Living donor liver transplantation for hepatocellular carcinoma: A single-center preliminary report



Liver transplantation (LT) is the treatment of choice for early hepatocellular carcinoma (HCC) in patients with end-stage liver disease but is limited by the availability of donor organs. Living donor liver transplantation (LDLT) represents an alternative therapeutic option for patients with disease confined to the liver. Between April 1998 and December 2003, 537 patients underwent liver transplantation in our center. Thirty patients with HCC and associated terminal cirrhosis and 4 patients with tumor recurrence after liver resection who underwent LDLT were reviewed. Nineteen patients (55.8%) met the Milan criteria for LT, whereas 15 patients (44.2%) “exceeded” them. The overall survival rates at 1 and 2 years were 68% and 62%, respectively, with a median follow-up of 41 months (range, 17-64 months). Five patients (14.7%) died in the first 30 days after LDLT. Hospital mortality was significantly correlated with age >60 years. Four patients developed recurrence between 6 and 35 months after LDLT. Recurrence was significantly related to the presence of more than 3 tumor lesions in our series. In conclusion, LDLT is a promising treatment option for patients with HCC. Even longer follow-up and bigger patients' series are needed to fully assess the benefits of LDLT for HCC patients exceeding the Milan criteria. Liver Transpl 12:934–940, 2006. © 2006 AASLD.

Hepatocellular carcinoma (HCC) is one of the most common malignancies worldwide, accounting for about 1 million related deaths annually. Its incidence is estimated to have increased by 75% in the United States in the last decade and reflects the increasing prevalence of chronic viral hepatitis.1 An estimated 372,000 new cases of hepatocellular carcinoma are diagnosed each year, constituting 4.6% of all new human cancers.2 Liver transplantation (LT) is considered to be the treatment of choice for early HCC in patients with end-stage liver failure, but LT availability is limited by donor organ availability. Despite recent criticism, the Milan criteria are most often used when determining organ allocation for HCC patients being considered for deceased donor LT.3

Due to recent developments in the field, living donor liver transplantation (LDLT) has emerged as a potentially better alternative to deceased donor LTx for patients with HCC. Cost-effectiveness analyses have shown that in patients with standard indications, the benefits of LDLT are achieved when the waiting time for a deceased donor liver exceeds 7 months.4 Moreover, LDLT can provide an opportunity for patients with tumors exceeding current listing criteria for deceased LT, as well as for patients who meet the criteria but are not able to wait because of end-stage liver cirrhosis.5

We herein report our experience with adult LDLT for HCC.


HCC, hepatocellular carcinoma; LDLT, living donor liver transplantation; LT, liver transplantation.


Between April 1998 and December 2003, 537 patients underwent liver transplantation in our center. Eighty-one patients with HCC were evaluated for LDLT in this period of time. Twenty-one of them (26%) experienced dropout after the evaluation because of extending our tumor criteria (extrahepatic tumor disease or portal vein thrombosis, n = 18), or death during the evaluation because of deterioration of end-stage liver disease (n = 3). Thirty-two further potential recipients were excluded from LDLT during the evaluation because of advanced tumor disease, complications of end-stage liver cirrhosis, or contraindications for the available donors (small liver volume, medical reasons). The remaining 28 patients (35%) underwent LDLT.

The data of 34 patients with HCC in pathology who underwent LDLT were reviewed. This group of patients also encompassed 6 in whom HCC was clinically suspected prior to LDLT. In other words, the definitive diagnosis was postoperatively established but there was already a suspicion of HCC prior to LDLT (identification of a mass >2 cm in diameter in a cirrhotic liver in 2 imaging modalities combined with early contrast enhancement on computed tomography and elevated serum alpha fetoprotein).

Tumor-specific evaluation included abdominal ultrasonography, abdominal and thoracic computed tomography, hepatic computed tomography angiography (reconstruction increments of 1 mm for the arterial scan and 2 mm for the venous scans), bone scintigraphy and determination of alpha-fetoprotein. Additionally, magnetic resonance imaging of the abdomen was performed in 9 patients and total body positron emission tomography in 4 patients to refine unclear staging findings. The diagnosis of HCC was established according to the consensus statement from the European Association for the Study of the Liver.6 The clinical classification was based on the morphological description of the tumor according to the radiological findings of 2 independent radiologists. The number of tumors, lobar distribution of the tumors, maximal tumor diameter, and the presence or absence of macrovascular invasion or of extrahepatic tumor disease was recorded. The total tumor diameter for patients with multiple tumor nodules was calculated as the sum of the maximal diameter of each lesion in centimeters.

