High pathological risk of recurrence after surgical resection for hepatocellular carcinoma: An indication for salvage liver transplantation


  • Margarita Sala,

    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Josep Fuster,

    1. Surgery Department, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Josep M. Llovet,

    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Miquel Navasa,

    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Manel Solé,

    1. Pathology Department, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • María Varela,

    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Fernando Pons,

    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Antoni Rimola,

    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Juan Carlos García-Valdecasas,

    1. Surgery Department, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Concepció Brú,

    1. Radiology Department, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    Search for more papers by this author
  • Jordi Bruix

    Corresponding author
    1. Liver Unit, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic, University of Barcelona, Catalonia, Spain
    • BCLC Group, Liver Unit, Hospital Clínic i Provincial, Villarroel 170, 08036-Barcelona, Catalonia, Spain
    Search for more papers by this author
    • Telephone: 34-3-2279803; FAX: 34-93-2275792


Surgical resection and liver transplantation offer a 5-year survival greater than 70% in patients with hepatocellular carcinoma, but the high recurrence rate impairs long-term outcome after resection. Pathological data such as vascular invasion and detection of additional nodules predict recurrence and divide patients into high and low risk profile. Based on this, we proposed salvage liver transplant to resected patients in whom pathology evidenced high recurrence risk even in the absence of proven residual disease. From January 1995 to August 2003 we have evaluated 1,638 patients. Resection was indicated in 77 patients, but only 17 (22%) (all cirrhotics, 14 hepatitis C virus+) were optimal candidates for both resection and transplantation. Of them, 8 exhibited a high risk profile at pathology and were offered transplantation. Among the 8 high risk patients, 7 presented recurrence, compared with only 2 of the 9 at low risk (P = .012). Two of the high risk patients refused transplant and developed multifocal disease during follow-up. The other 6 were enlisted and all but 1 had tumor foci in the explant. Only 1 presented extrahepatic dissemination early after transplant and died 4 months later. The others are free of disease after a median follow-up of 45 months. Two recurrences were detected in low risk patients, 1 of them being transplanted 18 months after surgery. These data in a small series of patients confirm that pathological parameters identify patients at higher risk of recurrence, which allow them to be listed for liver transplantation without proven malignant disease. In conclusion, this policy is clinically effective and could further improve the outcome of resected patients. (Liver Transpl 2004;10:1294–1300.)

Hepatocellular carcinoma (HCC) is the 5th most common cancer worldwide and the 3rd most common cause of cancer-related death.1 Radical treatments for early hepatocellular carcinoma are surgical resection, liver transplantation (LT), and percutaneous treatments.2 However, there are no randomized control trials (RCTs) comparing these treatments, and the best option depends on the results obtained in observational studies. LT is the best option in patients with decompensated cirrhosis and single HCC ≤ 5 cm or showing up to 3 nodules, each of them ≤ 3 cm. However, there is a major controversy in cirrhotic patients with preserved liver function and solitary tumors. Survival after surgical resection in Child Pugh A patients without significant portal hypertension and normal bilirubin is similar to that obtained with LT,3 and the main difference lies in the higher tumor recurrence rate after resection—above 70% at 5 years vs. 15% after LT.3–5 This is the major argument used by some authors to support LT as the first treatment option.

Several studies have shown that the risk of recurrence might be predicted by the presence of microvascular invasion or additional nodules.3, 6–10 and, therefore, both parameters could be used to divide already resected patients into those with low recurrence risk and those with high risk. Interestingly, the recurrence rate in transplanted patients who present a high risk profile in the explanted liver is slightly increased. However, it is not prohibitive,3, 4, 11, 12 and this suggests that this subgroup of patients would have been better served by transplantation.

In our Unit, surgical resection is the first option offered to HCC patients without significant portal hypertension and normal bilirubin. Following the above reasoning, we proposed in 1995 that it could be worthwhile to propose enlisting for LT to those patients with solitary HCC who were treated by surgical resection and in whom the pathological assessment evidenced parameters of a high risk of recurrence. This novel policy differs from that applied by most of the groups who follow resected patients and consider salvage transplantation upon detection of recurrence.13–19 The present study describes the results we have obtained with this strategy.


HCC, hepatocellular carcinoma; LT, liver transplantation.

Patients and Methods

Between January 1995 and August 2003, 1,638 patients with HCC were diagnosed, staged, and treated in our Liver Unit following a previously published schedule.20 Cirrhotic patients with early HCC (single tumors ≤ 5 cm or 3 nodules ≤ 3 cm each) are considered for radical therapies. Resection is indicated for patients with single tumors, absence of significant portal hypertension, and normal bilirubin. Patients with significant portal hypertension, abnormal bilirubin, or 3 nodules ≤ 3 cm are considered for LT (if younger than 65 years old and without severe associated diseases). Percutaneous treatments are applied when surgery is precluded. For patients in an intermediate stage, the first option is arterial chemoembolization, and those diagnosed at an advanced stage are considered for phase II-IV trials. Finally, end stage patients (Okuda III,21 Performance status 3-422) receive symptomatic treatment.

