Liver Resection as a Bridge to Transplantation for Hepatocellular Carcinoma on Cirrhosis

Abstract

Objective:

To assess the viability of a strategy of primary resection with secondary liver transplantation (LT) for hepatocellular carcinoma (HCC) on cirrhosis

Summary Background Data:

LT is the optimal treatment of HCC with cirrhosis. Owing to organ shortage, liver resection is considered as a reasonable first-line treatment of patients with small HCC and good liver function, with secondary LT as a perspective in case of recurrence. The viability of such strategy, positively explored in theoretical models, is not documented in clinical practice.

Methods:

Among 358 consecutive patients with HCC on cirrhosis treated by liver resection (n = 163; 98 of whom were transplantable) or transplantation (n = 195), the feasibility and outcome of secondary transplantation was evaluated in a 2-step fashion. First, secondary LT for tumor recurrence after resection (n = 17) was compared with primary LT (n = 195), to assess the risk and the outcome of secondary LT in patients who effectively succeeded to be treated by this approach. Second, primary resection in transplantable patients (n = 98) was compared with that of primary LT (n = 195) on an intention-to-treat basis, to assess the outcome of each treatment strategy and to determine the proportion of resected patients likely to be switched for secondary LT. Transplantability of resected patients was retrospectively determined according to selection criteria of LT for HCC.

Results:

Operative mortality (≤2 months) of secondary LT was significantly higher than that of primary LT (28.6% versus 2.1%; P = 0.0008) as was intraoperative bleeding (mean transfused blood units, 20.7 versus 10.5; P = 0.0001). Tumor recurrence occurred more frequently after secondary than after primary LT (54% versus 18%; P = 0.001). Posttransplant 5-year overall survival was 41% versus 61% (P = 0.03), and disease-free survival was 29% versus 58% (P = 0.003) for secondary and primary LT, respectively.

Of 98 patients treated by resection while initially eligible for transplantation, only 20 (20%) were secondarily transplanted, 17 of whom (17%) for tumor recurrence and 3 (3%) for hepatic decompensation. Transplantability of tumoral recurrence was 25% (17 of 69 recurrences). Compared with primarily transplanted patients, transplantable resected patients had a decreased 5-year overall survival (50% versus 61%; P = 0.05) and disease-free survival (18% versus 58%; P < 0.0001), despite the use of secondary LT.

On a multivariate analysis including 271 patients eligible for transplantation and treated by either liver resection or primary LT, liver resection alone (P < 0.0001; risk ratio [RR] = 3.27) or liver resection with secondary LT (P < 0.05; RR= 1.87) emerged as negative independent factors of disease-free survival as compared with primary LT. A number of nodules > 3 (P = 0.002; RR= 2.02) and a maximum tumor size exceeding 30 mm (P < 0.0001; RR=1.93) were also predictive of lower disease-free survival.

Conclusions:

LT after liver resection is associated with a higher operative mortality, an increased risk of recurrence, and a poorer outcome than primary LT. In addition, liver resection as a bridge to LT impairs the patient transplantability and the chance of long-term survival of cirrhotic patients with HCC. Primary LT should therefore remain the ideal choice of treatment of a cirrhotic patient with HCC, even when the tumor is resectable.

Liver transplantation (LT) after resection of hepatocellular carcinoma (HCC) on cirrhosis is associated with a higher operative mortality, an increased risk of recurrence, and a poorer outcome than primary LT. In addition, the strategy of liver resection as a bridge to transplantation significantly impairs the transplantability and the chance of long-term survival of cirrhotic patients. Primary LT remains the optimal treatment of a cirrhotic patient with HCC, even when the tumor is resectable.

Hepatocellular carcinoma (HCC) represents worldwide the most common malignancy in male patients,¹ accounting for more than 90% of primary liver neoplasms and causing more than 1 million deaths annually.² Traditionally considered as frequent in Asia and Africa and infrequent in Western countries, its incidence has however substantially increased in Japan,³ in the United States,⁴ in the United Kingdom,⁵ and in France⁶ during the past 3 decades. The majority of HCC (80–90%) are associated with underlying cirrhosis, and the main risk factors are infection with hepatitis C virus (HCV), infection with hepatitis B virus (HBV), alcoholic cirrhosis, and hemochromatosis. Only 2 treatments are potentially curative: hepatic resection and transplantation.

