Data on liver resection for hepatocellular carcinoma (HCC) without cirrhosis are sparse. The present study was conducted to evaluate the indications and results of liver resection for HCC with regard to safety and efficacy.
Data on liver resection for hepatocellular carcinoma (HCC) without cirrhosis are sparse. The present study was conducted to evaluate the indications and results of liver resection for HCC with regard to safety and efficacy.
Data for patients who had liver resection for HCC without cirrhosis between January 1996 and March 2011 were retrieved retrospectively using a prospective database containing information on all patients who underwent hepatectomy for HCC. Patient and tumour characteristics were analysed for influence on overall and disease-free survival to identify prognostic factors by univariable and multivariable analysis.
The 1-, 3- and 5-year overall survival rates after resection with curative intent for HCC without cirrhosis were 84, 66 and 50 per cent respectively. Disease-free survival rates were 69, 53 and 42 per cent respectively. The 90-day mortality rate was 4·5 per cent (5 of 110 patients). Surgical radicality and growth pattern of the tumour were independent prognostic factors for overall survival. Disease-free survival after resection with curative intent was independently affected by growth pattern and by the number and size of tumour nodules.
Liver resection for HCC without cirrhosis carries a low perioperative risk and excellent long-term outcome if radical resection is achieved. Copyright © 2012 British Journal of Surgery Society Ltd. Published by John Wiley & Sons, Ltd.
Hepatocellular carcinoma (HCC) is a leading cause of cancer-related death worldwide and the burden of this devastating cancer is expected to increase in the coming years1. Owing to the high prevalence of hepatitis B infection and dietary exposure to aflatoxin, the highest incidence of HCC is observed in East Asia and sub-Saharan Africa2–4. However, in Western countries a notable increase in the incidence of HCC has been observed in recent decades, associated mostly with hepatitis C infection and alcoholic liver disease5. On the other hand, in a considerable number of patients HCC arises without underlying liver disease, frequently diagnosed accidentally at a locally advanced tumour stage6.
Multiple treatment modalities have been established for HCC over the past decades, and surgery—liver resection or transplantation—affords the best chance of long-term survival and cure. Liver transplantation is established as an effective and safe treatment for patients with HCC in cirrhosis, but many patients are not suitable candidates. For these patients, resection or ablation can serve as an alternative treatment7. Liver resection is currently considered the treatment of choice for patients with HCC without cirrhosis. Owing to generally advanced local growth of HCC in patients without cirrhosis, ablative approaches, such as radiofrequency ablation, are usually not applicable8, 9. Data on the outcome of liver resection for HCC are necessary to determine indications and suitable selection criteria, but are sparse for this patient group.
The present study was performed to evaluate the results of liver resection with regard to safety and efficacy as a treatment for HCC in patients without cirrhosis. Patient and tumour characteristics were analysed for their influence on overall and disease-free survival to identify prognostic factors that may serve as selection criteria for resection.
A prospectively collected institutional database was studied retrospectively to retrieve all patients who underwent hepatic resection for HCC between January 1996 and March 2011. From this data set, all patients were selected who had liver resection for HCC without cirrhosis. Hepatectomy was performed through a right-sided upper abdominal incision with additional upper abdominal median extension. Parenchymal transection was performed with the cavitron ultrasonic surgical aspirator (CUSA®; Valleylab, Boulder, Colorado, USA). During transection the central venous pressure was 5 cmH2O or less, and a Pringle manoeuvre was not performed routinely. Biliary drainage was used when reconstruction of the bile ducts was necessary.
Only patients with tumours with exclusive hepatocellular differentiation, according to the tumour node metastasis (TNM) classification10, were considered for the analysis. Patients with tumours of mixed hepatocellular and cholangiocellular differentiation, and those with fibrolamellar hepatocellular carcinoma were not included in the study. The diagnostic work-up included recent (less 6 months previously) gastroscopy and total colonoscopy as well as contrast-enhanced abdominal computed tomography and/or magnetic resonance tomography. Female patients had a pelvic examination if they had not had one recently. Evaluation of the lungs or thyroid was performed if indicated. Before resection, screening for hepatitis A, B and C virus infection was performed, and α-fetoprotein (AFP), carcinoembryonic antigen and cancer antigen 19-9 levels were measured in all patients.
