropriate surgical method based on CT avoided the occurrence of postoperative hepatic congestion. In addition, there is also a study preserved maximally the postoperative residual functional hepatic volumes by observing intraoperative ultrasound about the distribution of hepatic vein. As mentioned above, these studies provide an effective way to reduce the occurrence of complications and liver failure after hepatectomy. However the continuity and the intuitively of the organ still can’t displayed properly. Nowadays, increasingly higher request of surgical visualization is needed for the surgeons. Under the support of "863" project, we use medical image three-dimentional visualization system (MI-3DVS) which was researched and developed independently by our research center to perform three-dimentional reconstruction, the3D models can display the route, distribution and its relationship with lesions of individual veins system spatially and intuitively in virtual visible situation, according to the route of the portal vein and hepatic vein we can perform individual segmentation to locate the tumor, then those3D models were simulated slicing while the hepatic volumes were measured by the Freeform modeling system, appropriate transection plane would be determined by the surgeons,which proposed a feasible and safe method for the selection of minor or major hepatectomy.This study had its values clinically for diagnosis improvement of hepatic carcinomas, direction for actual manipulations of hepatic surgery, conservations of more normal hepatic tissues and reduction of postoperative complications.Objective1. To study the characteristics and variations of individual digitized hepatic vein and portal vein imaging by MI-3DVS reconstruction.2. To study the effects of variation of hepatic vein and portal vein on liver segmentation.3. A comparative study of clinical efficacy of surgical scheme guided by CT image versus three-dimensional images.4. To study the mutual instruction between real surgery and simulation surgery.Methods1.Patients and materialsThis retrospective study included81cases of hepatectomy with the assist of MI-3DVS from2008.06to2012.09in our department.The male/female ratio was of61/20, and the median patient age was of46years (range,12-81years). The tumors included hepatocellular carcinoma (n=66), intrahepatic cholangiocarcinoma (n=3), hepatocellular adenoma (n=1), hepatic focal nodular hyperplasia (n=4) and hepatic cavernous hemangioma (n=7).41cases of the patients were complicated with cirrhosis. Liver function of all the patients were of Child-Pugh A class. Patients with malignant tumor had no hilar or distant nodal metastases, no invasion of adjacent organs, no distant organ metastasis.2.Equipments and Materials①PHILIPS BRILLIANCE64-slice helical CT and image post-processing workstation Mxview;②binocular tube high pressure injector and contrast agent;③HP blade servers;④computer;⑤DICOM CT image Viewer;⑥ACDSee10.0image processing and conversion software;⑦Medical Image3D Visualization System (MI-3DVS);⑧FreeForm Modeling System and PHANTOM.3.The instruments and parameters of CT sets, the segmentation and the three-dimensional reconstruction of liver, tumor and liver vessels are the same with the literature (Fang Chihua, et al. Study on the application of value of digital medical technology in the operation on primary liver cancer, Chin J Surg, April1009,47(7):513-516).4.Individual segmentationIn accordance to the characteristics of hepatic and portal veins as well as ’the presence of self-sustaining blood supply and blood outflows’, in the following, the livers were individually segmented into8segments. If there was a inferior right hepatic vein, and the inferior right hepatic vein beyond the right portal vein plane, two additional segments of the right lobe were divided by the inferior right hepatic vein, the liver was segmented into10segments; if the inferior right hepatic vein was under the right portal vein plane, one additional segment of the right lobe was divided by the inferior right hepatic vein, the liver was segmented into9segments; If there was additional seg.IV vein,the seg.IV vein divided the left medial segment into two segments, the liver was divided into9segments; if additional seg.IV and inferior right hepatic vein both appeared, the liver was divided into10or11segments. Color the each segment for distinction,.5.Simulated resection and actual surgeryOmnidirectional rotation of the3D model to observe the distribution of intrahepatic venous system as well as the relationship between the tumor and the veins. Simulated resection of the3D model with the assist of free-form modeling system. Major hepatectomy was defined as resection of3or more than3segments whereas minor hepatectomy was the resection of2or less than2segments. For patients estimated could perform minor hepatectomy, draw the pre-removal line follow the segmentation result to perform anatomical liver resection or non anatomical liver resection, for the resected liver volume is small, we do not need to calculate the residual liver volume. For major hepatectomy, total and residual functional hepatic volumes were measured before and after simulated major hepatectomy, and the residual ratio was calculated (residual functional hepatic volume/total functional hepatic volume). If the residual ratio is less than50%after virtual anatomical liver resection, non anatomical liver resection based on the spatial relationship of tumors and vessels should be performed instead of following the rule of2cm away from tumor boundary as the resection margin, to reserve more of the normal liver tissue and ensure the intraoperative safety under the condition of reducing the tumor burden so that every entire segment or part of segment after virtual operation has its own independent blood supply and drainage to avoid ischemia and congestion of liver. If the residual ratio is more than50%, the operation would be performed in accordance with this plan; otherwise portal vein ligation or TACE will be performed. The surgical scheme by MI-3DVS, compared with the real intraoperative anotomical situation, can guide the real opertion. By comparing the three-dimensional models with the anatomy of abdomonal organs, and comparing the virtual operation with the real surgical procedures, we assess the value of three-dimensional technique in hepatectomy.6.Follow-up and statisticsAll patients were regularly followed-up at an outpatient clinic and monitored for recurrence by serum alpha-feto protein (AFP) and liver ultrasound every3months. A suspected recurrence was confirmed