LT is the optimal treatment of HCC because it results in the widest possible resection margins for the cancer, removes the underlying cirrhotic liver tissue that is at risk for the development of de novo HCC and restores normal hepatic function. With the refinements of selection criteria,^7–9 several studies have shown that liver transplantation for early-stage HCC yields excellent results that are superior to resection^7,10–13 and equal to transplantation of candidates with nonmalignant diseases.^7–9,14 However, the limited availability of donor organs makes LT less available to individual patients.^15–17

Accordingly, hepatic resection remains an important therapeutic option. Significant improvement of survival after resection has been achieved within the past decade as a result of advances in the diagnosis and surgical management of HCC.^18–21 However, the long-term prognosis remains poor, with 5-year overall survival rates of 33% to 44%7,17,18,22 and 5-year cumulative recurrence rates of 80% to 100%.^12,23,24

Owing to the limited organ supply, a compromise gathered by many liver units is to resect selected patients with an expected hope of long-term survival, mainly those Child-Pugh A with single HCC, and to reserve LT for those with impaired liver function and/ or small oligonodular HCC according to currently accepted tumor selection criteria. For those patients who would present hepatic recurrence after resection, the possibility would remain further, to be transplanted.

With 2 other European centers, we conducted a preliminary study showing that this strategy of so-called “salvage” transplantation for recurrence in appropriately selected patients, could result in survival similar to that achieved by primary transplantation for HCC.²⁵ This pragmatic attitude is able to avoid unnecessary LT in long-term survivors of resection and could result in a significant sparing of organs,²⁶ but its viability remains to be explored. The aim of this study was to assess the risk and outcome of transplantation after liver resection in comparison with that of primary LT and to reappraise on an intention-to-treat basis whether the concept of salvage transplantation is both attractive in theory and applicable in practice.

PATIENTS AND METHODS

From November 1984 to December 2000, 495 consecutive patients with documented HCC were treated by hepatic resection (n = 271) or LT (n = 224) at our institution. We excluded from the analysis 137 patients with noncirrhotic liver (resection, n = 108; transplantation, n = 29) to focus the study on 358 patients with HCC on superimposed cirrhosis (resection, n = 163; transplantation, n = 195) (Fig. 1).

FIGURE 1. Overview of the Paul Brousse experience of hepatocellular carcinoma in the study period with outcome after resection and distribution of the study groups.

The 163 cirrhotic patients treated by liver resection during the study period were selected mainly when they had 1 or 2 nodules, good hepatic function, and preoperative and intraoperative absence of macroscopic vascular invasion and of extrahepatic tumor (preoperative chest and abdominal computerized tomography [CT] and bone scintigraphy). All resections were potentially curative. Among this group, 65 patients (42%) were retrospectively considered as nontransplantable because of their age > 65 years (n = 42), to a maximum tumor size exceeding 65 mm (n = 18) or to a large and multinodular tumor (ie, > 3 nodules and > 3 cm; n = 5). according to the selection criteria used for transplantation in the same period. The remaining 98 patients (60%) were potentially transplantable but treated by liver resection at a time when resection was always primarily preferred to LT, when feasible or more recently, because the situation of organ shortage restrict the indications of LT. Postoperative follow-up included liver function tests, dosage of serum alfa feto protein (AFP) and hepatic ultrasound on a 4-month basis, and thoracic CT once a year. As LT was available all long the study period, the policy was to consider LT for patients who would present liver recurrence or hepatic deterioration after resection. Accordingly, among the 98 transplantable patients (Group LR), 20 (20%) should be subsequently transplanted, of whom 17 (17%) for tumor recurrence and 3 (3%) for hepatic decompensation.

