Preoperative predictors of microvascular invasion in multinodular hepatocellular carcinoma

Available online at www.sciencedirect.com

EJSO 39 (2013) 858e864

www.ejso.com

Preoperative predictors of microvascular invasion in multinodular hepatocellular carcinoma W.-C. Zhao a,c, L.-F. Fan b,c, N. Yang b, H.-B. Zhang b, B.-D. Chen b, G.-S. Yang b,* b

a Department of Hepatobiliary Surgery, PLA Navy General Hospital, No. 6 Fucheng Road, Beijing 100048, China Fifth Department of Hepatic Surgery, Eastern Hepatobiliary Surgery Hospital, Second Military Medical University, No. 225 Changhai Road, Shanghai 200438, China

Accepted 25 April 2013 Available online 11 May 2013

Abstract Background: The preoperative predictors of microvascular invasion (MVI) in multinodular hepatocellular carcinoma (HCC) are currently unclear. Methods: We retrospectively analyzed 266 patients who underwent potentially curative resection of multinodular HCC. MVI was diagnosed on pathological examination in 64 patients. Preoperative risk factors for MVI were identified and survival curves were analyzed. Results: Patients with MVI had significantly lower overall and recurrence-free survival rates than those without MVI (overall survival, 1 year: 86% vs. 71%, 3 years: 58% vs. 16%; recurrence-free survival, 1 year: 69% vs. 12%; 3 years: 48% vs. 12%; both P < 0.001). Multivariate analysis showed that serum alpha-fetoprotein (AFP) level >400 mg/L (odds ratio [OR] ¼ 3.732, P ¼ 0.016), serum gammaglutamyltransferase (GGT) level >130 U/L (OR ¼ 19.779, P < 0.001), total tumor diameter >8 cm (OR ¼ 5.545, P ¼ 0.010), and tumor number >3 (OR ¼ 11.566, P ¼ 0.007) were independent predictors of MVI. A scoring system was constructed, and the MVI rate was significantly higher in patients with a score of 3 than those with a score of <3 (64.1% vs. 10.9%, P < 0.001). Overall and recurrence-free survival rates were significantly lower in patients with a score of 3 (both P < 0.001). Conclusions: Serum AFP level >400 mg/L, serum GGT level >130 U/L, total tumor diameter >8 cm, and tumor number >3 were preoperative predictors of MVI in patients with multinodular HCC. In patients with a high risk of MVI and well-preserved liver function, anatomic resection may be worth considering. Ó 2013 Elsevier Ltd. All rights reserved. Keywords: Hepatocellular carcinoma; Multinodular; Microvascular invasion

Introduction Hepatocellular carcinoma (HCC) is the third leading cause of cancer mortality worldwide, and the highest reported rates of HCC are in East Asia.1,2 Despite recent advances in imaging examinations, HCC frequently presents at an advanced stage, for many reasons including a lack of early symptoms. Approximately 40% of patients have multiple lesions at the time of diagnosis.3 Many studies have reported that hepatic resection is the optimal treatment for selected patients with multinodular HCC,4e7 but survival rates remain poor with a 5-year recurrence rate of 70e75%.4,5 * Corresponding author. Tel./fax: þ86 21 81875292. E-mail address: [email protected] (G.-S. Yang). c W.-C. Zhao and L.-F. Fan contributed equally to this work. 0748-7983/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejso.2013.04.003

Intrahepatic embolism secondary to vascular invasion is one of the most common reasons for HCC recurrence. HCC has a propensity to invade the portal vein8 resulting in intrahepatic spread via macroscopic or microscopic tumor emboli. Such intrahepatic spread may be diffuse, difficult to detect, and difficult to treat effectively.9 Vascular invasion is therefore a prognostic factor for survival after surgical resection. Macroscopic vascular invasion has an important influence on treatment selection and can be detected on imaging examinations, but identification of microvascular invasion (MVI) requires histological examination, which limits its usefulness for preoperative assessment of prognosis.10 Identification of preoperative markers of MVI could help the selection of appropriate surgical procedures in patients with multinodular HCC. Several recent studies have reported on the risk factors for MVI in patients undergoing resection for HCC, but

