Impact of Three-Dimensional Reconstruction Technique in the Operation Planning of Centrally Located Hepatocellular Carcinoma

Transcription

1 Impact of Three-Dimensional Reconstruction Technique in the Operation Planning of Centrally Located Hepatocellular Carcinoma Chi-hua Fang, MD, Hai-su Tao, MM, Jian Yang, MD, Zhao-shan Fang, MM, Wei Cai, MM, Jun Liu, MD, Ying-fang Fan, MD BACKGROUND: The aim of this retrospective study was to compare the outcomes of operations based on 3- dimensional (3D) operation planning with non 3D-assisted operations in the treatment of centrally located hepatocellular carcinoma. STUDY DESIGN: From April 2008 to March 2014, 116 patients with centrally located hepatocellular carcinoma received surgical treatment in our department. Among these cases, a total of 60 patients received resection with operation planning based on 3D reconstructions (group A); the remaining 56 received treatment with the aid of traditional imaging (group B). Threedimensional surgical planning, including the classification system for centrally located hepatocellular carcinoma, was elaborated in the study. RESULTS: Compared with group B, group A was linked to shorter operation time ( minutes vs minutes; p ¼ 0.028) and lower rate of hepatic inflow occlusion (51.7% vs 71.4%; p ¼ 0.029). No differences were found in surgical methods, intraoperative blood transfusion, and intraoperative blood loss. The groups were similar in their rates of complications, except that group B was more liable to have Clavien Grade III to V complications (3.3% vs 14.3%; p ¼ 0.048). In addition, a significant difference in ascites was found across the 2 cohorts (2 in group A and 8 in group B; p ¼ 0.048), and the 2 groups also differed significantly in total bilirubin ( g/l vs g/l; p ¼ 0.032) and albumin ( g/l vs g/l; p ¼ 0.033). CONCLUSIONS: Compared with non 3D-assisted operations, the operation planning based on 3D reconstruction is a more effective and reasonable method in the treatment of centrally located hepatocellular carcinoma. In addition, the classification system may facilitate the 3D operation planning. (J Am Coll Surg 2015;220:28e37. Ó 2015 by the American College of Surgeons) Disclosure Information: Nothing to disclose. Support: This paper was supported by the National High Technology Research and Development Program of China (863 Program) (Grant No. 2006AA02Z346 and 2012AA021105), the Natural Science Foundation of Guangdong Province, China (Grant No ), the strategic cooperation project jointly funded by Guangdong Province and the Chinese Academy of Sciences (Grant No. 2010A ), the Science and Technology Project of Guangdong Province, China (Grant No.2011t and 2011B ), the National Natural Science Foundation of China (Grant No ), and the Industrial Technology Research and Development Program of Guangdong Province (Grant No. 2012A ). Received June 29, 2014; Revised August 17, 2014; Accepted September 29, From the Department of Hepatobiliary Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou, Guangdong Province, China. Correspondence address: Chi-hua Fang, MD, Department of Hepatobiliary Surgery, Zhujiang Hospital, Southern Medical University, Guangzhou, Guangdong Province , China. fangch_dr@163.com Asian countries account for nearly 78% of the roughly 600,000 cases of HCC (hepatocellular carcinoma) reported globally each year. 1 The leading cause is chronic hepatitis B virus infection in Eastern Asia, including China, Hong Kong, Indonesia, Korea, and Taiwan. Chronic hepatitis B virus carriers in Asia account for the majority of the 3.6 billion all over the world. 2 Although opinions are widely divided on the treatment of HCC, surgery still dominates in the comprehensive treatment of liver cancer. It can provide opportunities for cure in some patients and can prolong the overall survival time of patients who have surgery. Because the central location of hepatocellular carcinoma can be tricky, it is traditionally resected by right lobectomy (segments V, VI, VII, and VIII), left lobectomy segments II, III, and IV, and extended right/left lobectomy. ª 2015 by the American College of Surgeons Published by Elsevier Inc ISSN /14