The 195 patients transplanted for HCC in the study period were selected according to the following pretransplant criteria: age < 70 years, absence of metastatic lymph nodes or evidence of extrahepatic spread, absence of macroscopic vascular invasion, no history of other malignant tumor from at least 5 years and from 1992, date of our proposal of new selection criteria,⁷ absence of large multinodular tumor (ie, maximum size >3 cm and number of nodules > 3). Indication of LT relied mainly on the technical unresectability of HCC or to an altered liver function (Child Pugh B or C). Preoperative staging routinely included hepatic ultrasound, chest and abdominal CT, and bone scintigraphy to look for any extrahepatic tumor. All LT procedures were potentially curative. Patients with HCC were given no priority in the waiting list as compared with other patients.

Design of the Study

The strategy of secondary transplantation was evaluated in a 2-step fashion. First, secondary LT for tumor recurrence after resection (Group LT2, 17 patients) was compared with Primary LT for HCC (Group LT1, 195 patients) to assess the risk and the outcome of patients who effectively succeeded to be treated for HCC by a secondary LT. Survival in each group was calculated from the time of transplantation.

Second, primary resection in patients eligible for transplantation (Group LR, 98 patients) was compared with primary LT (Group LT1, 195 patients) on an intention-to-treat basis, to assess the outcome of each treatment strategy and to determine the proportion of resected patients likely to be switched for secondary LT. Survival in each group was calculated from the time of the primary procedure (primary resection or primary transplantation). Patients with secondary LT were included in the resection group, and their survival was calculated from the time of resection.

Statistical Analysis

χ² test and analysis of variance (ANOVA) were respectively used to compare proportions and means between groups. Survival probabilities were estimated using the Kaplan-Meier method. The log-rank test was used to compare patient survivals. The stepwise Cox proportional hazard regression model was used to assess the independent factors of patient survival (overall and disease-free). Two multivariate analyses were performed: 1 in the whole group of transplanted patients, and 1 in the whole group of patients eligible for LT (resected and/or transplanted). A P value < 0.05 was considered statistically significant.

RESULTS

Comparison of Secondary Versus Primary Liver Transplantation

Patient and Tumor Characteristics Before Secondary Versus Primary Transplantation

At time of liver resection, the majority of patients secondarily transplanted had virus-related cirrhosis (76%). Five patients received a treatment of HCC before liver resection: 4 had previous transarterial chemoembolizations (TACE), 1 had previous alcohol injection. Three patients were treated for complicated portal hypertension (one portacaval shunt, 1 devascularization, 1 sclerosis of esophageal varices). All liver resections were done by a transabdominal approach. Hepatectomy was limited to fewer than 3 segments in 15 of the 17 patients (88%). Two patients underwent a repeat hepatectomy for localized tumor recurrence before the indication of LT. The median time from resection to transplantation was 2.2 years (0.3–8.1).

At time of transplantation, there were no differences between primary and secondary transplantation groups for age, gender, severity, and etiology of underlying cirrhosis (Table 1).

TABLE 1. Comparison of Patient and Tumor Characteristics Between Primary and Secondary Transplantation at the Time of Transplantation

Tumor characteristics before LT were similar in terms of number of nodules and of maximum tumor size. On the analysis of the specimens, the proportion of patients with more than 3 nodules was higher in Group LT2 than in Group LT1 (53% versus 25%; P = 0.03). Whereas the time interval from the diagnosis of cirrhosis to that of HCC was similar between groups, the mean time interval from the diagnosis of HCC to LT was significantly longer in Group LT2 than in Group LT1 (9.2 ± 9.8 versus 23.6 ± 18.3 months; P < 0.0001). During the waiting time for LT, most patients received a treatment of their HCC (TACE and/or local treatment) without difference between the 2 groups.

Mortality and Complications After Secondary Versus Primary Transplantation

Operative mortality (≤2 months) was significantly higher for secondary than for primary transplantation (23.5% versus 2.1%; P < 0.0001) (Table 2). Four patients died in Group LT2: 2 from cardiac arrhythmia, and 2 from sepsis with multiple organ failure. Among the 4 patients who died in Group LT1, 1 died of primary nonfunction, 1 from primary dysfunction complicated by renal and respiratory insufficiency, 1 from aspergillus infection, and 1 from cardiac failure related to pulmonary arterial hypertension. Intraoperative bleeding assessed by the number of transfused units of packed red blood cells was also increased in Group LT2 (mean, 16.7 ± 12.4 versus 10.9 ± 8.3; P = 0.009). By contrast, the operative time and the incidence of postoperative complications were similar between groups (Table 3).