During surgery, assessment was completed by ultrasonography of the liver. Regional lymphadenectomy was performed for accurate determination of the nodal status and to avoid insufficient surgical treatment in the event that pathological evaluation might reveal mixed or cholangiocellular differentiation of the tumour. Lymphadenectomy was performed along the common hepatic artery down to the coeliac trunk and within the hepatoduodenal ligament. Frozen-section biopsies of lymph nodes were taken routinely, but more for orientation regarding disease extent than to dictate the further course of the operation. Para-aortic lymph nodes were dissected only when infiltrated macroscopically by tumour and when lymph nodes around the coeliac trunk were positive. Surgical procedures aimed to achieve curative resection. Major resection was defined as excision of three or more liver segments, and minor resection as excision of less than three segments. Staging was performed according to the seventh edition of the International Union Against Cancer (UICC) classification of malignant tumours10.
Patients were followed up according to the guidelines of the German Cancer Society by physical examination, conventional radiography of the chest, abdominal ultrasonography and AFP serum assays every 6 months.
Surviving patients were contacted by telephone or in person. If necessary, information on patient survival and recurrence was obtained from the general practitioner.
Overall and disease-free survival were calculated from the date of resection. Comparison of categorical and continuous variables was performed using the χ2 test and Wilcoxon test respectively. Survival data were compared with the log rank test. Variables with a significant influence on survival in the univariable analysis were entered into multivariable analysis using a step-by-step Cox regression model. Survival analysis and univariable analysis were undertaken with the SPSS® software program version 20 (IBM, Armonk, New York, USA), which was also used to generate Kaplan–Meier curves. Differences were considered significant when P < 0·050.
Between January 1996 and March 2011, 110 patients underwent liver resection for HCC without cirrhosis at a tertiary referral centre. A further 26 patients had liver resection for HCC with cirrhosis. During the same period, a total of 725 liver resections were performed for malignant primary and secondary liver tumours, resulting in a proportion of HCC without cirrhosis of 15·2 per cent (Table 1).
|No. of patients (n = 110)|
|Age at surgery (years)*||68 (35–91)|
|Sex ratio (M:F)||85:25|
|No. of tumour nodules|
|Largest tumour diameter (cm)*||7 (0·8–28)|
|Lymph vessel infiltration|
|Tumour capsule infiltration|
|Pathological tumour stage|
|Ligamental lymph nodes|
|Preop. α-fetoprotein level (ng/ml)*||10 (0–62 315)|
The diagnosis of HCC was made due to symptoms in 56 (50·9 per cent) of the 110 patients. The major complaints were upper abdominal tension (33 patients), weight loss (27), lack of appetite (19) and/or drop in performance (19). Underlying liver disease was detected in seven patients, six of whom had hepatic steatosis in 10–30 per cent of hepatocytes. Four patients had additional liver fibrosis (stage F1 according to the Scheuer classification11) and one patient had fibrosis without steatosis (stage F2). The liver damage in these seven patients was caused by alcoholic steatohepatitis in five patients and by non-alcoholic steatohepatitis in two.
For tumour downsizing, eight patients with large tumours adjacent to central vascular or biliary structures received transarterial chemoembolization before resection. In two patients who underwent hepatectomy in the 1990s, ethanol injection into the HCC was performed before resection. No patient received neoadjuvant chemotherapy and/or radiotherapy. Five patients received adjuvant chemotherapy after non-curative resection.
The surgical procedures in the study were hemihepatectomy in 41 patients (31 right, 10 left), extended hemihepatectomy in 25 patients (18 right, 7 left), a right trisectionectomy in one patient, plurisegmentectomies in nine patients, left lateral resections in six, segmentectomies in 13 and atypical resection in 15 patients. Thus, 60·9 per cent (67 of 110) were major liver resections.
The median follow-up for all patients was 20·3 (range 0·2–164·2) months.
Overall 1-, 3- and 5-year survival rates for the entire study population were 77·7, 60·0 and 46·6 per cent respectively. Curative (R0) resection was achieved in 98 patients (89·1 per cent). Microscopically positive margins (R1 resection) or gross residual disease (R2) were detected in nine (8·2 per cent) and three (2·7 per cent) patients respectively. Overall survival rates differed significantly with respect to surgical radicality (P < 0·001). The overall 1-, 3- and 5-year survival rates after R0 resection were 84, 66 and 50 per cent respectively. Respective rates after R1 resection were 29, 14 and 14 per cent respectively, whereas no patient with gross residual disease was alive at 1 year after resection. Analysis of the different time periods did not reveal significant differences in the overall survival rates between time frames (data not shown).