W.-C. Zhao et al. / EJSO 39 (2013) 858e864

few focused on patients with multinodular HCC. The aim of this study was to identify the predictive factors for MVI in patients with multinodular HCC. Materials and methods Patients Between January 2004 and December 2008, 453 patients with multinodular HCC underwent surgical resection at the Department of Surgery, Eastern Hepatobiliary Surgery Hospital, Second Military Medical University (Shanghai, China). We regarded a tumor with an adjacent nodule as a single tumor when the nodule was attached to the main tumor.4,5 Patients who met all of the following criteria were included in this study: (1) no macroscopic vascular invasion; (2) no extrahepatic metastasis; (3) potentially curative resection, defined as complete removal of all detectable tumor with tumor-free margins confirmed by histopathological examination4,5; and (4) no preoperative anti-HCC therapy. A total of 266 patients (51 females, 215 males; mean age 50.16 years, age range 27e76 years) were enrolled. The baseline clinical characteristics of the patients are shown in Table 1. All patients had Child-Pugh class A liver function. Fifty-six patients were Barcelona Clinic Liver Cancer stage A and 210 patients were stage B. Serum hepatitis B virus surface antigen was positive in 254 patients. There was no evidence of liver disease or infection in the remaining 12 patients, and no patients had a history of any other chronic liver disease. The study protocol was approved by the clinical research ethics committee of our hospital. Diagnosis of HCC and preoperative assessment HCC was diagnosed if at least one of the following criteria was fulfilled4,7,11: (1) focal hepatic lesions with contrast enhancement during the arterial phase and washout during the portal venous phase, detected on two different imaging modalities such as enhanced spiral computed tomography (CT) and magnetic resonance imaging (MRI); (2) focal hepatic lesions detected on imaging examinations and serum alpha fetoprotein (AFP) level >400 mg/L; (3) rapid enlargement of hepatic lesions without typical characteristics of HCC in a patient at high risk for HCC; or (4) histological evidence of HCC on fine needle aspiration biopsy. None of the patients in this study underwent preoperative fine needle aspiration biopsy for diagnosis. Definitive diagnosis was by histopathological examination after resection. Microvascular invasion was defined as invasion of HCC through the vascular endothelium, visible only on microscopy.12 All patients underwent preoperative laboratory testing including a complete blood count and evaluation of liver function by measurement of serum levels of total bilirubin, direct bilirubin, indirect bilirubin, albumin, globulin, prealbumin, alanine aminotransferase, aspartate aminotransferase,

859

Table 1 Baseline clinical characteristics at the time of diagnosis of HCC. Variable

Mean/n

Age(y) Sex(M:F) HBsAg-positive TBIL>17.1 (mmol/L) ALB <35 (g/L) Prealbumin<170(mg/L) PT > 13 s (n) ALT > 44U/L (n) AST > 38 U/L (n) ALP > 129U/L GGT > 64U/L AFP > 400 mg/L BCLC Staging Stage A Stage B Milan criteria Within Milan criteria Exceeding Milan criteria TNM staging (7th) T2 T3a Total tumor diameter > 8(cm) The largest tumor size > 5(cm) Number of tumors 2e3 >3 PLT < 100  109/L (n) Microvascular invasion Death in 3 months after operation

50 (median, range:27e76) 215:51 254 (96%) 84 (32%, range: 4.9e33) 14 (5%, range: 28.6e53) 45 (17%, range:21e500) 76 (29%, range: 10.4e17.6) 91 (34%, range: 12.9e235) 152 (57%, range: 12.8e207) 36 (14%, range: 36e368) 164 (62%, range: 13e1000) 104 (40%, range: 2.5e4537) 56 (21%) 210 (79%) 56 (21%) 210 (79%) 195 (73%) 71 (27%) 54 (20%, range:1.4e23) 66 (25%, range:0.7e21) 248 (93%) 18 (7%) 72 (27%) 64 (24%) 7 (3%)