2 Vol. 220, No. 1, January 2015 Fang et al Operation Planning Based on Three-Dimensional Reconstruction 29 Nevertheless, if 60% to 80% of the liver parenchyma is removed, patients will suffer from a risk of significant blood loss and postoperative liver failure, especially for those with cirrhosis or poor preoperative liver function. 3-5 One way to minimize the volume of the resected liver is to remove the central hepatic segments (Couinaud s segments IVA, IVB, V, and VIII), and preserve the functional parenchyma to prevent postoperative liver failure. The first description of central (medial segmentectomy), or now mesohepatectomy, was reported in Although mesohepatectomy retains more functional liver tissue, there are still risks of postoperative liver failure because of intraoperative blood loss and time spent controlling intraoperative blood flow. In addition, there may be bilateral damage to key structures, which inevitably has a negative effect on the residual liver function. 7 Importantly, there are certain groups of patients who cannot tolerate mesohepatectomy (resection of segments IV, V and VIII I) because they have posthepatitic cirrhosis. In China, 80% to 90% of liver cancer patients suffer from different degrees of posthepatitic cirrhosis, which require individualized operations to preserve more liver parenchyma. 8 To achieve individualization in the surgical treatment of centrally located hepatocellular carcinoma, 3D reconstruction technology is helpful. Although digital medicine is not recognized by many, it has gained acceptance because it combines computer technology and medicine. The 3D reconstruction technique was introduced to our department about 10 years ago According to our long-term experience using the 3D reconstruction technique in surgery, an individualized preoperative surgical planning process was developed and used for patients with centrally located hepatocellular carcinoma. This retrospective study aimed to assess the effects of individualized hepatectomy with preoperative planning based on a 3D reconstruction technique for centrally located hepatocellular carcinoma. METHODS Ethics statement The Ethics Committee of Zhujiang Hospital, Southern Medical University, approved this retrospective study and supervised procedures. Patients In the Department of Hepatobiliary Surgery of Zhujiang Hospital, the medical records of 116 patients with centrally located hepatocellular carcinoma, who received curative liver resection from April 2006 to March 2014, were retrospectively analyzed. The patients with centrally located hepatocellular carcinoma were carefully selected using the following criteria: 1. Postoperative histologic examination of the tumor indicated hepatocellular carcinoma; 2. Preoperative CT and/or MRI were used to determine the size of the tumor, with a maximum diameter >3 cm; 3. Tumors were situated mainly in the central segments of the liver (Couinaud segments IV, V, and VIII I); 4. There was no intrahepatic or extrahepatic disseminated disease or tumor thrombus in the main vein trunk; 5. Liver function was evaluated using Child-Pugh Criteria (Child-Pugh grades A and B); 6. The liver remnant after liver resection contained more than 50% of the functional liver volume on CT volumetric assessment. 12 All patients were informed of treatment details including procedure, risks, and complications, and then were grouped according to whether they had accepted treatment and had preoperative planning using the 3D reconstruction technique. As a result, 60 patients were assigned to group A because they had surgical therapy with preoperative planning based on the 3D reconstruction technique, and the remaining 56 patients were put into group B because they received liver resection only, with preoperative planning based on traditional imaging modalities. To ensure operative quality and safety, all surgical therapies and postoperative management were overseen by the same surgical team. The technological process of surgery planning based on 3D reconstruction 1. Enhanced CT scanning and CT scanning parameters setting. Data were collected by Philips Brilliance 256-MDCT scanner. 2. Collecting 2D image data and storing data. The 2D images collected by Philips Brilliance 256-MDCT scanner with predefined scanning parameters were processed by MxliteView DICOM Viewer. 3. Images segmentation and 3D reconstruction. The Medical Image Three-Dimensional Visualization System (MI-3DVS, proprietary software developed by the authors [software copyright No.:2008SR18798].) allowed segmentation and 3D reconstruction of the CT images, in which thin-sliced CT data were imported into the software to facilitate their automatic registration. 4. Surgery planning based on a 3D reconstruction technique. The reconstructed models were exported as Standard Template Library files and imported to the Free Form Modeling System (SensAble Technologies, Inc), on which the spatial distribution of the anatomic structure, hepatic artery blood supply, types of hepatocellular carcinoma, and the variation of hepatic artery were all presented.