TABLE 2. Comparison of Perioperative Data Between Primary and Secondary Transplantation

TABLE 3. Comparison of Outcome of Primary and Secondary Transplantation for HCC

Outcome After Secondary Versus Primary Transplantation

Tumor recurrence occurred in 7 of 13 posttransplantation survivors (54%) of a secondary LT versus 34 of 191 posttransplantation survivors of a primary LT (18%; P = 0.001; Table 3). The duration of posttransplantation follow-up was similar between groups (mean, 49± 50 versus 51 ± 46 months, respectively; P = NS). Intrahepatic recurrence accounted for the majority of the cases.

Posttransplant overall survival rate was lower after secondary LT than after primary LT (41% versus 61% at 5 years; P = 0.03). Disease-free survival was 29% versus 58%, respectively (P = 0.003; Fig. 2).

FIGURE 2. Comparison of survival between primary and secondary transplantation for HCC on cirrhosis. (A) overall survival; (B) disease-free survival.

Prognostic Factors of Survival After LT

At multivariate analysis of the whole transplanted group, a previous liver resection emerged as an independent risk factor for both overall survival (P = 0.003; risk ratio [RR] =3.8; CI, 1.6–9.2) and disease-free survival (P = 0.0002; RR = 4.9; CI, 2.1–11.1) with 3 other factors: a maximum tumor size > 30 mm (P = 0.004; RR = 2.1; CI, 1.3–3.6), the absence of total tumor necrosis (P = 0.03; RR = 2.6; CI, 1.1–6.0), and the presence of vascular invasion (P = 0.02; RR = 2.0; CI, 1.1–3.7) at pathology of the specimen.

Comparison of Primary Resection ± LT Versus Primary LT in Patients Initially Eligible for Transplantation

Patient Data and Tumor Characteristics of Resected Patients Eligible for LT Versus Transplanted Patients at Time of First Surgery

In the resection group, patients were slightly older but had less severe and less virus-related cirrhosis than transplanted patients (Table 4). Preoperative tumor characteristics were similar in term of size, of serum levels, of alfa feto protein but multinodular tumors (> 3 nodules) were almost exclusively observed in the transplantation group. These findings were confirmed on the analysis of specimens.

TABLE 4. Comparison of Patient and Tumor Characteristics Between Primary Resection in Transplantable Patients and Primary Transplantation

Transplantability of Resected Patients

The initial transplantability of the resected population in relation to preoperative tumor characteristics was 60% (98 of 163). As 20 patients were subsequently transplanted for tumor recurrence or hepatic decompensation, the final transplantability rate was 20% (20 of 98). Among the 98 transplantable patients, 75 patients (77%) presented a recurrence 52 of whom were intrahepatic (53%), 1 was extrahepatic (1%), and 22 were both intrahepatic and extrahepatic (22%). Twenty of the 52 intrahepatic recurrences (38.5%) were multinodular (>3 nodules). The transplantability of tumor recurrence was therefore 23% (17 of 75).

Outcome of Resected Versus Transplanted Patients

Survival after resection tended to be better for transplantable than for nontransplantable patients (52% versus 39% at 5 years, respectively; P = 0.09; Fig. 3).

FIGURE 3. Comparison of survival after liver resection for HCC on cirrhosis between patients initially eligible versus noneligible for transplantation.

Compared with primarily transplanted patients (Group LT1), transplantable resected patients (Group LR) had a decreased overall survival: 50% versus 61% at 5 years and 28% versus 50% at 10 years (P = 0.05), although the resection group also included secondary LT. The difference was still higher for disease-free survival that was 18% versus 58% at 5 years and 12% versus 48% at 10 years for Group LR and Group LT1, respectively (P < 0.0001; Fig. 4).