Abbreviations: HBsAg: hepatitis B virus surface antigen; HBV: hepatitis B virus; AFP: alpha fetoprotein; TBIL: total bilirubin; ALB: albumin; PT: prothrombin time; ALT: alanine aminotransferase; AST: aspartate aminotransferase; ALP: alkaline phosphatase; GGT: gammaglutamyltransferase; BCLC: Barcelona Clinic Liver Cancer; PLT: platelet. TNM: 7th edition of the American Joint Committee on Cancer (AJCC) TNM staging system.

alkaline phosphatase, and gamma-glutamyltransferase (GGT). Prothrombin time and activated partial thromboplastin time were also measured to evaluate liver function and surgical safety. Tumor marker levels including AFP, carcinoembryonic antigen, carbohydrate antigen 19-9, and alpha fucosidase were measured to identify tumor origin. Kidney function was evaluated by measurement of serum levels of urea nitrogen, creatinine, uric acid, and electrolytes. All patients underwent upper gastrointestinal endoscopy to screen for portal hypertension and hemorrhage, and chest X-ray to screen for lung metastasis. The sizes and locations of tumors were evaluated by abdominal ultrasonography, enhanced CT, or MRI. Patients older than 60 years or who had a history of pulmonary or cardiovascular disease underwent pulmonary function testing and cardiovascular Doppler ultrasonography to screen for contraindications to hepatic resection. Hepatic resection The criteria for hepatic resection were: Child-Pugh class A liver function, technically feasible resection, no

860

W.-C. Zhao et al. / EJSO 39 (2013) 858e864

macroscopic malignant portal vein or vena cava invasion, no distant metastasis detected on preoperative evaluations, and sufficient estimated residual liver volume. Patients with resectable multinodular HCC underwent immediate resection without preoperative anti-HCC therapy. All patients underwent localized nonanatomic resection as previously described.13 Tumors were resected en bloc in 78 patients (29.32%) and by two or more concomitant liver resections in 188 patients (70.68%). Intraoperative ultrasonography was used during all procedures to detect nonvisible, nonpalpable nodules and to ensure that all detectable tumor tissue was completely removed. Resection margins were determined by histological examination. Follow-up All patients were followed up monthly for the first 3 months with surveillance for recurrence by measurement of serum levels of AFP and liver enzymes, complete blood count, and CT or MRI scan. If there was no recurrence, routine examinations continued every 3 months. Tumor recurrence was defined as a new lesion on imaging examinations with typical characteristics of HCC, increasing serum AFP level, or rapid enlargement of a lesion without typical characteristics of HCC. If tumor recurrence was detected, second hepatectomy, transcatheter arterial chemoembolization (TACE), or locoregional ablation such as radiofrequency ablation was recommended depending on liver function, liver volume, tumor size, and tumor location. Statistical analysis Continuous data were expressed as mean  standard deviation and compared using the unpaired t-test. Categorical data were compared using the c2 test or Fisher’s exact test as appropriate. Data regarding tumor number, size, and location were obtained from preoperative imaging examinations. Survival analyses were performed using the KaplaneMeier method, and the differences between curves were analyzed using the log-rank test. Overall survival was computed from the day of surgery to the day of death or to the most recent follow-up visit. Recurrence-free survival was computed from the day of surgery to the first follow-up visit at which there was clear evidence of recurrence or to the most recent follow-up visit. Deaths not related to HCC were included in the overall survival analyses but not in the recurrence-free survival analyses. A P value of less than 0.05 was considered significant. All statistical analyses were performed using SPSS 18.0 (SPSS Inc, Chicago, IL, USA). Construction of scoring system All potential predictive factors for MVI were dichotomized, and the two categories for each factor were compared using the chi-square test. Multivariate logistic