3 30 Fang et al Operation Planning Based on Three-Dimensional Reconstruction J Am Coll Surg According to the distribution area of hepatic vein and the shape of the portal vein, MI-3DVS could segment the liver individually (Couinaud s anatomic segments). Then we classified the cases based on a special classification method (Fig.1). The volume measurement tool attaching to the FreeForm Modeling system can measure the volume of liver and tumor in the model. After the virtual surgery through the FreeForm Modeling System, we calculated the percentage of the volume of the remnant liver and decided on the appropriate operation. 9,13,14 Classification system for centrally located hepatocellular carcinoma based on 3D reconstruction Concentrating on the involvement of resected segments and the anatomic location of tumors relative to the principal hepatic vascular structures, centrally located hepatocellular carcinoma was divided into 5 subtypes by a classification system based on liver anatomy and resection method. Type I As shown in Figure 1A, this type of hepatocellular carcinoma occupies the liver parenchyma of segments V, VIII, or both. They are characterized by their close proximity to or even direct violation of the adjacent portal vein. They do not adhere to or compress the right hepatic vein trunk. Complete excision of these lesions is required in the resection of segments V, VIII partial irregular resection of segment IV. Type II As shown in Figure 1B, this type of hepatocellular carcinoma invades the liver parenchyma of segments IVa, Figure 1. Classification of centrally located hepatocellular carcinoma. (A) Type I; (B) type II; (C) type III; (D) type IV; (E) type V. The Roman numerals indicate Couinaud s classification of liver segments. IVC, inferior vena cava; LHV, left hepatic vein; LPV, left portal vein; MHV, middle hepatic vein; PV, portal vein; RAPV, right ahead portal vein; RHV, right hepatic vein; RPPV, right posterior portal vein.

4 Vol. 220, No. 1, January 2015 Fang et al Operation Planning Based on Three-Dimensional Reconstruction 31 IVb, or both. It is characterized by its close proximity to or even direct violation of the left hepatic vein trunk. In addition, it does not adhere to or compress the left hepatic vein trunk. For these reasons, the solution is to completely resect segments IVa, IVb partial irregular resection of segments V and VIII. Complete excision of these lesions was required in the section of IVa, IVb partial irregular resection of segments V and VIII. Type III As shown in Figure 1C, this type of hepatocellular carcinoma occupies the most liver parenchyma of segments IV, V, and VIII. These lesions are characterized by their wide and deep invasion of the parenchyma, or their close proximity to the middle hepatic vein. In addition, the liver function is not abnormal, so enough liver can be reserved. These lesions require mesohepatectomy (resection of segments IV, V, and VIII I). Type IV As shown in Figure 1D, this type of hepatocellular carcinoma occupies the most liver parenchyma of segments V, VIII, and IV. The lesions are characterized by their close proximity to, or a direct violation of, the left/right portal vein trunk or the left/right hepatic vein. Liver function is not abnormal and enough liver can be reserved. These lesions require a traditional operative procedure including right lobectomy (segments V, VI, VII, and VIII), left lobectomy (segments II, III, and IV), and extended right/left lobectomy. Type V As shown in Figure 1E, this type of hepatocellular carcinoma occupies the superficial liver parenchyma of segments V, VIII, and IV. The lesions are characterized by not being close to either the portal branch or the hepatic vein. In this situation, irregular resection of liver parenchyma is an option. Perioperative and operative data Preoperative clinical data, surgical variables, and postoperative clinical outcomes were analyzed in order to identify independent variables and compare the outcomes of the 2 groups. Patient demographics and clinical characteristics collected included age, sex, cirrhosis, viral hepatitis, preoperative transarterial chemoembolization, platelets, alphafetoprotein, live function (prothrombin time, Child-Pugh classification, albumin, total bilirubin, alanine aminotransferase, and aspartate transaminase) and the number and size of tumors based on preoperative CT/MR. Surgical variables included the method of hepatic resection, operative time, hepatic inflow occlusion, intraoperative blood transfusion, intraoperative blood loss, and characteristics of the tumors. Postoperative clinical outcomes included length of stay, postoperative complications, perioperative mortality (defined as mortality within 30 days) and postoperative examination index. Statistical