FIGURE 4. Comparison of survival between primary resection with possible transplantation and primary transplantation for HCC on cirrhosis (A) overall survival and (B) disease-free survival.

Prognostic Factors of Survival for All Patients Eligible for Transplantation and Treated by Liver Resection or Transplantation

Univariate analysis

In addition to the modality of treatment, 5 factors were evaluated, including age, sex, number of tumors, maximum tumor size, and serum levels of alfa feto protein. A number of tumors > 3 (P = 0.03) and a maximum tumor size > 3 cm (P = 0.0001) were significantly associated with a lower overall survival. Patient age (P = 0.007), the number of tumors (P = 0.002) and a maximum tumor size > 3cm (P < 0.0001) were significantly associated with a lower disease-free survival.

Multivariate analysis

Multivariate analysis included 271 patients eligible for transplantation and treated either by liver resection or primary LT. A number of nodules > 3 (P = 0.017), a maximum tumor size exceeding 30 mm (P = 0.002), and liver resection alone as compared with primary transplantation (P = 0.004) were independently associated to a lower overall survival (Table 5).

TABLE 5. Multivariate Analysis: 271 Patients Eligible for Transplantation (Resected or Transplanted)

A number of nodules > 3 (P = 0.002), a maximum tumor size exceeding 30 mm (P < 0.0001), liver resection alone compared with primary transplantation (P < 0.001), and secondary transplantation compared with primary transplantation (P < 0.05) emerged as negative independent factors of disease-free survival.

DISCUSSION

This study shows that LT after liver resection is associated to a higher operative mortality, an increased risk of recurrence, and a poorer outcome than primary LT. In addition, it is addressed to a limited proportion of patients treated by resection while initially eligible for LT

LT is theoretically the optimal treatment of HCC because it is the only one treating both the tumor and the underlying liver cirrhosis and providing by liver replacement the best oncological radicality and the best prevention of de novo tumors and of life-threatening complications of cirrhosis. Its superiority on liver resection has been shown in terms of incidence of recurrence and survival.^7,10–13 However, some series suggest equivalent results as LT^27–29 and the critical shortage of donor livers has limited the use of LT as a realistic option for many patients.^16,17,30 Therefore, liver resection has been considered as a reasonable first-line treatment of patients with small HCC and good liver function, with the perspective of LT as a second-line treatment in case of recurrence.^16,26,29

The rationale of this strategy is based on 3 assumptions. First, a proportion of patients may survive without recurrence for 5 and even 10 years without the need of LT. Second, a “salvage” LT if needed will have the same risk as primary LT. Third, the outcome of secondarily transplanted patients will be similar to that of primarily ones.

For Child Pugh class A patients with small HCC; hepatic resection is now associated with a reduced mortality^20,21 and a 5-year overall survival rate of up to 70%.^23,31–34 A proportion of 27% to 36% of patients may survive without recurrence for 5 years^12,18,23 and even 10 years (22%).²³ However, the prognostic factors are often insufficient to predict accurately which individual patient will really benefit from such a long survival. They may be different with regards to early or late recurrence.³⁵ On a probability estimation, it could be argued that a corresponding 5-year disease-free survival rate of 35% compares very unfavorably to that of around 70% with LT.^8,16,36 In addition, long-term outcome of patients with HCC should also concern 10-year survival, a point when results of resection and transplantation are still more divergent. The current 10-year survival after LT for HCC and cirrhosis in the European Liver Transplantation Registry (ELTR) is around 50%³⁷ compared with a maximum rate of 22% after resection.²³