regression analysis was performed to identify independent predictive factors for MVI. All factors with a P value of less than 0.1 in the univariate analyses were entered into the multivariate logistic regression analysis, and four significant predictive factors were identified. The scoring system was then developed using the point system described by Sullivan et al.14 Points were assigned to each significant predictive factor according to the relative magnitude of its partial regression coefficient (b) in the multivariate model. We defined the b value of one independent factor (We chose b value of the factor AFP > 400 mg/L as the constant in results section) randomly as the constant for the scoring system, corresponding to one point. Therefore, the point assigned to this risk factor was 1 and the points assigned to the other three risk factors were their quotients when divided by this constant. To calculate the score easily during clinical practice, all points were rounded to the nearest integer.14 The overall predictive score of an individual patient was calculated by summating the points for all the relevant risk factors. The ability of the overall predictive score to predict MVI was evaluated using receiver operating characteristic (ROC) curve analysis. Results Short-term outcomes The mean period of postoperative hospital stay was 10.6 days (range 5e28 days). The overall perioperative morbidity rate was 28.95% (n ¼ 77). The most common complications were plural effusion and ascites. Two patients developed bile leakage and one patient developed transitory arrhythmia. The in-hospital and 90-day mortality rates were 0 and 2.26% (n ¼ 6), respectively. One patient died of acute severe hepatitis, and five patients died of liver failure. Long-term outcomes The mean overall follow-up period was 30 months (range 3e81 months). The median overall survival time was 38 months (range 3e81 months) and the 1-, 3-, and 5-year overall survival rates were 83, 49, and 34%, respectively. The median recurrence-free survival time was 19 months (range 1e79 months) and the 1-, 3-, and 5-year recurrence-free survival rates were 54, 39, and 31%, respectively. A total of 177 patients (67%) were diagnosed with tumor recurrence during the follow-up period (57 in the MVI group and 120 in the non-MVI group). The sites of recurrence are shown in Table 2. We divided recurrence into four types according to the site as described by Couinaud’s classification system: near the surgical margin (type 1, n ¼ 33); in the same lobe or ipsilateral hemiliver (type 2, n ¼ 48); in the contralateral hemiliver (type 3, n ¼ 41); or scattered throughout the remaining liver (type 4, n ¼ 55). Patients with MVI had a high rate of type 2

W.-C. Zhao et al. / EJSO 39 (2013) 858e864

861

Table 2 Comparisons of characteristics between patients with and without MVI.

Survival of patients with and without MVI

Variables

MVI (n ¼ 64)

Without MVI (n ¼ 202)

P value

Sex, m/f Age < 40 y TBIL > 17.1 mmol/L ALB < 35 g/L PLT < 100  109/L ALT > 44U/L AST > 38 U/L ALP > 129 U/L GGT > 64 U/L GGT > 130 U/L AFP > 400 mg/L Total tumor size > 8 (cm) Size of largest nodule > 5 (cm) Tumor number > 3 Locateda In same segment In same sector In same hemiliver Site of recurrence tumorsb Near the surgical margin In the same hemiliver In the opposite hemiliver Throughout the remaining liver

55/9 13 (20%) 21 (41%) 4 (6%) 16 (25%) 16 (25%) 44 (69%) 8 (13%) 42 (66%) 28 (44%) 48 (75%) 36 (56%) 34 (53%) 11 (17%)

160/42 39 (19%) 37 (16%) 10 (5%) 56 (28%) 75(37%) 108 (53%) 28 (14%) 102 (51%) 38 (19%) 56 (28%) 18 (9%) 32 (16%) 7 (3%)

0.233 0.860 0.014 0.748 0.700 0.075 0.031 0.938 0.034 <0.001 <0.001 <0.001 <0.001 <0.001

MVI was diagnosed on postoperative histopathological examination in 64 of the 266 patients. In patients with MVI, the 1- and 3-year overall survival rates were 71 and 16%, respectively; and the 1- and 3-year recurrence-free survival rates were 12 and 12%, respectively. In the remaining 202 patients, the 1- and 3-year overall survival rates were 86 and 58%, respectively; and the 1- and 3-year recurrence-free survival rates were 69 and 48%, respectively. Patients with MVI had significantly lower overall and recurrence-free survival rates than those without MVI (both P < 0.001, Fig. 1A and B).