analysis Statistical analysis was performed using SPSS 20.0 for Windows. Continuous data were presented as mean standard deviation, while categorical variables were presented as n (%). For statistical analysis, categorical variables were compared using the chi-square test or Fisher s exact test, and continuous variables were compared with the Student s t-test. In all cases, statistical significance was defined as p < RESULTS Patient demographics and clinical characteristics Patient demographics, clinical characteristics, and preoperative medical imaging outcomes are shown in Table 1. The 2 groups shared similarities in age and sex distributions. The majority of the patients had cirrhosis and hepatitis B, and the 2 groups did not significantly differ according to preoperative liver function, alphafetoprotein values, or percentage of patients with liver cirrhosis and viral hepatitis. Preoperative evaluations were performed in all patients with ultrasonography, CT, and magnetic resonance, and 3D reconstruction from CT scan data was done in patients in group A. The 2 groups had similar tumor size and number of tumors based on preoperative imaging outcomes. Outcomes of 3D reconstruction Three-dimensional reconstruction models clearly displayed all the structures, including the bile duct, hepatic artery, hepatic vein, portal vein, and size, shape, number, and location of the tumors. The 3D reconstruction model could be amplified, rotated, and hyalinized to clarify the anatomic character of tissue structure with omnidirectional, multiple-angle, and multilevel views. In order to conduct the virtual surgery, we needed to locate the positions and determine the shapes of tumors and the distributions of the intrahepatic vessel system, which helped to identify and map the variations of the intrahepatic vascular structures and their relationship to tumors. 10 Operation planning in 3D reconstruction models An 81-year-old man was diagnosed with a liver containing lesions at physical examination 1 week before. Laboratory tests showed: albumin (ALB), 39.2 g/l; alanine

5 32 Fang et al Operation Planning Based on Three-Dimensional Reconstruction J Am Coll Surg Table 1. Preoperative Clinical Data of the Two Groups Variables Group A (n ¼ 60) Group B (n ¼ 56) p Value Age, y, mean SD Sex, male/female 52/8 50/ Chronic hepatitis, n (%) Hepatitis B 51 (85.0) 41 (73.2) Hepatitis C 2 (3.3) 2 (3.6) Hepatitis Bþ Hepatitis C 2 (3.3) 1 (1.8) Negative 9 (15.0) 14 (25.0) Cirrhosis, n (%) 41 (68.3) 34 (60.7) Alcohol intake, n (%) 22 (36.7) 17 (30.4) TBIL, mmol/l, mean SD ALT, U/L, mean SD AST, U/L, mean SD Platelets, 10 9 /L, mean SD ALB, g/l, mean SD Prothrombin time, s, mean SD AFP > 20 mg/l, n (%) 41 (68.3) 37 (66.1) Child-Pugh, n (%) A 56 (93.3) 53 (94.6) B 4 (6.7) 3 (5.4) Preoperative imaging* Tumor sites, n (%) Single 46 (76.7) 41 (73.2) Multiple 14 (23.3) 15 (26.8) Tumor size, cm, mean SD Preoperative TACE, n (%) 6 (10.0) 5 (8.9) *Based on preoperative CT and/or MRI. AFP, alpha fetoprotein; ALB, albumin; ALT, alanine aminotransferase; AST, aspartate transaminase; TACE, transcatheter arterial chemoembolization; TBIL, total bilirubin. aminotransferase (ALT), 39 U/L; aspartate transaminase (AST), 38 U/L; total bilirubin (TBIL), 7.7 mmol/l; alpha-fetoprotein (AFP), mg/l; hepatitis B surface antigen (þ), hepatitis B e antibody (þ), and hepatitis B core antibody (þ). Epigastric CT enhancement scanning showed that an approximately 7.5 cm 7.0 cm low density shadow was found in the left interior lobe and right anterior lobe in the CT scan. In the arterial phase, the lesion was heterogeneously enhanced and had disordered vascular enhancement points. In the vein phase, the lesion dispersed fast, and patches of low density shadow could be observed (Fig. 2A). Clinical diagnoses included hepatocellular carcinoma; liver function, Child-Pugh A; and carriers of hepatitis B. After the original CT data had been reconstructed by MI-3DVS, preoperative planning and assessment of resectability were done using the 3D model. As the 3D model displayed (Fig. 2B), the tumor was located mainly in segments IV, V, and VIII, compressing the middle hepatic vein. It had not yet invaded the left or right portal vein trunks. Therefore, the tumor was classified as a lesion of Type III. As in Figure 2C and Figure 2D, the virtual operation was carried out and the percentage of the liver remnant volume was calculated. Depending on the outcomes of the virtual surgery and calculation, mesohepatectomy was determined as a treatment for the patient. Intraoperative conditions are shown in Figures 2E and 2F. Surgery methods and intraoperative data As shown in Tables 2 and 3, the 2 groups were not different in terms of operative methods, intraoperative blood transfusion, and intraoperative blood loss, but group A had a slightly shorter operative time and a lower rate of hepatic inflow occlusion. There were also some significant differences in the mean tumor diameters between the 2 groups (group A, mean diameter cm; group B, mean diameter cm), which might be caused by the selection of samples or patients preference for new technology when they were in a bad situation. Groups were similar in the number of tumors,