LDLT was considered only for those patients with HCC confined to the liver. Patients with extrahepatic tumor manifestation or with radiological suspicion of macroscopic vascular invasion of the portal vein or of the hepatic veins were excluded. The number and size of the tumors did not constitute exclusion criteria per se.

Twenty-eight patients with clinically evident HCC who underwent LDLT fell retrospectively into 3 categories: (1) HCC patients with liver cirrhosis meeting the Milan criteria for tumor classification (n = 14, 9 of them in Child-Turcotte-Pugh Class B and 5 in Child-Turcotte-Pugh Class C), (2) HCC patients with liver cirrhosis exceeding the Milan criteria (n = 10, 7 of them in Child-Turcotte-Pugh Class B and 3 in Child-Turcotte-Pugh Class C), or (3) HCC patients without cirrhosis with a recurrent HCC after previous liver resection (n = 4).

Additionally, in the subgroup of patients with suspicion of HCC prior to LDLT, 3 patients exceeded the Milan criteria due to the presence of multilobar small tumors (all 3 in Child-Turotte-Pugh Class B), and 3 were within the Milan criteria (2 in Child-Turcotte-Pugh Class B and 1 in Child-Turcotte-Pugh Class C).

The demographic data of the transplanted patients are shown in Table 1. Prior to admission to our center, 13 patients underwent transarterial chemoembolization and 4 received radiofrequency ablation as tumor-specific bridging treatments. In all these patients, transarterial chemoembolization or radiofrequency ablation were performed in other institutions before the possibility of LDLT was discussed.

Table 1. Demographic Data of 34 Patients With HCC Who Underwent LDLT
  1. Abbreviations: HBV, hepatitis B virus; HCV, hepatitis C virus.

Age (years)Median, 53.4 years (range, 18-65)
Follow-up period (months)Median, 33 months (range, 2-63)
Pretransplant indication 
 HBV-liver cirrhosis + HCC10
 HCV-liver cirrhosis + HCC14
 Alcoholic-liver cirrhosis + HCC3
 Cryptogenic liver cirrhosis + HCC3
 Recurrent HCC4

A right hepatectomy, comprising segments 5 to 8, without inflow occlusion, was performed in all donors, as previously reported.7 Careful pathologic study (1-cm slices) of all livers removed was performed by the same pathologist.

No patient was lost to follow-up. Follow-up studies included computed tomography scans of the abdomen and chest and measurement of alpha-fetoprotein levels every 4 months during the first 2 years after transplantation, every 6 months during the third year, and yearly thereafter.

Statistical Analysis

The categorical variables were analyzed by Fisher exact test. Survival was calculated using the Kaplan-Meier method, with statistical significance determined by the log-rank test; P < 0.05 was considered significant.


Recipient morbidity and mortality is shown in Table 2. The median duration of the intensive care unit stay was 9.5 days. Five recipients (14.7%) died during the first 30 postoperative days due to acute pancreatitis (n = 2), small-for-size syndrome leading to multi-organ failure (n = 2), and cardiac failure (n = 1). The overall survival rates at 1 and 2 years were 68% and 62%, respectively (Fig. 1) with a median follow up of 41 months (range, 17-64 months).

Table 2. Major Morbidity and Mortality in Patients Who Underwent LDLT for HCC
MorbidityPatients (n)30 Days' Mortality (n)Overall Hospital Mortality (n)
Heart failure111
Pulmonary embolism1-1
Acute pancreatitis222
Hepatic artery thrombosis1--
Biliary complications4--
Embolic stroke affecting to brainstem infarction1-1
Fungal pneumonia2-1
Methicillin-resistant Staphylococcus aureus sepsis2-1
Figure 1.

Overall survival of 34 HCC patients after LDLT.