8 (13%) 18 (28%) 40 (65%)

20 (10%) 38 (19%) 76 (41%)

9 (33%) 23 (36%) 4 (7%) 21 (54%)

24 25 37 34

a b

(20%) (21%) (31%) (28%)

0.555 0.111 0.005 0.042 <0.001 0.019 0.001

According to Couinand’s system. According to Couinand’s system.

recurrence. The rate of type 4 recurrence was higher in patients with MVI than those without MVI, but this difference was not statistically significant. The rate of recurrence in the contralateral hemiliver was higher in patients without MVI than those with MVI. Of the 177 patients with recurrence, 19 (11%) underwent second hepatic resection, 121 (68%) underwent TACE alone, 29 (16%) underwent TACE combined with locoregional ablation, and 8 (5%) underwent locoregional ablation alone.

Predictive factors for MVI Clinical variables were compared between the 64 patients with MVI and the 202 patients without MVI (Table 2). Patients with MVI had higher serum levels of total bilirubin (P ¼ 0.014), alanine aminotransferase (P ¼ 0.075), aspartate aminotransferase (P ¼ 0.031), AFP (P < 0.001), and GGT (P < 0.001), as well as larger tumor size (P < 0.001) and a higher number of tumors (P < 0.001), than patients without MVI. Tumors were more commonly limited to one hemiliver in patients with MVI than in patients without MVI (P ¼ 0.005). Multivariate logistic regression analysis showed that serum AFP level >400 mg/L (OR ¼ 3.732, P ¼ 0.016), serum GGT level >130 U/L (OR ¼ 19.779, P < 0.001), total tumor diameter >8 cm (OR ¼ 5.545, P ¼ 0.010) and tumor number >3 (OR ¼ 11.566, P ¼ 0.007) were independent risk factors for MVI (Table 3). Ability to predict MVI As multivariate regression analysis indicated a different level of importance for each predictive factor, it was

Figure 1. The cumulative survival curves of patients with and without microvascular invasion. (A) Overall survival; (B) recurrence-free survival (both P < 0.001).

862

W.-C. Zhao et al. / EJSO 39 (2013) 858e864

Table 3 Risk factors for microvascular invasion on univariate and multivariate analyses. Variables AFP > 400 mg/L TBIL > 17.1 mmol/L GGT > 130U/L ALT > 44U/L AST > 38U/L Largest nodule > 5 cm Total tumor size > 8 cm Tumor number > 3 Located in hemiliver

Results of univariate analysis

Results of multivariate analysis

Score

OR

95%CI

P

b

5.989 2.667 16.857 0.553 1.915 2.746 3.950 4.615 2.679

2.655e13.511 1.262e5.837 6.586e43.144 0.293e2.041 1.055e3.477 1.225e6.157 1.797e8.682 1.587e13.426 1.220e5.885

<0.001 0.014 <0.001 0.066 0.033 0.014 0.001 0.005 0.014

1.317

3.732

1.274e10.932

0.016

1

2.985

19.779

5.888e66.440

<0.001

2

1.713 2.448

5.545 11.566

1.519e10.245 2.595e51.548

0.010 0.007

1 2

OR

95%CI

P

Abbreviations: OR: odds ratio; CI: confidence interval; b: partial regression coefficient.

inappropriate to assess the combined predictive ability of the factors by a simple combination of scores. We therefore constructed a scoring system that accounted for the b value of each independent predictive factor (Table 3).14 Each patient’s overall predictive score was then calculated by summation of the scores of the independent predictive factors, giving a theoretical range for the overall predictive score of 0e6. The overall predictive score was 0 in 117 patients, 1 in 30 patients, 2 in 56 patients, 3 in 30 patients, 4 in 8 patients, 5 in 21 patients, and 6 in 4 patients. A higher score indicated a higher risk of MVI. ROC curve analysis showed the predictive ability of this score (Fig. 2). The optimal cutoff score was 3, with a sensitivity of 64% and specificity of 89% for MVI. In high-risk patients (score 3; n ¼ 63), the MVI rate was 64%, and in low-risk patients (score <3; n ¼ 203), the MVI rate was 11% (P < 0.001). In high-