6 Vol. 220, No. 1, January 2015 Fang et al Operation Planning Based on Three-Dimensional Reconstruction 33 Figure 2. Operative procedure planning based on 3D reconstruction technique. (A) Venous phase of CT enhancement scan. (B) 3D reconstruction model (liver was hyalinized). (C) Virtual surgery based on 3D reconstruction model. In the virtual surgery, the portal vein was sutured and repaired after the lesion had been resected. 1, ligation of small branches of portal vein; 2, ligation of small branches of portal vein; 3, inferior vena cava; 4, left portal vein; 5, splenic vein. (D) According to the operation planning, the residual volume of liver was calculated. (E) During the operation, we designed the transection line and marked it with an electrocoagulation knife. (F) During the operation, the portal vein was sutured and repaired after the lesion was resected. 1, portal vein. encapsulation, satellite nodules, vascular invasion, and tumor differentiation (Table 3). Postoperative complications and clinical outcomes There was no difference in the number of postoperative complications, while the patients of group B were more susceptible to Clavien III (needed procedural intervention), Clavien IV (needed ICU care), or Clavien V (death) complications (3.3% vs 14.3%; p ¼ 0.048). The cases of bile leakage (1 in group A and 6 in group B) and intra-abdominal abscess (1 in group A and 5 in group B) were lower in group A than in group B, but not significantly. The patients who had all received major hepatectomy (extended hepatectomy or mesohepatectomy) were treated with full abdominal drainage. The 2 groups had a significant difference in the number of ascites (2 in group A and 8 in group B; p ¼ 0.048). As shown in Table 4, no difference was found in the 2 groups for other complications. There were 3 perioperative deaths: 1 from respiratory failure after serious pulmonary infection in group A; 1 from multiple organ failure after liver failure, and another from sepsis after intraabdominal abscess in group B. The hospital mortality of group B was 3.6%, while group A s was 1.7% (Table 4). No significant differences were found in the 2 groups with respect to alanine aminotransferase, aspartate transaminase, or prothrombin time. What s more, hemoglobin was lower in group A than in group B, but not

7 34 Fang et al Operation Planning Based on Three-Dimensional Reconstruction J Am Coll Surg Table 2. Methods of Hepatic Resection in the Two Groups Group A (n ¼ 60) Group B (n ¼ 56) Variables, operative method n % n % p Value Segment IV partial irregular resection Segment V, VIII partial irregular resection Irregular resection Right/left extended hepatectomy Right/left semi-hepatectomy Mesohepatectomy significantly. Group B showed a significantly higher total bilirubin ( g/l vs g/l; p ¼ 0.032) and a significantly lower albumin ( g/l vs g/l; p ¼ 0.033) (Table 5). DISCUSSION Individualized surgery for centrally located hepatocellular carcinoma There have been worldwide controversies about the treatment of centrally located hepatocellular carcinoma, with focus on the surgical method, hepatic inflow occlusion method, and intraoperative margin selection. In Asia, however, surgeons now popularly tend toward protective liver resection because patients with centrally located hepatocellular carcinoma often suffer from cirrhosis after hepatitis In the past, many of them died after radical hepatectomy, which challenged hepatobiliary surgeons to improve the resection rate in order to reduce the operation-induced mortality. 8 Fortunately, individualized mesohepatectomy has evolved from routine hepatectomy because digital medical science has broadened our horizon by minimizing the operative scope clinically and renewing the concept of tumor treatment theoretically. The curative treatment of HCC has been significantly improved by radically extending the resection range instead of by locating the tumor, specifying its size, and precisely removing it for the comprehensive considerations of reasonable reservation of liver function and general condition. In spite of even more prospects of mesohepatectomy, the special location of centrally located HCC, mainly occupying the parenchyma between the hepatic veins and the bifurcation of the portal vein, perplexes hepatobiliary surgeons as to what the spatial relationship between the tumors and the