In the 98 patients who had an R0 resection, primary tumour stage had an effect on survival (P < 0·001). The 1-, 3- and 5-year overall survival rates of the 76 patients with T1 and T2 tumours were 90, 76 and 61 per cent respectively, whereas the respective rates in the 22 patients with T3 and T4 tumours were 63, 34 and 20 per cent. In addition, invasion of lymphatic vessels by tumour cells had a strong impact on overall survival after resection: 1-, 3- and 5-year overall survival rates of 76 patients with HCC without invasion of lymphatic vessels were 89, 72 and 57 per cent respectively, whereas 22 patients with tumours with invasion of lymphatic vessels had 1- and 3-year survival rates of 64 and 40 per cent respectively, with no patient alive at 5 years after resection (P = 0·001) (Fig. 1a). A significant influence on overall survival after curative resection of HCC without cirrhosis was also found for the growth pattern of the tumour: 1-, 3- and 5-year overall survival rates after resection of HCC in 70 patients with a nodular growth pattern were 84, 75 and 57 per cent respectively, compared with 84, 39 and 31 per cent in 28 patients with tumours with an invasive or diffuse growth pattern (P = 0·043).
Patients who died after surgery were excluded from the analysis of prognostic factors for overall survival after liver resection for HCC without cirrhosis. In the univariable analysis, multicentric tumours (P = 0·002), tumour diameter greater than 5 cm (P = 0·007), positive resection margin (P < 0·001), lymph and blood vessel infiltration by the tumour (P < 0·001 and P = 0·001 respectively), a diffuse or infiltrative growth pattern of the tumour (P = 0·001), primary tumour stage T3 or T4 (P < 0·001) and lymph node metastasis within the hepatoduodenal ligament (P = 0·027) were predictors of poor long-term survival (Table S1, supporting information). In the multivariable analysis, only surgical radicality (P = 0·025) and growth pattern of the tumour (P = 0·003) were identified as independent prognostic factors for overall survival after liver resection for HCC without cirrhosis (Table 2).
|No. of tumour nodules (solitary versus multicentric)||1·40 (0·16, 12·10)||0·759|
|Largest tumour diameter ( ≤ 5 cm versus > 5 cm)||2·35 (0·84, 6·53)||0·102|
|Surgical radicality (R0 versus R1/2)||2·18 (1·10, 4·30)||0·025|
|Vascular infiltration (yes versus no)||0·88 (0·44, 1·78)||0·725|
|Growth pattern (diffuse/infiltrative versus nodular)||0·38 (0·20, 0·71)||0·003|
|Pathological tumour stage (T1/2 versus T3/4)||3·13 (0·33, 29·89)||0·322|
|Lymph node metastasis (yes versus no)||0·73 (0·29, 1·85)||0·503|
|Lymph vessel infiltration (yes versus no)||1·90 (0·81, 4·45)||0·139|
Excluding postoperative death within 90 days, overall 1-, 3- and 5-year disease-free survival rates after curative liver resection for hepatocellular carcinoma without cirrhosis were 69, 53 and 42 per cent respectively. Primary tumour stage had a significant influence on disease-free survival: 1-, 3- and 5-year disease-free survival after curative resection in 73 patients with T1 and T2 tumours were 79, 59 and 51 per cent respectively, compared with 36, 31 and 15 per cent in 21 patients with T3 and T4 tumours. The growth pattern of the HCC had a significant impact on disease-free survival. Sixty-six patients with tumours of nodular growth pattern had 1-, 3- and 5-year disease-free survival rates of 77, 64 and 52 per cent respectively, whereas 28 patients with diffuse or infiltrative growing tumours had respective rates of 50, 25 and 17 per cent (P < 0·001). The number of tumour nodes had a significant effect on disease-free survival: 67 patients with solitary HCC had 1-, 3- and 5-year disease-free survival rates of 80, 62 and 53 per cent respectively, compared with 43, 31 and 19 per cent in 27 patients with multicentric HCC (P < 0·001). In addition, infiltration of lymphatic vessels by tumour cells had a strong influence on disease-free survival rates: 1-, 3- and 5-year rates in 75 patients with HCC without lymphatic vessel infiltration were 72, 58 and 46 per cent respectively, whereas the 1- and 3-year disease-free survival rates after curative resection of tumours with infiltration of lymphatic vessels (19 patients) were 55 and 27 per cent respectively, with no patient yet surviving recurrence-free for more than 4 years after resection (P = 0·026) (Fig. 1b).