With regards to the risk of transplantation after resection, it seems logical to expect a higher risk than that of primary LT. Postoperative adhesions on the site of hepatectomy may increase the difficulties of LT and the context of portal hypertension and progressive liver failure usually increases operative bleeding. The difficulties could still be increased by additional treatments that resected patients are likely to receive before LT such as TACE, alcohol injection, local treatments, or repeat resection. Accordingly, in our study, 2 of the 17 secondarily transplanted patients had been operated for complicated portal hypertension, and 2 were previously submitted to repeat liver resection. Operative bleeding was increased in secondary compared with primary transplant procedures and operative mortality was higher. Fifty percent (4 of 8) of all operative deaths observed in our series of transplanted HCC were from patients previously submitted to liver resection while these patients only represented 7% of the overall population of transplanted HCC. Although similar to patients treated by primary LT in terms tumor characteristics, patients with secondary LT had a higher risk of recurrence (54% versus 18%), possibly related to the longer history of HCC as shown by the longer time interval from the diagnosis to LT in these patients. Survival was also significantly impaired after secondary LT (39% versus 61% at 5 years). It may be however that the risk could be decreased by transthoracic or laparoscopic approaches associated with less postoperative adhesions or by alternative treatments such as percutaneous radiofrequency.

The viability of the strategy of “salvage” LT also relies on the absence of a too high drop out of resected patients from LT. Because we considered this as a key point in assessing the good strategy to adopt for a patient likely to benefit from the 2 options, primary LT or primary resection and secondary LT, we conducted an intend-to-treat analysis in our group of resected patients. Using similar criteria of selection as those of LT,⁷ 58% of our resected patients were considered transplantable. Although they were submitted to a close follow up by the same team and LT was available during the whole study period, the final transplantability was only 20% overall and 23% in case of tumor recurrence, a proportion significantly lower than the theoretical 40% to 80% transplantability rate proposed by Majno et al in the Markov-based decision model²⁶ or the 79% rate proposed by Poon et al²³ in a retrospective analysis of his series of resected patients. Our proportion agrees more closely to that of the Barcelona group where less than 10% of patients undergoing resection were bridged to LT during the last 10-year period.³⁴ Also, the first multicentric analysis of patients transplanted after liver resection that we did with 2 other large European centers only cumulated 12 patients.²⁵ Similarly, only 2 of 53 patients with transplantable recurrence were transplanted in the series of Poon et al.²³ These data support our findings that in practice secondary transplantation is addressed to a minority of resected patients.

Secondary LT could be also indicated for liver deterioration or for complications of cirrhosis. This was the case in 3% of our resected patients. In these circumstances, the rapid evolution of liver disease or a sudden life-threatening complication could hinder a potential LT by the impossibility to obtain an available organ on time. In the Poon et al series,²³ of the 6 of 130 resected patients (5%) who developed deterioration of liver function and were candidates for salvage transplantation, all died of cirrhotic complications 2 to 5 months after decompensation, a time that indeed precludes any possibility to be transplanted in most centers.

The waiting time for LT is indeed a critical factor in this period of organ shortage. After a median waiting time of 6 months, the probability of drop out for patients with HCC was 23% in Barcelona³⁴ and reached 30 to 50%³⁸ in other centers when the waiting time exceeds 1 year. This is concordant with that reported by Lau et al³⁹ that many untreated patients will become ineligible for LT in view of a 1-year probability of tumor progression of 70%, vascular invasion of 21%, and extrahepatic spread of untreated tumors of 9%.

This low transplantability rate would not have any clinical relevance if the outcome of patients treated by this strategy was similar to that of first transplanted patients. However, we show that even with the rescue of secondary LT, transplantable patients treated by primary resection had 2-fold less probability to be disease-free at long term than patients treated by primary LT (34% versus 62% survival at 5 years; 21% versus 53% at 10 years).

Therefore, the estimation assuming that every resected patient if transplantable could further benefit from LT is largely hindered by the risk of patient drop out for nontransplantable recurrence, for tumor progression during the waiting time of LT, or for a sudden life-threatening complication of cirrhosis. In addition the feasibility of salvage transplantation is closely related to early diagnosis of intrahepatic recurrence and therefore to a strict surveillance after resection.