major intrahepatic vascular network can be before they carry out their operative plans. To solve the problem, the 3D reconstruction technique is an unquestioned option for its strengths over other imaging techniques in terms of individualized hepatectomy. Traditional imaging techniques and 3D reconstruction technique Primary imaging techniques for the diagnosis of HCC include ultrasonography, CT, MR, PET-CT, and Table 3. Key Indices of the Operation and Characteristics of Tumors Found Variables Group A (n ¼ 60) Group B (n ¼ 56) p Value Operative time, min, mean SD * Hepatic inflow occlusion, n (%) 31 (51.7) 40 (71.4) 0.029* Intraoperative blood transfusion, ml, mean SD Intraoperative blood loss, ml, mean SD Number of tumors, n (%) y Single 45 (75.0) 40 (71.4) Multiple 15 (25.0) 16 (28.6) Tumor size, cm, mean SD y * Encapsulation, n (%) y 41 (68.3) 41 (73.2) Satellite nodules, n (%) y 13 (21.7) 14 (25.0) Vascular invasion, n (%) y 14 (23.3) 11 (19.6) Tumor differentiation, n (%) Low 6 (10.0) 8 (14.3) Middle 50 (83.3) 42 (75) high 4 (6.7) 6 (10.7) *Significant difference. Based on intraoperative findings.

8 Vol. 220, No. 1, January 2015 Fang et al Operation Planning Based on Three-Dimensional Reconstruction 35 Table 4. Postoperative Clinical Outcomes Variables Group A (n ¼ 60) Group B (n ¼ 56) p Value Length of stay, d, mean SD Postoperative complications, n (%) 11 (18.3) 15 (26.8) Clavien IeII, n (%)* 9 (15.0) 7 (12.5) Clavien IIIeV, n (%)* 2 (3.3) 8 (14.3) y Wound infection, n (%) 2 (3.3) 2 (3.6) Bile leakage, n (%) 1 (1.7) 6 (10.7) Intra-abdominal abscess, n (%) 1 (1.7) 5 (8.9) Pleural effusion, n (%) 8 (13.3) 10 (17.9) Pulmonary infection, n (%) 4 (6.7) 5 (8.9) Hemorrhage, n (%) 1 (1.7) 3 (5.4) Ascites, n (%) 2 (3.3) 8 (14.3) y Other, n (%) 2 (3.3) 3 (5.4) Perioperative mortality, n (%) 1 (1.7) 2 (3.6) *Clavien-Dindo classification. Significant difference. Clavien grade I, complications including any deviation from the postoperative course that do not require any intervention; grade II, complications require pharmacologic therapy, including blood transfusions; grade III, complications require procedural intervention under either local or general anesthesia; grade IV, complications indicate single- or multi-organ failure requiring ICU admission; grade V, complications result in death. angiography; all of these are 2D based. Surgeons obtain visual imaging data on the relationship of organs or tissues, depending only on their experienced judgment from the 2D modalities. To many young doctors who lack experience, however, a sound judgment in some complicated cases based on the 2D modalities is surely in question, let alone the operator-dependent ultrasound and digital subtraction angiography (DSA). Besides the problem of its invasiveness, digital subtraction angiography is rarely used alone in the diagnosis of HCC, although it is largely used in the auxiliary diagnosis of HCC and in hepatic artery embolism chemotherapy of HCC. 19 To date, the 2D imaging data by CT or MRI scanning is transferable into 3D visualization by way of segmentation and reconstruction technologically. 20 Three-dimensional reconstruction models facilitate acquisition of intuitionistic and omnidirectional information about the hepatic parenchyma, bile duct system, and tumors in great detail, and what s more, they eliminate concern about the vagueness and instability of CT and MRI images. More importantly, 3D reconstruction has power in evaluating preoperative blood supply, impaired function of the drainage area, resection scope, and necessity of vascular reconstruction, which are extraordinarily important to complicated patients with distortion of the hepatic vascular anatomy, extended hepatectomy, and mesohepatectomy. Therefore, 3D reconstruction can improve operative accuracy and security by rational planning of the hepatectomy and the operation path. 11,21 Classification for centrally located hepatocellular carcinoma In China, major liver resection would not be tolerated by a large number of HCC patients, so it makes sense to manage the patients with centrally located hepatocellular carcinoma by following the principle of maximum preservation of liver parenchyma. 22 Despite the modified mesohepatectomy procedures that have been reported, there is still no reliable classification system that can help resolve the operation strategy for centrally located HCC. The right anterior lobe Table 5. Postoperative Laboratory Indices of the Two Groups Variables Group A (n ¼ 60) Group B (n ¼ 56) p Value TBIL, mmol/l, mean SD * ALB, g/l, mean SD * ALT, U/L, mean SD AST, U/L, mean SD HGB, g/l, mean SD Prothrombin time, s, mean SD *Significant difference. ALB, albumin; ALT, alanine aminotransferase; AST, aspartate transaminase; HGB, hemoglobin; TBIL, total bilirubin.