risk patients, the 1- and 3-year overall survival rates were 67 and 11%, respectively; and the 1- and 3-year recurrence-free survival rates were 17 and 11%, respectively. In low-risk patients, the 1- and 3-year overall survival rates were 87 and 59%, respectively; and the 1- and 3-year recurrence-free survival rates were 68 and 49%, respectively. The overall and recurrence-free survival rates were significantly lower in high-risk patients than in lowrisk patients (both P < 0.001, Fig. 3A and B). Discussion The long-term survival of patients with multinodular HCC who have undergone hepatic resection is strongly influenced by tumor recurrence in the remnant liver. Recurrence in the liver results primarily from intrahepatic dissemination via the portal circulation, and macrovascular invasion is considered to be the most powerful predictor of prognosis in patients with HCC who have undergone hepatic resection or liver transplantation.15,16 This study also found that microvascular invasion was a risk factor for early recurrence after resection of multinodular HCC. Our survival outcomes were similar to survival outcomes reported in previous studies.4,5 In patients with MVI, the recurrence rate was 76% at 3 months and 88% at 1 year after resection, indicating a high likelihood of undetectable preoperative micrometastases. Predicting MVI preoperatively

Figure 2. Receiver operating characteristic curve analysis of the overall predictive score for prediction of microvascular invasion. The area under curve was 0.832 (95% CI: 0.744, 0.920; P < 0.001). The optimal cut-off score was 3, with a sensitivity of 64% and specificity of 89%.

Accurate preoperative diagnosis of MVI is not considered possible. Several studies have examined the risk factors for MVI in HCC patients undergoing resection,17,18 but most of these included pathological characteristics such as tumor differentiation and intrahepatic metastasis, which cannot be determined preoperatively. The preoperative predictors of MVI remain unclear, especially in patients with multinodular HCC. In this study, we retrospectively analyzed 266 patients with multinodular HCC and identified serum AFP level >400 mg/L, serum GGT level >130 U/L, total tumor diameter >8 cm, and

W.-C. Zhao et al. / EJSO 39 (2013) 858e864

863

Figure 3. The survival curves of patients with scores of 3 and <3. (A) Overall survival; (B) recurrence-free survival (both P < 0.001).

tumor number >3 as independent predictive factors for MVI. All these tumor-associated parameters can be determined preoperatively. Several previous studies reported that tumor size and number, which are important parameters of tumor burden, were independent risk factors for MVI.15,17e19 The risk of MVI increased with increasing tumor size.15,19 Simultaneous detection of more than three lesions usually indicates intrahepatic metastasis rather than multicentric tumor. Serum AFP level has been shown to be a marker of HCC, and has been reported to correlate with factors associated with tumor invasiveness such as MVI and differentiation.19 GGT is produced by HCC cells and secreted into the bloodstream. Serum GGT level is considered to be a valuable marker of HCC in patients with a low serum AFP level.20,21 Serum GGT level is highly correlated with tumorigenesis, and is a sensitive biomarker of hepatocyte damage.20 However, this parameter is easily overlooked. Recent evidence confirmed that serum GGT level is associated with long-term survival and vascular invasion in HCC patients. A study including 219 HCC patients showed that higher serum GGT levels were associated with larger tumor size, increased risk of vascular invasion, more advanced tumor stage, and lower serum albumin level.22 Carr et al.21 found that serum GGT level >110 U/mL was one of the most significant predictors of survival. A study by Zeng et al.23 demonstrated that pretreatment serum GGT level was an independent risk factor in HCC patients with macroscopic portal vein or inferior vena cava invasion who underwent external-beam radiation therapy. In the present study, only a significantly increased serum GGT level (>130 U/L, twice the upper limit of the normal reference range) was found to be strongly associated with MVI. This may be because of the relatively low sensitivity of serum GGT level for detection of HCC, and the direct secretion of GGT into the bloodstream by the cells of tumor emboli.20