In practice, when a patient is likely to be treated by resection or transplantation, a 2-stage strategy should not be applied systematically, but rather to a very selected group of patients with a low risk of recurrence after resection. Accordingly, patients with factors that adversely affect the disease-free survival after resection should be excluded such as those with oligonodular tumors, ie, 2 or 3 tumor nodules,²³ Child Pugh score B or C,^40,41 serum alpha-fetoprotein levels > 32 ng/ml³⁵ or serum aminotransferase greater than twice the normal levels,^42,43 hepatitis C-related HCC,^44,45 microscopic vascular invasion.^23,35 Primary transplantation would be a more appropriate option for this subset of patients. For the remnant patients with small-size, single HCC developed on Child Pugh A nonactive, non–virus C cirrhosis, it may be that a 2-stage procedure have some viability. However, in this setting, local treatments such as radiofrequency are also efficient in treating the tumor with the advantage of being mini-invasive, safer, and less compromising for a secondary LT. Indeed, such strategy of 2-stage procedure is increasingly used even in patients planned to be transplanted during the waiting time of LT to avoid tumor progression.^46,47,48

The strategy of secondary transplantation might also work best with a reduction of waiting time potentially offered by the increasing use of living related transplantation. Living donor transplantation (LRLT) eliminates the factor of waiting time. A Markov model comparing living donor versus cadaveric transplantation for early HCC showed substantial gains in life expectancy and cost-effectiveness when the waiting list for LT exceeded 7 months.⁴⁹ Another Markov model comparing intention-to-treat LT versus living donor for Child Pugh class A cirrhotic patients with nonresectable small (3–5 cm) HCC showed substantial benefit of survival for living donor transplant.⁵⁰ Therefore living donation may facilitate a more rapid performance of secondary LT. On the other hand, it increases the availability of organs in such a manner that access to primary transplantation is easier and probably more beneficial for the patients.

In conclusion, LT after liver resection is associated with a higher operative mortality, an increased risk of recurrence, and a poorer outcome than primary LT. In addition, liver resection as a bridge to LT impairs the patient transplantability and the chance of long-term survival of cirrhotic patients with HCC. Primary LT should therefore remain the ideal choice of treatment of a cirrhotic patient with HCC, even when the tumor is resectable.

ACKNOWLEDGMENTS

We are indebted to Vincent Karam and Valérie Delvart for providing information from the European Liver Transplant Registry and their assistance in the statistical evaluation of the data.

Discussion

Dr. Ronald P. DeMatteo (New York, New York): In the first part of your presentation it is not clear at what time you began measuring survival. In the abstract you used the phrase ‘posttransplantation survival.‘ Did you start measuring from the time of diagnosis or the time of transplantation for the first part of your presentation? If you measured from transplantation, that would obscure the potential benefit of performing a resection first.

The next thing that you didn’t touch upon was why there was such a high perioperative mortality rate in the patients who underwent salvage transplantation. It appears that you had a 29% perioperative death rate and now it is 23% with the addition of 3 extra patients from the time that you submitted the abstract. The question is, what operations did these patients have as their resection? Were there major hepatectomies? Or were there other factors that led to technical difficulties in these patients?

Lastly, the mean number of nodules in the group that underwent resection before transplantation was double that of the transplantation alone group. Perhaps the resection group was destined to have a poorer outcome.

Professor Rene Adam (Villejuif, France): Thank you, Dr. DeMatteo, for your question. For the calculation of survival, the date of transplantation was taken as reference to compare the 2 groups of primary and secondary transplantation. In the intention-to-treat analysis, survival was calculated from the date of the primary procedure, ie, resection for the primary resection group (including secondary transplantation), and transplantation for the primary transplantation group.

I agree with you that the peri-operative mortality observed in our secondarily transplanted patients was high. This obviously contrasts with the acceptable mortality observed in our whole series of transplantation for HCC, that was lower than 4% (8 out of 239 patients). Indeed half of the total number of deaths observed in our series (4/8), occurred in patients treated by a previous liver resection. This reflects in our view the higher technical difficulties of transplantation in these patients in relation to previous dissection of the liver and of the hepatic pedicle, as well as to postoperative adhesions which in the context of portal hypertension and liver failure, were responsible of higher intraoperative bleeding. The increased mortality could also have been related to the fact that 2 patients with secondary transplantation group were previously submitted to repeat liver resection and other 2 had surgery of portal hypertension prior to transplantation.