9 36 Fang et al Operation Planning Based on Three-Dimensional Reconstruction J Am Coll Surg (segments V, VIII) and the left internal lobe (segment IV) of the liver are independent from each other with respect to the portal vein, artery, and bile duct. According to the literature and clinical experience, segment IV is supplied mainly by the branch of the main trunk of the left side of the Glisson system. In order to work out a rational operation plan and better treatment, centrally located HCC was classified based on the anatomy of the liver. Centrally located HCC can be divided into 5 types, each characterized by its appropriate treatment depending on its involvement in the resected segment, the anatomic location, and the proximity of the lesions to the principal hepatic vascular structures. 23 Analyses of the clinic outcomes In this study, the 2 groups shared similarities in demographics, clinical characteristics, and operative methods. However, there were some differences between the 2 groups in intraoperative data, postoperative complications, and clinical outcomes. The shorter operative time in group A as compared with group B ( minutes vs minutes; p ¼ 0.028) may be due to the rational preoperative planning and sensible choice of operation path. That outcomes of preoperative planning were consistent with intraoperative findings may have contributed to the accuracy and fluency of the operation, which undoubtedly shortened operative time. Group A had a lower rate of hepatic inflow occlusion as compared with group B (51.7% vs 71.4%; p ¼ 0.029); this is explained by the fact that the 3D model may facilitate location of tumors and identification of their anatomic relationship with the adjacent key vascular system. As compared with group B, group A was not susceptible to bile leakage (1 in group A and 6 in group B) or intra-abdominal abscess (1 in group A and 5 in group B), and patients were less likely to have Clavien grade III to V complications. This can be due to reasonable preoperative planning and choice of operation path. Regarding clinical indices, there were significant differences across the 2 cohorts in ascites (2 in group A and 8 in group B, p ¼ 0.048), total bilirubin ( g/l vs g/l; p ¼ 0.032), and albumin ( g/l vs g/l; p ¼ 0.033). These findings prove that the 3D reconstruction technique used for treatment of centrally located HCC by hepatectomy may achieve the ultimate purpose of protecting liver function. Consequently, operation planning based on 3D reconstruction models may be an appropriate and sensible choice for centrally located HCC patients. Limitations and prospect This study is limited by its retrospective nature and to selective bias in general. Also, the data from the patients were short-term, which is believed to impair the study to a certain degree. Therefore, additional long-term research would be worthwhile in order to reach a more assertive conclusion about the effects of both kinds of operations in the treatment of centrally located HCC. Finally, the classification was roughly done mainly in accordance with the spatial distribution of tumor and its adjacent relationship with important vascular system structures; therefore, quantitative indicators are deemed reliable and usable for further study. CONCLUSIONS Although the surgical strategy for centrally located HCC remains controversial, individualized hepatectomy may be a safer choice, especially when the preservation of liver function is critical for centrally located HCC patients. The findings of this study show that operation planning based on 3D reconstruction models may be a reasonable and effective method for treating selected patients with centrally located HCC. Author Contributions Study conception and design: C-H Fang, Tao, Fan Acquisition of data: Tao, Z-S Fang, Cai Analysis and interpretation of data: Tao, Liu Drafting of manuscript: Tao, Yang Critical revision: C-H Fang, Tao, Yang Acknowledgment: The authors wish to thank the Imaging Center of Zhujiang Hospital for their superb technical and equipment assistance in data collection and analysis and the Statistic Departments of Southern Medical University for statistical assistance. We also would like to express our gratitude to Prof. Wu Rangke from the School of Foreign Studies, Southern Medical University, for his contribution to the English translation of this paper. REFERENCES 1. Management of hepatocellular carcinoma in Asia: consensus statement from the Asian Oncology Summit Raoul J. Natural history of hepatocellular carcinoma and current treatment options. Sem Nuclear Med 2008;38:13e Hu R, Lee P, Chang Y, et al. Treatment of centrally located hepatocellular carcinoma with central hepatectomy. Surgery 2003;133:251e Lee JG, Choi SB, Kim KS, et al. Central bisectionectomy for centrally located hepatocellular carcinoma. Br J Surg 2008;95: 990e995.