We constructed a scoring system to evaluate the predictive abilities of the above factors. ROC curve analysis showed that this scoring system had a high predictive value for MVI. Patients with a high score (3) had an MVI rate of 64%, which was significantly higher than in patients with a low score (<3). Comparisons of survival curves showed that the scoring system was useful for risk stratification. The results indicate that these risk factors are useful for preoperative prediction of MVI. Surgical strategies for patients with or without MVI Identifiction of predictors of MVI could improve the selection of surgical procedures in patients with multinodular HCC. All the patients in this study underwent local nonanatomic resection. The 3-month rate of tumor recurrence was nearly 80% in patients with MVI, indicating that complete resection of all detectable tumors may not remove all intrahepatic metastases in these patients. Univariate analysis showed that in patients with MVI, multiple tumors were more commonly located in one hemiliver compared with patients without MVI (65% vs. 41%, P ¼ 0.005). Recurrence in the ipsilateral hemiliver was more frequent in patients with MVI than those without MVI (36% vs. 21%, P ¼ 0.011). We therefore hypothesize that anatomic resection such as hemihepatectomy may provide curative treatment in patients with MVI. Anatomic hepatic resection in HCC patients helps to eradicate intrahepatic vascular metastases resulting from vascular invasion, and a recent report indicated that this procedure may be superior to local nonanatomic resection.24 Anatomic resection is generally recommended for patients with well-preserved liver function. In patients at high risk of MVI, anatomic resection removes the liver tissue that is most likely to contain micrometastases, and may thereby reduce the risk of recurrence. However, anatomic resection may lead to a decreased liver remnant size and an increased risk of

864

W.-C. Zhao et al. / EJSO 39 (2013) 858e864

postoperative liver failure,25 especially in patients with multinodular HCC. Shi et al.24 found that micrometastases were found at >1 cm from the primary tumor in very few patients without MVI. In this study, patients without MVI had a 1-year recurrence rate of only 31%, indicating that limited nonanatomic resection with a surgical margin of >1 cm may be sufficient in these patients.26 Anatomic resection is therefore worth considering in patients with a high risk of MVI who have well-preserved liver function, while limited nonanatomic resection may be suitable for patients with a low risk of MVI. This study is limited by its retrospective, single-center design and limited sample size. The scoring system should be validated in a large-scale, randomized, controlled trial. Further studies are therefore needed before definitive conclusions can be reached. Conclusion In conclusion, MVI is a well-known predictor of poor survival after hepatic resection for multiple HCC. Serum AFP level >400 ug/L, serum GGT level >130 U/L, total tumor diameter >8 cm, and tumor number >3 were identified as independent predictors of microvascular invasion. Anatomic resection aimed at eradicating micrometastases may be worth considering in patients with a high risk of MVI and well-preserved liver function. Conflict of interest statement The authors declare no conflicts of interest. References 1. Parkin DM. Global cancer statistics in the year 2000. Lancet Oncol 2001;2:533–43. 2. Altekruse SF, McGlynn KA, Reichman ME. Hepatocellular carcinoma incidence, mortality and survival trends in the United States from 1975e2005. J Clin Oncol 2009;27:1485–91. 3. Sun Y, Liang BL, Zhang XH, Shen J, Xie BK. Magnetic resonance imaging manifestation of 500 patients with primary hepatic cell tumor. Chin J Cancer 2002;21:509–13. 4. Ng KK, Vauthey JN, Pawlik TM, et al. Is hepatic resection for large or multinodular hepatocellular carcinoma justified? Results from a multiinstitutional database. Ann Surg Oncol 2005;12:1–10. 5. Ishizawa T, Hasegawa K, Aoki T, et al. Neither multiple tumors nor portal hypertension are surgical contraindications for hepatocellular carcinoma. Gastroenterology 2008;134:1908–16. 6. Ho MC, Huang GT, Tsang YM, et al. Liver resection improves the survival of patients with multiple hepatocellular carcinoma. Ann Surg Oncol 2009;16:848–55. 7. Lin CT, Hsu KF, Chen TW, et al. Comparing hepatic resection and transarterial chemoembolization for Barcelona clinic liver cancer stage B hepatocellular carcinoma: change for treatment of choice? World J Surg 2010;34:2155–61.