Postoperative deaths within 90 days were excluded from the analysis of prognostic factors for disease-free survival after curative resection of HCC without cirrhosis (94 patients). In the univariable analysis, number of tumour nodes (P < 0·001), largest tumour diameter (P = 0·017), growth pattern of the tumour (P < 0·001), primary tumour stage (P < 0·001) and lymph vessel infiltration (P = 0·026) significantly affected disease-free survival (Table S2, supporting information). The multivariable analysis identified number of tumour nodules (P = 0·032), largest tumour diameter (P = 0·029) and growth pattern (P < 0·001) as independent prognostic factors for disease-free survival (Table 3).
|No. of tumour nodules (solitary versus multicentric)||0·23 (0·06, 0·88)||0·032|
|Largest tumour diameter ( ≤ 5 cm versus > 5 cm)||3·03 (1·12, 8·18)||0·029|
|Growth pattern (diffuse/infiltrative versus nodular)||0·26 (0·14, 0·49)||< 0·001|
|Pathological tumour stage (T1/2 versus T3/4)||0·67 (0·16, 2·93)||0·597|
|Lymph vascular infiltration (yes versus no)||1·71 (0·81, 3·60)||0·156|
The postoperative 90-day mortality rate after liver resection for HCC without cirrhosis was 4·5 per cent (5 of 110 patients). Four of the five patients developed liver failure with subsequent multiple organ failure after major liver resection; one patient died from pulmonary embolism after a minor resection. In the last 4 years there were no postoperative deaths. Postoperative complications developed in 34 patients (30·9 per cent). The most frequent complications were bilioma (15 patients), wound infection (12), pleural effusion (11) and pneumonia (5).
In the whole study population, tumour recurrence was detected in 54 patients (49·1 per cent) within the follow-up period. After curative resection, 45 (46 per cent) of the 98 patients developed recurrent disease. The recurrence was localized most frequently within the liver, lungs and/or bones. Intrahepatic tumour recurrence was treated by repeat liver resection in 13 patients, and three had resection of extrahepatic metastases. Most patients with unresectable recurrence received palliative chemotherapy. In the last 5 years, sorafenib was the most frequent chemotherapeutic agent used for palliative chemotherapy.
Surgical therapy—liver resection or transplantation—remains the ‘gold standard’ in the curative treatment of HCC8. Liver resection is the treatment of choice for patients with HCC without cirrhosis, and provides excellent results. In the present study, the overall survival rate of patients with HCC without cirrhosis was 50 per cent at 5 years after curative resection, comparable with the results of previous studies12–15. In early tumour stages the results achieved by liver resection are even better. The 5-year overall survival rate after hepatectomy for solitary tumours or small multicentric HCCs exceeded 60 per cent in the present study. Similar long-term survival cannot be achieved by non-surgical treatment16. Liver resection also achieved long-term survival and cure for patients with locally advanced HCCs in the present study, and the 5-year overall survival rate after resection of large multicentric HCCs exceeded 20 per cent. In keeping with this, in a study from Japan the 5-year survival rate after resection of HCC with invasion of major branches of the portal vein or the inferior vena cava was also 20 per cent17. Further, in an analysis of 53 patients with HCC larger than 10 cm in diameter, the survival rate 5 years after resection was 35 per cent, significantly superior to that among non-surgically treated patients18. In addition, a study from Italy reported a 5-year survival rate after resection of large solitary and multicentric HCCs of 56·1 and 33·6 per cent respectively19. Importantly, long-term survival and cure are observed only rarely with any other treatment, and even with innovative approaches, such as selective internal radiation therapy, long-term survival is achieved only exceptionally20. This emphasizes the central role of liver resection in treating HCC without cirrhosis and indicates that in patients with resectable tumours a liver resection should be aimed for.
Owing to increasing life expectancy, an increasing number of patients now develop HCC at an advanced age and adequate treatment for these patients is necessary21. Liver resection is presently a safe treatment modality that can be performed with low perioperative morbidity and mortality rates below 5 per cent15, 22, as also shown in the present study. Liver resection is also safe in elderly patients23 and, therefore, age is not a criterion for excluding patients with HCC from liver resection.
Suitable criteria for the selection of patients with HCC for resection are important to determine the best individual treatment with the highest therapeutic benefit for each patient. In the present study, the growth pattern of HCC was shown to have an important influence on overall and disease-free survival. HCCs with a nodular growth pattern show a clear boundary between tumour and liver parenchyma, whereas those with an infiltrative or diffuse growth pattern have an invasive character at the border with the surrounding liver. Thus, it would appear that HCCs with a nodular growth pattern are less aggressive, resulting in a better prognosis after curative resection, irrespective of primary tumour stage. This is in keeping with experience of solitary large HCCs, which have a relatively good prognosis independent of primary tumour size, and have been presumed to be a new subtype of HCC24. Thus, the prognostic significance of growth pattern in the present study supports the importance of biological markers for prognosis, and indicates that suitable biological prognostic markers may improve the selection of patients for surgical treatments.