by dynamic CT scan and hepatic angiography. The minimum follow-up period for all patients was3months. The survival time was calculated from the day of operation until the last follow-up or until December25,2012. The survival rate was analyzed by Life Table, with SPSS13.0, one statistical software. The central tendency of datas, such as the intraoperative bleeding and blood transfusion, were described by Median.Result1.Varitions of hepatic vein and portal vein 3D models of hepatic vein and portal vein are solid, distinct and vivid. In the81cases, the majority of left and middle hepatic shared the same trunks (50/81,61.7%) and for the rest cases, the left, middle and right hepatic veins respectively converged into inferior caval veins. The right hepatic veins in24cases (29.6%) were viewed to be small and short, with outflows from other surrounding hepatic veins,10of which (12.3%), for example, manifested with greater posterior hepatic veins,8(9.9%) with right accessory hepatic veins and6(7.4%) with compensatory outflows of right liver blood from middle hepatic veins.34of them (41.9%) presented section VI hepatic veins:The segment VI veins in18of them (22.2%) ran into left hepatic veins, those of12(14.8%) into middle hepatic veins and those of4(4.9%) directly into inferior vena cava. The portal vein of81cases patients can be divided into4types:Ⅰ type is the common type, there were64cases, the portal vein was seprated with the left branch and the right branch in the hilar.; type Ⅱ, there were10cases, The portal trunk was divided into the left branch of portal vein (LPV), the right anterior portal branches, the right posterior portal branches trifurcated; type Ⅲ, there were6cases, The branch of right anterior portal vein start from the left trunk, the branch of right posterior portal vein start from the main trunk independently;Type Ⅳ, there were1case, lack of left branch of portal vein.2. Individual digital liver segmentationThe3D models of liver, hepatic vein and portal vein were displayed to perform individual segmentation according to the anatomical features of hepatic and portal veins as well as the specific changes from the space-occupying lesions. From the models, the liver was segmented into8segments in42cases, which accorded to Couinaud hepatic segmentation; due to presences of section Ⅳ veins and posterior right hepatic veins, the liver was segmented into11segments in1cases and into10segments in6cases and into9segments in31cases; due to atrophy or nonvisulization of portal veins from intrahepatic space-occupying and crushing, into7segments in1case.3.Surgical planView the relationship between the tumor and the target structures (the liver, hepatic vein, portal vein) through the3D models.The3D model clearly revealed the size and location of the tumor as well as the relationship between the tumor and peripheral vessels. Individual segmentation provided spatial message for accurate hepatectomy.With the assist of steroscopically and directly3D models,5patients assessed need perform major hepatectomy by CT were instead of mionr hepatectomy by MI-3DVS.57patients received minor hepatectomy, draw the pre-removal line follow the segmentation result to perform anatomical liver resection or non anatomical liver resection with the assist of MI-3DVS.24patients received major hepatectomy,10patients performed anatomical hemihepatectomy,1patient who assessed need perform right hemihepatectomy performed non anatomical right hemihepatectomy and caudate lobectomy and one patient performed non anatomical hemihepatectomy, volume calculation showed that the median residual ratio was52%(range,40%-82%);12patients performed non anatomical relatively-reduced major hepatectomy, volume calculation showed that the median residual ratio was55%(range,44%-79%). The residual ratio of7patients were under50%, all of which were with normal liver, estimated that the surgical procedure would be safe.4. Surgical plans based on3D models versus that based on CT imagesA comparative study of liver resection range and operation mode in surgical scheme versus the actual operation was undertaken. The coincidence rate between the surgical planning based on3D images and the real operation process is98.7%(80/81); while that between the surgical planning based on CT images and the real operation process is82.7%(67/81). These two results were compared by Ki-square test gaining p<0.05(two-sided test), suggesting the surgical planning based on3D images is better than that based on CT images significantly.5.ReaI surgery and follow-up90-day mortality rate was0.5patients with malignant tumor were lost to follow up. Of all the patients with malignant tumor,1-year survival rate was83%in which performed minor hepatectomy,77%in relatively-reduced major hepatectomy and75in ordinary major hepatectomy. Under real-time guidance of MI-3DVS, surgeons can easily identify the veins of target area, depending mainly on size and location. Veins, in the area where they are simulated cut, are the structures used for tracking resection progress in actual surgery. Actual surgery was performed according to the simulated resection. Surgeons were able to better anticipate and control the next important veins, in addition, assurance of a good safety margin around tumours was increased. A free resection margin was obtained in all patients. Postoperative morbidity was observed in12(14.8%) cases, all of which had pleural effusion requiring a tap. No hepatic congestion was found in postoperative CT review, and no liver failure occurred after hepatectomy.Conclusions1. Reconstructed models of hepatic veins and portal veins by MI-3DVS based on64-slice helical CT hepatic vein scanning data are clear, solid and vivid. We can estimate the variations of hepatic veins and portal veins preoperation.It could provide a good platform for preoperative planning, surgical rehearsal, assessment of surgical risk, selection of surgical approaches and clinical teaching.2. The application of3D technology can avoid some risks of blood circulation disorder in residual liver tissue resulting from two-dimensional CT-aided operation design, The three-dimensional models reconstructed by MI-3DVS can accurately directly present internal hepatic lesions as well as the adjacency of normal liver tissues, hepatic venins and portal veins. Individual segmentation provided spatial message for accurate hepatectomy, proposed a visualized proof led to optimal operation plan in minor or major hepatectomy. It can decrease surgical risk,reserve as much normal liver tissues as possible, and help prevent postoperative hepatic failure.