Univariate analysis of patients and tumor characteristics is demonstrated in Table 3. Four patients developed tumor recurrence at 6, 23, 35, and 56 months after LDLT. Among these, the first of them died 10 months after LT because of multifocal tumor recurrence, and the second died 63 months after LDLT. Recurrence-free survival rates at 1 and 3 years (hospital mortality included) were 64% and 56%, respectively (Fig. 2). When the patients who were lost during the first 3 months after LDLT were excluded, the corresponding 1- and 3-year recurrence-free survival was as high as 88% and 75%, respectively. The sites of the recurrence were the lungs only, the lungs and the adrenal gland in 2 patients, the lungs and the lacrimal gland-orbital region in another patient, and pelvic soft tissue and bone in the forth patient. Treatment for the recurrent tumor included adrenalectomy and systemic chemotherapy in the first patient, resection of lung metastases and chemotherapy in the second, lung and facial surgery plus chemotherapy in the third, and systemic chemotherapy alone in the fourth patient. Two patients are still alive in good general condition with pulmonary metastases and pelvic soft tissue and bone metastases 62 and 58 months after LDLT, respectively. Recurrence was significantly associated with the presence of more than 3 tumor lesions (P = 0.025 with Fisher exact test). One- and 3-year recurrence-free survival were 66% and 62%, respectively, among patients who met the Milan criteria, compared with 60% and 47%, respectively, among those who did not.

Table 3. Relationship Between Various Factors and Overall Mortality/Tumor Recurrence After LDLT
Patient groupsTotal (n)Overall Mortality (n)P Value for Overall Mortality (Fisher exact test)Recurrence (n)P Value for Recurrence (Fisher exact test)
  1. NOTE: Tumor number >3 cm was significantly related with tumor recurrence in our series (P = 0.025).

  2. Abbreviations: NS, not significant; TACE, transarterial chemoembolization; RFA, radiofrequency ablation; CLIP, Cancer of the Liver Italian Program Investigators; MELD, model for end-stage liver disease; AFP, alpha-fetoprotein; p-TNM, postoperative tumor-node-metastasis; UICC, Union Internationale Contre le Cancer; UCSF, University of California San Francisco; BCLC, Barcelona Clinic Liver Cancer.

Age  NS NS
 <60 years267 3 
 >60 years86 1 
HCC patients within the Milan criteria  NS NS
 Child-Turcotte-Pugh B93 0 
 Child-Turcotte-Pugh C52 1 
HCC patients beyond the Milan criteria  NS NS
 Child-Turcotte-Pugh B74 3 
 Child-Turcotte-Pugh C32 0 
HCC recurrence in non cirrhotic liver after prior resection  NS NS
 Within Milan criteria20 0 
 Beyond Milan criteria21 0 
Cirrhotic patients with a suspicion of HCC prior to LDLT  NS NS
 Within Milan criteria30 0 
 Beyond Milan criteria31 0 
Child-Turcotte-Pugh class  NS NS
 B217 3 
 C94 1 
Pre-LT therapy (TACE, RFA)  NS NS
 Positive179 3 
 Negative174 1 
Radiological Milan criteria  NS NS
 Within154 2 
 Beyond199 2 
 0-1 Points123 1 
 2-3 Points2210 3 
 <223010 4 
 >2243 0 
 <400 ng/mL266 1 
 >400 ng/mL87 3 
 Stage I93 1 
 Stage II72 0 
 Stage III115 3 
 No TNM possible183 0 
Post-LT tumor diameter  NS NS
 <5 cm157 0 
 >5 cm196 4 
Post-LT tumor number  NS 0.025
 ≤3227 0 
 >3126 4 
Tumor characteristics  NS NS
 Solitary134 2 
 Multifocal219 2 
Tumor grading  NS NS
 Grade I114 1 
 Grade II144 2 
 Grade III22 1 
 No grade possible73 0 
Microvascular invasion  NS NS
 Positive42 1 
 Negative3011 3 
Pathological Milan criteria  NS NS
 Within196 0 
 Beyond157 4 
UCSF criteria     
 Beyond137 3 
BCLC staging     
 B94 2 
 C22 0 
 D31 0 
Figure 2.

Recurrence-free-survival of HCC patients after LDLT.

In 7 patients, although vital tumor was found in pathology, an exact histopathological tumor staging and grading could not be determined due to extensive areas of necrosis after transarterial chemoembolization or radiofrequency ablation. In all of these patients, a diagnosis of HCC was also confirmed histologically according to the results of a percutaneous biopsy achieved prior to bridging treatments.

Correlating the radiological findings prior to LDLT with the pathological analysis of the explanted livers, 20 patients (59%) were correctly evaluated for the number of tumor lesions (10 of them continued to meet and 10 to exceed the Milan criteria). In the remaining 14 patients, pathology gave a new assessment of tumor characteristics: 10 overestimates (6 of them concerning the number of tumor lesions and the rest four concerning the tumor size) and 4 underestimates (3 of them concerning the number of tumor lesions and 1 concerning vascular invasion). The problem of correctly evaluating the clinical tumor stage of patients with HCC and cirrhosis has already been discussed.8 Of the 10 overestimated patients, who would have been excluded from deceased LT by the Milan criteria, 3 died in the first 3 months after LT, and 1 died 6 months after LT due to relapsing pneumonia. Six of them are alive without recurrence at a median interval of 36 months (range, 16-50 months).