10 Vol. 220, No. 1, January 2015 Fang et al Operation Planning Based on Three-Dimensional Reconstruction Cheng C, Yu M, Wu T, et al. Surgical resection of centrally located large hepatocellular carcinoma. Chang Gung Med J 2012;35:178e Scudamore CH, Buczkowski AK, Shayan H, et al. Mesohepatectomy. Am J Surg 2000;179:356e Stratopoulos C, Soonawalla Z, Brockmann J, et al. Central hepatectomy: The golden mean for treating central liver tumors? Surg Oncol 2007;16:99e Liang Li-jian. Preoperative evaluation and decision of complicated hepatectomy. Chinese J Pract Surg 2010;30:645e Fang CH, Liu J, Fan Y, et al. Outcomes of hepatectomy for hepatolithiasis based on 3-dimensional reconstruction technique. J Am Coll Surg 2013;217:280e Lang H, Radtke A, Liu C, et al. Extended left hepatectomymodified operation planning based on three-dimensional visualization of liver anatomy. Langenbecks Arch Surg 2004;389: 306e Fang CH, Yang J, Fan Y, et al. The research of virtual hepatectomy. Chinese J Surg 2007;45:753e Chen X, Qiu F, Lau W, et al. Mesohepatectomy for hepatocellular carcinoma: a study of 256 patients. Int J Colorectal Dis 2008;23:543e Fang CH, Lu CM, Huang YP, et al. Study on the application of value of digital medical technology in the operation on primary liver cancer. Chin J Surg (Chin) 2009;47:523e Yang J, Fang C, Fan Y, et al. To assess the benefits of medical image three-dimensional visualization system assisted pancreaticoduodenctomy for patients with hepatic artery variance. Int J Medical Robotics Computer Assisted Surg doi.org/ /rcs Otsubo T. Control of the inflow and outflow system during liver resection. J Hepato-Biliary-Pancreatic Sci 2012;19: 15e Yu W, Rong W, Wang L, et al. R1 Hepatectomy with exposure of tumor surface for centrally located hepatocellular carcinoma. World J Surg 2013;37:2394e Machado MA, Herman P, Machado MC. Intrahepatic Glissonian approach for pedicle control during anatomic mesohepatectomy. Surgery 2007;141:533e Chen P. A novel technique for central hepatectomy: Maintain the blood supply and biliary drainage on one side and the blood supply from the portal vein on the other. Exp Ther Med 2014;7:51e Lin-Xue Qian, Yu-Jiang Liu, Hui-Guo Ding, et al. Advances in imaging diagnosis of small hepatocellular carcinoma. World Chinese J Digestology 2010;18:479e Zhao H, Yao JL, Wang Y, et al. Detection of small hepatocellular carcinoma: comparison of dynamic enhancement magnetic resonance imaging and multiphase multirow-detector helical CT scanning. World J Gastroenterol 2007;13: 1252e Orimo T, Kamiyama T, Yokoo H, et al. Usefulness of artificial vascular graft for venous reconstruction in liver surgery. World J Surg Oncol 2014;12: Torzilli G, Palmisano A, Procopio F, et al. A new systematic small for size resection for liver tumors invading the middle hepatic vein at its caval confluence: mini-mesohepatectomy. Ann Surg 2010;251:33e Qiu J, Wu H, Bai Y, et al. Mesohepatectomy for centrally located liver tumours. Br J Surg 2013;100:1620e1626.