Invasion of lymphatic vessels and spread to locoregional lymph nodes were observed frequently in the present study. This may result from the systematic evaluation of lymph vessels during the pathological examination of the tumours and routine regional lymphadenectomy during resection. Interestingly, the role of lymphatic infiltration for progression of HCC has hardly been investigated to date. In the present study, tumour cell infiltration of lymphatic vessels had a significant effect on overall and disease-free survival that has not been described previously. Even in recently published large cohorts of patients with HCC without cirrhosis who underwent resection, the prognostic importance of lymph vessel infiltration was not analysed15, 22. Recent results indicate that tumour-associated lymphangiogenesis could play an important role in the progression of HCC. In an experimental setting it was shown that the expression of a prolymphangiogenic growth factor increases the intrahepatic and extrahepatic spread of HCC by induction of tumour-associated lymphangiogenesis25. Furthermore, a high density of intratumoral lymph vessels was shown to be an independent prognostic factor for poor disease-free survival in human HCC26. Therefore, invasion of tumour cells in lymphatic vessels may not only be an important prognostic factor in HCC, but targeting tumour-associated lymphangiogenesis may also represent a suitable strategy for developing new therapeutic approaches. In fact, blocking of prolymphangiogenic vascular endothelial growth factor D was shown to slow down progression of HCC in an experimental setting25. However, clinical experience with such approaches is lacking. Thus, the clinical importance of lymph vessel infiltration may currently mainly be in its role as a factor affecting survival.
Although some tumour characteristics have significant influence on the prognosis after resection, long-term survival and cure can be achieved by hepatectomy in a considerable number of patients, even if the tumours show these prognostically relevant factors. Therefore, against the background of the limited efficacy of non-surgical treatments16, the presence of these characteristics may not exclude patients from curative resection. Thus, most of the above clinicopathological characteristics do not represent suitable criteria by which to select patients with HCC without cirrhosis for resection. In contrast, surgical radicality was an independent prognostic factor for survival in the present study. Patients who had a non-curative resection rarely survived for more than 1 year. Therefore, surgical radicality is decisive for the prognosis, and the feasibility of a curative resection represents the most relevant criterion for selecting patients for liver resection.
In patients with HCC in underlying cirrhosis, liver transplantation is considered to be the best treatment option, at least in early tumour stages27. Liver transplantation achieves excellent long-term results with low recurrence rates and 5-year survival rates of around 80 per cent in early stages of HCC in cirrhosis, defined by the Milan criteria27. Furthermore, liver transplantation provides the opportunity not only to remove the tumour but also to cure the underlying cirrhosis, eliminating the risk of future de novo HCCs in the cirrhotic liver remnant. Owing to low perioperative mortality rates and efficient immunosuppressive protocols with few side-effects, liver transplantation is an effective and safe treatment for patients with HCC in cirrhosis27. However, as a consequence of advanced tumour stage or age, persistent liver injury or severe co-morbidity, many patients are not suitable candidates for liver transplantation16. Alternative treatment modalities are necessary in these patients, in whom liver resection may be curative. In addition, in patients with large and multiple tumour nodes, long-term survival can be achieved by liver resection despite a higher risk of recurrence28, 29. However, the perioperative risk is considerable in patients with liver cirrhosis, because of the insufficient remnant liver function. In particular, patients with portal hypertension are at risk of developing postoperative liver failure, and this excludes many patients with HCC in cirrhosis from having a resection30, 31. Therefore, only a limited number of patients with HCC in cirrhosis are suitable candidates for resection, as found in the present study. Strategies to improve the function of the future liver remnant after resection represent a crucial step in a broader application of hepatectomy in patients with HCC in cirrhosis who are not suitable for transplantation.
The most important limitation of the present study is the size of the study population. Although the present cohort represents one of the largest series from Western countries published to date, the number of patients remains limited. In particular, in some subgroups with very small numbers the outcome of single patients may have had a significant influence on the statistical analysis, resulting in overestimation or underestimation of the prognostic importance of some of the patient or tumour characteristics that frequently serve as selection criteria for resection.
Additional supporting information may be found in the online version of this article:
Table S1 Univariable analysis of prognostic factors for overall survival after liver resection for hepatocellular carcinoma without cirrhosis (Word document)
Table S2 Univariable analysis of prognostic factors for disease-free survival after curative resection for hepatocellular carcinoma without cirrhosis (Word document)
Please note: John Wiley & Sons Ltd is not responsible for the functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
A.T. and C.B. are joint first authors of this article.
Disclosure: The authors declare no conflict of interest.