During this period of time, 77 cirrhotic patients with HCC were suitable for surgical resection, but only 17 (22%) were optimal candidates for both resection (first option) and LT (≤ 65 years old without severe associated conditions). These patients were offered resection as a first option and constitute the population of this cohort analysis to assess the efficacy of the above mentioned salvage strategy.

Sixteen patients were male, mean age was 55 ± 7 years, and the etiology of underlying cirrhosis was hepatitis C virus in 14 cases. All patients had a preserved liver function (all belonged to Child Pugh A group) and did not have significant portal hypertension. According to preoperative staging, all had single tumors that in 9 patients were ≤ 30 mm and in 8 ranged between 30 and 50 mm. The mean tumor size was 30 ± 9 mm (range 15–50 mm) (Table 1).

Table 1. Characteristics of the Patients
VariablesOverall (n = 17)Low Risk (n = 9)High Risk (n = 8)P
  1. Abbreviations: HCV, hepatitis C virus; HBV, hepatitis B virus; AFP, alpha-fetoprotein; AST, aspartate aminotransferase; ALT, alanine aminotransferase; HVPG, hepatic venous pressure gradient; ns, not statistically significant.

  2. NOTE: Numbers expressed as mean ± standard deviation.

Age (years)55 ± 754 ± 8.556 ± 5ns
Gender (M/F)16/18/18/0ns
Etiology of cirrhosis   ns
AFP (ng/mL; <10/11 − 100/101 − 400/>400)10/ 1/ 5/ 16/ 1/ 2/ 04/ 0/ 3/ 1ns
Bilirubin (mg/dL).9 ± .1.9 ± .1.8 ± .1ns
Prothrombin activity (%)88 ± 1089 ± 1188 ± 10ns
Albumin (g/dL)43 ± 443 ± 343 ± 4ns
AST (IU/L)67 ± 4073 ± 5161 ± 24ns
ALT (IU/L)100 ± 65107 ± 8492 ± 36ns
HVPG (mmHg)6.3 ± 2.55.5 ± 2.57.5 ± 2ns
Mean tumor size (mm)30 ± 929 ± 1230 ± 6ns
 ≤30 mm954 
 31–50 mm844 
Pathologic characteristics    
 Mean tumor size (mm)32 ± 1329 ± 1335 ± 12ns
 Differentiation degree   ns
 Microvascular invasion   As per design
  Yes7/ 107 
 Additional nodules   As per design

Diagnosis of HCC was performed by needle biopsy in 12 cases and by noninvasive criteria in 5 cases (2 coincidental imaging techniques or 1 imaging technique with increased alpha-fetoprotein). Preoperative staging included abdominal ultrasound, dynamic computed tomography, and/or magnetic resonance imaging. Additionally, a preoperative hemodynamic study with measurement of hepatic venous pressure gradient was also performed in order to exclude patients with significant portal hypertension (hepatic venous pressure gradient ≥ 10 mmHg).

Surgical technique included intraoperative ultrasound to exclude additional nodules, localize the tumor, and perform an anatomical resection.

Resected liver specimens were serially sliced in .5 cm thick slices and fixed in formalin. Representative samples of tumor, nontumoral tissue, and surgical margins were embedded in paraffin for microscopic examination. Tumor number and size were confirmed on gross inspection, and microscopic analysis determined the presence of vascular invasion, microscopic tumor satellites, tumor differentiation, and status of the resection margin.

According to the pathological criteria, the patients were divided into 2 groups: patients with high risk of recurrence if they had microvascular invasion and/or additional nodules or satellites and patients with low risk of recurrence if they did not have any of these parameters (Table 1). Patients of the high risk group were offered enlistment for liver transplantation even in the absence of tumoral disease and followed every 3 months after enlistment by means of clinical examination, alpha-fetoprotein, abdominal ultrasound and computed tomography scan. Patients of the low risk group were advised to attend regular follow-up every 6 months. Additional diagnostic techniques were performed upon suspicion of recurrence to confirm malignancy, stage the disease, and indicate treatment, which is based on the same strategy as depicted for the primary tumor.

Statistical Analysis

Baseline characteristics of the patients are expressed as mean ± SD. Comparison between groups was done by using the Student's t-test for quantitative variables, and the χ2 test or the Fisher test for qualitative variables. Follow-up length is expressed as median (range). Follow-up was computed as starting from the resection date for all patients and was maintained until death or last visit before December 15, 2003.

The calculations were done by the SPSS package (SPSS 10.0, 1989–1995, Chicago, IL).