8. Fan ST, Ng IOL, Poon RT, Lo CM, Wong J. Hepatectomy for hepatocellular carcinoma: the surgeon’s role in long-term survival. Arch Surg 1999;134:1124–30. 9. Portolani N, Coniglio A, Ghidoni S, et al. Early and late recurrence after liver resection for hepatocellular carcinoma: prognostic and therapeutic implications. Ann Surg 2006;243:229–35. 10. Choi KK, Kim SH, Choi SB, et al. Portal venous invasion: the single most independent risk factor for immediate postoperative recurrence of hepatocellular carcinoma. J Gastroenterol Hepatol 2011;26: 1646–51. 11. Vivarelli M, Guglielmi A, Ruzzenente A, et al. Surgical resection versus percutaneous radiofrequency ablation in the treatment of hepatocellular carcinoma on cirrhotic liver. Ann Surg 2004;240:102–7. 12. Andreana L, Burroughs AK. Treatment of early hepatocellular carcinoma: how to predict and prevent recurrence. Dig Liver Dis 2010; 42S:S249–57. 13. Yang T, Zhang J, Lu JH, Yang GS, Wu MC, Yu WF. Risk factors influencing postoperative outcomes of major hepatic resection of hepatocellular carcinoma for patients underlying liver disease. World J Surg 2011;35:2073–82. 14. Sullivan LM, Massaro JM, D’Agostino Sr RB. Presentation of multivariate data for clinical use: the Framingham Study risk score functions. Stat Med 2004;23:1631–60. 15. Cucchetti A, Piscaglia F, Grigioni AD, et al. Preoperative prediction of hepatocellular carcinoma tumour grade and micro-vascular invasion by means of artificial neural network: a pilot study. J Hepatol 2010; 52:880–8. 16. Lim KC, Chow PK, Allen JC, et al. Microvascular invasion is a better predictor of tumor recurrence and overall survival following surgical resection for hepatocellular carcinoma compared to Milan criteria. Ann Surg 2011;254:108–13. 17. Kaibori M, Ishizaki M, Matsui K, Kwon AH. Predictors of microvascular invasion before hepatectomy for hepatocellular carcinoma. J Surg Oncol 2010;102:462–8. 18. Kim BK, Han KH, Park YN, et al. Prediction of microvascular invasion before curative resection of hepatocellular carcinoma. J Surg Oncol 2008;97:246–52. 19. McHugh PP, Gilbert J, Vera S, Koch A, Ranjan D, Gedaly R. Alphafetoprotein and tumor size are associated with microvascular invasion in explanted livers of patients undergoing transplantation with hepatocellular carcinoma. HPB (Oxford) 2010;12:56–61. 20. Whitfield JB. Gamma glutamyl transferase. Crit Rev Cl Lab Sci 2001; 38:263–355. 21. Carr BI, Pancoska P, Branch RA. Low alpha-fetoprotein hepatocellular carcinoma. J Gastroenterol Hepatol 2010;25:1543–9. 22. Ju MJ, Qiu SJ, Fan J, et al. Preoperative serum gamma-glutamyl transferase to alamine aminotransferase ratio is a convenient prognostic marker for child-pugh A hepatocellular carcinoma after operation. J Gastroenterol 2009;44:635–42. 23. Zeng ZC, Fan J, Tang ZY, et al. Prognostic factors for patients with hepatocellular carcinoma with macroscopic portal vein or inferior vena cava tumor thrombi receiving external-beam radiation therapy. Cancer Sci 2008;99:2510–7. 24. Hasegawa K, Kokudo N, Imamura H, et al. Prognostic impact of anatomic resection for hepatocellular carcinoma. Ann Surg 2005; 242:252–9. 25. Tanaka K, Shimada H, Matsumoto C, et al. Anatomic versus limited nonanatomic resection for solitary hepatocellular carcinoma. Surgery 2008;143:607–15. 26. Poon RT, Fan ST, Ng IO, Wong JW. Significance of resection margin in hepatectomy for hepatocellular carcinoma. Ann Surg 2000;231: 544–51.