With respect to the number of tumoral nodules in the 2 groups, no difference was shown on the pretransplant setting. A higher number was evidenced on pathology of the liver specimens from secondary as compared with primary transplantation. This could be related to the longer history of HCC in the former patients, as reflected by their longer time interval from diagnosis of HCC to transplantation.

Dr. Henry A. Pitt (Milwaukee, Wisconsin): My comment relates both to this presentation and to the one earlier from Memorial. My question is what whole role ablation plays in these patients.

We recently reviewed our experience in Milwaukee over a decade and looked at all options for our patients with hepatocellular carcinoma. Like both groups, we found that the best results were achieved either with transplant or with surgical resection. Now, over that period of time, many ways to ablate these tumors have been used including alcohol, cryotherapy, more recently radiofrequency and chemoembolization. We have tended to use ablation in patients with more advanced Child’s class and more advanced tumors. We found in a multivariate analysis that Child’s class was an independent predictor. However, the results of ablation were surprisingly good, and we could not actually find a statistically significant difference in survival between our surgically treated patients and our ablated patients even though the ablated patients were theoretically worse patients.

In your presentation it seemed as though you used ablation preoperatively in your transplant patients. What is the role of ablation as a bridge to transplant? In which patients do you ablate and not do either resection nor transplantation?

Professor Rene Adam (Villejuif, France): To answer your question, Dr. Pitt, only 4 patients were treated by an ablative technique prior to transplantation in the whole series. Of course there is no possible comparison between the objective of an ablative technique and that of transplantation. A place exists for a local treatment before transplantation to avoid tumor progression mainly when the expected waiting time for transplantation is long. However, this treatment has only a local effect and is reserved to patients with small HCC visible at ultrasound and likely to be treated percutaneously. Our preferred approach has been to prevent tumor progression with a regional treatment such as transarterial chemoembolization. This was made both before resection and transplantation provided that the patient was not Child Pugh C.

Dr. Ronald W. Busuttil (Los Angeles, California): I enjoyed your paper very much and thank you for providing me with the manuscript.

Most of the questions that I had, have already been answered. But I would like to echo the results that you have presented today, because at UCLA our results were almost identical. We have only resected as a bridge to transplant only 6 patients out of 250 that we have transplanted for HCC. And of that very small series of six, we have a 50% two-year survival. So it has not been good in our hands either.

Two brief questions: First of all, do you think that there is any role in the cirrhotic patient with HCC for resection? And how would you identify that patient? Secondly, what do you think the role is for living related transplantation, which is being used more and more in this country to get people transplanted quickly? What are your views on that?

Professor Rene Adam (Villejuif, France): Your question, Dr. Busuttil, concerning the role of resection is really a key point. Obviously, the place of resection for HCC in view of our results is limited if the objective for any patient is to provide a treatment with optimal survival expectancy and minimal risk of recurrence. However, we should keep in mind that first, not all the patients could benefit from transplantation owing to lack of donors and second, that our results do not eliminate the possibility for a subset of very selected patients to benefit at long term from resection. We could hypothesize that this may concern patients with a single HCC less than 5 cm in size, without vascular invasion, developed on a nonactive, nonhepatitis C, well compensated cirrhosis in the absence of portal hypertension… Of course all these conditions make that the number of such patients is much more limited than that of the overall group of resectable patients. It may be that in this group, a strategy of two-stage procedure - resection and the transplantation - could work. But it remains to be demonstrated.

With respect to your second question, living related transplantation in my view, will probably increase the proportion of patients transplanted for HCC. A high proportion of such patients are currently excluded from the possibility to be transplanted on the basis of too restrictive criteria, owing to organ shortage. In view of the current alternatives to cadaveric liver transplantation, I think that the most important point to discuss is the reasonable cut-off of expected survival that we could admit to allow a patient with HCC to be transplanted. I would propose that an expected survival of at least 50% at 5 years is acceptable to justify the risk of living donation. According to this, the patients for whom both resection and transplantation are feasible, are probably those having the optimal chance of long-term survival. Living donation could offer them the possibility of having at once the best therapy.