According to the pathological study of the resected specimen, patients were divided into 2 groups: patients with high risk of recurrence (n = 8) if they had additional nodules and/or microvascular invasion and patients with low risk of recurrence when these 2 parameters were absent (n = 9) (Fig. 1). The characteristics of both groups are depicted in Table 1. The resection border was tumor-free in all cases.

Figure 1.

Results of the proposed decision policy for HCC patients treated by surgical resection after stratifying them according to the risk of recurrence based on pathology. TACE, transarterial chemoembolization; PEI, percutaneous ethanol injection; mo, months.

High Risk Patients

Amongst the 8 patients with high risk of recurrence, 6 patients accepted to be enlisted for LT and 2 refused. Of the 6 patients enlisted, 2 developed tumor recurrence while waiting. One patient presented a single tumor at 12 months, and percutaneous ethanol injection was applied as adjuvant treatment. This patient was transplanted 4 months after. The other patient developed multifocal tumor recurrence 6 months after resection and is currently still waiting for LT. The resected specimen of this patient showed a single HCC 40 mm in size, with microvascular invasion and without additional nodules. Recurrence was not suspected prior to LT in the other 4 patients enlisted because of risk (median time from resection to LT: 12 months, range 9–19). Median time elapsed between imaging studies and LT was 53 days (range 14–93).

Characteristics of the explanted liver in the 5 high risk patients that have been transplanted are depicted in Table 2. Only 1 patient had no tumoral recurrence in the explanted liver. This patient was transplanted 9 months after resection and presented both adverse pathological findings. The remaining 4 patients presented tumor recurrence. The patient whose tumor recurrence was treated by percutaneous ethanol injection showed viable tumor in the treated foci and 1 additional tumor nest. The other 3 patients were transplanted without any evidence of tumor recurrence, but the explanted liver showed recurrent tumor foci less than 2 cm, associated with microvascular invasion in 2 patients.

Table 2. Characteristics of the High Risk Transplanted Patients
No. PatientResected HCCTime (months)*Explanted LiverRecurrenceFollow-Up (months)
Size (mm)Vascular InvasionAdditional NodulesTumorSize (mm)Vascular InvasionAdditional Nodules
  • Abbreviations: HCC: hepatocellular carcinoma.

  • *

    Time from resection to liver transplantation.

  • Recurrence after liver transplantation.

  • Time from liver transplant to the end of follow-up.

  • §

    Patient with tumor recurrence after liver transplantation who died 4 months after.

  • Patient treated percutaneously before liver transplantation.

  • Patient with high risk of recurrence in whom explanted liver did not show tumor recurrence. This patient died 84 months after liver transplantation due to recurrent HCV cirrhosis.

  • NOTE: The sixth patient is still awaiting liver transplantation.


At the end of follow-up (median follow-up 26 months, range 4–84), 2 of the transplanted patients had died. One patient (the 1 without tumor nests in the explanted liver) died 7 years after LT because of recurrent hepatitis C virus cirrhosis without HCC recurrence. The other patient, whose explanted liver showed multiple foci of HCC and microvascular invasion, died 4 months after LT due to peritoneal dissemination of HCC and intrahepatic spread.

Finally, the 2 patients who refused enlistment for LT developed a large multinodular recurrence not amenable for radical therapies at 4.5 and 50 months, respectively. They died at 10 and 18 months after recurrence due to tumor progression.

Low Risk Patients

According to our strategy, these 9 patients were not offered to be enlisted for LT after resection because of their low recurrence risk. However, 2 of them had recurrence during follow-up. One patient, initially having a solitary HCC of 4 cm encapsulated and well differentiated, developed a multinodular HCC 19 months after resection and was treated by transarterial chemoembolization. He is alive 55 months after resection. The 2nd patient, operated on because of a solitary HCC of 2 cm without capsule and moderately differentiated, developed a solitary tumor recurrence and was enlisted 18 months after resection. LT was performed 9 months later and he is alive and free of recurrence 28 months after LT (55 months after resection).

The remaining 7 patients are alive and without HCC recurrence with a median follow-up of 55 months (range 19–103).