From the 53 patients excluded from LDLT after presentation/evaluation, 21 patients died within 9 months and 17 patients were lost to follow-up, whereas 15 patients are currently alive and receiving palliative treatment in a median follow-up of 8 months.


Liver transplantation is now acknowledged as the best therapeutic option for patients with early, irresectable HCC.9–11 LDLT for HCC is a controversial option, discussed with animosity in the literature.12–16 Since the introduction of the model for end-stage liver disease by United Network for Organ Sharing in February 2002 to allocate livers in the United States and its adaptation in the case of HCC patients with cirrhosis, a higher priority was given to HCC patients meeting the Milan criteria in the this country.10, 17 A corresponding priority does not currently exist within the “Eurotransplant area” (Austria, Belgium, Germany, Luxembourg, the Netherlands, and Slovenia), where allocation is mainly based on waiting time and on disease severity as defined by Child-Turcotte-Pugh classification among other criteria. Consequently, these patients are subject to long waiting times prior to receiving deceased donor organs, potentially resulting in dropouts due to tumor progression. Because of this, attempts are routinely made to treat identified HCC in transplant candidates to prevent disease progression while they are waiting.18–20

LDLT for HCC in cirrhosis is a treatment option for these patients. The main theoretical advantage is the very quick availability of an organ with optimal graft quality. The possibility of extending indications for LT beyond classic criteria is an extraordinary chance for these patients. The timely transplant, in contrast to deceased LT, reduces the effect of the “observation time” and the natural selection, and it could result in a better selection of patients. Patients who initially meet the Milan criteria but experience dropout due to tumor progression while on the waiting list, patients outside current listing criteria at the time of presentation, and patients who meet the criteria but are not able to wait because of end-stage liver cirrhosis may be considered in this debate.

The Milan criteria are derived from deceased organ allocation. The situation is quite different in LDLT, where the powerful incentive of the donor to donate provides a timely graft. The current international trend toward “expanding” the existing Milan criteria in the case of LDLT is based on several factors: (1) Current imaging techniques do not always allow appropriate staging of small HCC <2 cm, with high rates of false-negative results8, 21–23; (2) HCC is often multifocal; the difficulty of differentiating a multifocal HCC from regenerative nodules in an end-stage cirrhotic liver can make it impossible to allocate according to the Milan criterion “2 to 3 tumors all ≤3 cm”8, 21–23; (3) there is still no imaging method that can estimate microvascular invasion, a factor correlated with bad prognosis and quick recurrences in most studies17, 22, 24; and (4) many transplant centers based on pathological findings report acceptable overall and recurrence-free outcomes in patient groups exceeding the Milan criteria.12–14, 16, 24–27 For all of these reasons, LDLT is becoming progressively utilized in several centers using “extended indications.”

Vascular invasion, tumor size >5 cm, tumor number >3, and poor differentiation are reported to have a significant positive impact on HCC recurrence after LT in various series.8, 11, 15, 24–30 In our series, tumor number >3 was significantly related with recurrence (Table 3), showing that conventional Milan criteria maintain their significant influence also in the LDLT setting. Although the present series present the longest median follow-up reported for similar series, recurrence rate was 17% (4/24 patients), which was lower than the one reported by Gondolesi et al. (6/28 patients, 22%) and by Axelrod et al. (4/11 patients, 36%), and similar to that reported by Todo et al. (40/281 patients, 14%).12, 14, 31

The early mortality rate was 14.7%, which is similar to the rates reported in the literature by centers performing LDLT for HCC as “extended indications.”12–13 A perioperative mortality of 22% (8/36 patients) and of 11% (35/316 patients) was reported by Gondolesi et al. and by Todo et al., respectively.12, 14 Hospital mortality was significantly greater for patients aged >60 years in our series (P = 0.01 with Fisher exact test).

An obvious limitation of this report is its “preliminary character” and the small number of patients included. Further analysis in needed to evaluate the current selection criteria for patients with HCC undergoing LDLT, especially for patients with extended indications. However, according to our results, patients over 60 years should be very critically selected or excluded.