Tumor recurrence is the major drawback after surgical resection of HCC, and this is used by some authors to support LT as the first therapeutic option for these patients.19, 23, 24 However, transplantation applicability and outcomes are curtailed by the shortage of donors. This creates a waiting time during which the HCC may progress and, when analyzed according to intention to treat the survival of the patients, might be less than that offered by surgical resection.3, 25, 26 Because of this, most of the groups consider resection as the first treatment approach for patients who would be candidates for both resection and LT.13–18, 20 In such a scenario, the main issue is how to manage the risk of recurrence and to treat its development. Despite some encouraging results with Interferon,27 acyclic retinoids,28 radiation,29 and adoptive immunotherapy,30 there is no effective therapy to prevent tumor relapse.31 Thus, in the majority of centers the established clinical practice is to carefully follow the resected patients and, upon detection of recurrence, consider the potential indication of so-called salvage LT.15 This approach has been heavily criticized for years because initial studies suggested that LT after resection would offer poorer results,19 but the data recently published by Belghiti et al.13 indicate that the survival after salvage LT is not significantly lower. The explanation for this discrepancy might be related to the biological selection process that the patients have undergone until reaching LT. This does not refer to a better identification of candidates by imaging techniques in order to exclude desperate patients with extensive multifocal recurrence, but to the fact that only those patients with a less aggressive recurrence would become candidates for salvage LT. It is well known that there are 2 major pathways leading to recurrence: tumor dissemination prior to operation and de novo tumor development in an oncogenic cirrhotic liver.32–40

In most cases, recurrence will be related to the first mechanism (dissemination) and will very likely appear early during follow-up as a multifocal involvement.32, 39, 41–43 This dissemination nature of recurrence and its faster progression mean that patients will not be candidates for salvage surgery. In fact, if enlisted upon detection of recurrence, they will experience a higher rate of exclusion while waiting, and even if they do reach transplantation, their outcome will be dismal.

By contrast, those infrequent patients in whom dissemination has not taken place will most likely develop solitary de novo tumors and in that way become candidates for successful salvage therapy. While the retrospective analysis of the database by Poon et al. suggests that the majority of patients with recurrence could become candidates for salvage LT, this is not supported at all by the few available data in clinical practice3, 13, 14, 16, 18, 19, 44 and even by their own previous publications.39, 45 As a whole, less than 20% of the patients are candidates for salvage surgery,46–56 and this low applicability was the major criticism of the Markov analysis on salvage LT performed by Majno et al. who considered an 80% applicability rate upon recurrence detection.15

According to all these comments, it is clear that the policy to wait for recurrence to develop and then indicate treatment is less than optimal. The risk of recurrence can be accurately predicted by pathologic examination of the resected tissue. Microvascular invasion and presence of satellites or additional intrahepatic neoplastic sites are thought to be related to unrecognized tumor spread prior to resection, and interestingly, those transplanted patients in whom pathology examination depicts these pathologic high risk parameters do not present a prohibitive rate of disease recurrence during follow-up.11, 12 Therefore, for these patients LT should be considered the best primary treatment.

Following this reasoning, we proposed a more active attitude offering enlistment for LT to those patients who after initial HCC resection would prove to bear this pathologic high risk profile. The results of this study demonstrate the efficacy of this novel policy. During the period of the investigation, we operated on 77 patients and 17 of them qualified as candidates both for resection and LT. According to our treatment strategy, they were offered resection as primary treatment, and based on the pathologic findings, 8 patients were classified at high risk and 9 at low risk. All but 1 of those at high risk showed recurrence either prior to LT (n = 2) or in the explanted liver in the absence of disease on imaging techniques (n = 3). Unfortunately, 2 patients did not accept to be enlisted, and when recurrence was detected, it was recognized as multifocal and LT could not be indicated as a salvage procedure. These 2 subjects support the aforementioned concept suggesting that recurrence due to dissemination is very unlikely to be detected at a stage when salvage LT might be feasible. In addition, we have also evidenced that this policy offers an adequate long-term outcome. Only 1 patient developed massive or extensive tumor dissemination after LT and died 4 months after the operation. The other 4 patients have been followed for a median of 45 months and show no tumor recurrence, there being only 1 death at 84 months because of recurrent hepatitis C virus cirrhosis.

On the other hand, the outcome of patients classified as low risk has also been encouraging. Only 2 out of the 9 patients in this group have recurrence. One of them was amenable to be enlisted for LT (successfully performed after 9 months waiting time), while the other was discarded due to multifocal disease and was treated by transarterial chemoembolization. The remaining 7 patients are all alive and free of disease after a median follow-up of 55 months. They may develop recurrence during follow-up, but this will very likely be the result of a de novo tumor in the cirrhotic liver. Accordingly, this group of patients will benefit from a conservative approach with regular surveillance that allows them to skip or delay the risks associated to LT. If HCC occurs, they will become optimal candidates for salvage LT.

It could be argued that the number of patients in which the need of LT is avoided is very limited, and therefore, it makes no sense to maintain resection as the first option. However, any effort to optimize the use of the limited pool of cadaveric donors is worth undertaking.

In the future, the availability of effective adjuvant therapies may effectively prevent recurrence, and/or molecular profiling57 will refine the risk assessment even prior to tumor resection. While these advancements take place, our active salvage transplantation policy according to risk appears to be an effective treatment policy for patients with surgical HCC. Obviously, confirmation of our data by other groups should be available prior to unequivocally recommending this policy in conventional clinical practice.