We searched the MEDLINE database (PubMed) and the Cochrane Library to identify studies evaluating down-staging and locoregional therapy for hepatocellular carcinoma (HCC) before liver transplantation (LT). The search was restricted to studies written in English and published from 1995 to 2010. The key words for the search included down-staging, liver transplantation, hepatocellular carcinoma, locoregional therapy, staging, treatment outcomes, radio frequency ablation, and transarterial chemoembolization. All published articles and abstracts were reviewed and screened by 2 members of the working group independently (F.Y.Y. and S.B.), and relevant articles and abstracts were selected for further review by the group. The group agreed on 5 key questions for addressing this topic, and each of the 5 members of the working group led the review and summary of an assigned question. The summary was reviewed and approved by all members of the working group. The level of evidence was graded according to the standards of the Oxford Centre for Evidence-Based Medicine and was approved by all group members.
DEFINITION OF DOWN-STAGING AND RADIOGRAPHIC RESPONSES TO DOWN-STAGING TREATMENTS
The term down-staging has been used loosely with HCC to describe the effects of neoadjuvant therapy before LT. We define down-staging as reducing the size of a tumor with locoregional therapy specifically to meet acceptable criteria for LT. The responses of tumors to down-staging treatments should be based on radiological measurements of the sizes of viable tumors (by contrast-enhanced computed tomography or magnetic resonance imaging), and the measurements should not include areas of necrosis resulting from locoregional therapy. This follows the principle of tumor response to locoregional therapy in the European Association for the Study of the Liver guidelines.1 The complete response of a tumor nodule to treatment with a lack of tumor enhancement is equivalent to the obliteration of the tumor. According to this principle, patients with 4 lesions can achieve successful down-staging and meet the Milan criteria (up to 3 lesions)2 if 1 or more of the tumor nodules show a complete response to treatment with no viable tumors.
WHAT ARE THE GOALS AND EXPECTED OUTCOMES OF DOWN-STAGING?
The goal of HCC down-staging is to decrease the tumor burden so that the acceptable criteria for LT are met in patients whose initial tumor size and number exceeded these criteria. Acceptable criteria are determined by the expectation of good survival after LT, and the same criteria are often used to limit access to deceased donor LT. The principle of down-staging involves the selection of patients whose tumors have a more favorable biology and who respond well to treatment and also do well after LT.3 Thus, down-staging serves mainly as a selection tool for LT, but for the individual tumor, the down-staging treatment may not necessarily change the outcome of LT. The same principle of down-staging applies to living donor LT when the initial tumor burden exceeds the acceptable limits for the transplant center that is performing living donor LT. Clinically, the most relevant outcome parameters for down-staging are tumor recurrence after LT and survival, which includes posttransplant survival and intention-to-treat survival.
In terms of posttransplant survival, the Milan criteria2 remain the gold standard against which other criteria and down-staged HCC should be measured. The posttransplant survival rates for patients with down-staged tumors should be comparable to or only slightly below those achieved by patients with HCC meeting the Milan criteria before LT. The reported 5-year survival rates after LT for patients with HCC meeting the Milan criteria are 70% to 80%. We propose the goal of a 5-year posttransplant survival rate of 60% to 70% after tumor down-staging.
Intention-to-treat survival has been used in a number of studies to evaluate outcomes after HCC down-staging.3–12 This outcome measure starts from the time at which a patient with HCC is subjected to down-staging treatments before LT listing and takes into consideration tumor progression beyond acceptable limits (ie, changes in the tumor size or number, vascular invasion, or extrahepatic tumor dissemination) that leads to delisting or dropout from the LT waiting list. Dropout from the waiting list depends not only on the tumor biology but also on the time on the wait-list for deceased donor LT in a particular transplant center or region. In this context, HCC down-staging may allow the selection of patients whose tumors have a more favorable biology and who respond well to treatment and also do well after LT. At the same time, because of the greater initial tumor burden for patients undergoing the down-staging process, the dropout rate is expected to be greater than that observed in patients whose HCC is initially within the Milan criteria. As a result, the intention-to-treat survival rate with down-staging is also expected to be lower. Whether the intention-to-treat outcomes should be compared for those meeting the Milan criteria and those undergoing down-staging remains a debatable issue. Taking into consideration the shortage of organs, we should set a realistic and attainable goal for the intention-to-treat survival rate before we decide to perform transplantation for both patients with HCC within the Milan criteria and patients undergoing tumor down-staging. This is a highly complex issue because wait-list dropout depends not only on tumor-related factors but also on wait-list times and donor availability.
WHAT SHOULD BE THE ENDPOINT OR ENDPOINTS OF DOWN-STAGING?
Many transplant centers have continued to apply the Milan criteria2 for the selection of candidates for LT. In the United States, patients with HCC exceeding the Milan criteria are generally not eligible for priority listing for LT. Under these circumstances, it would be appropriate for the endpoint of tumor down-staging to be that the patients meet the Milan criteria and qualify for LT listing. Most of the published reports have used the Milan criteria as the endpoint of down-staging.3–8 The Response Evaluation Criteria in Solid Tumors (RECIST)9–11 and other criteria12, 13 have also been used to define the response to tumor down-staging, especially when eligibility for LT is not restricted to patients meeting the Milan criteria. Because we are defining the response to down-staging according to measurements of only the viable tumor after locoregional therapy, the RECIST approach does not fit well into this scheme: it defines treatment responses according to the measurement of the entire tumor (including both the ablated or necrotic portion and the viable tumor).14
Published data on the outcomes of tumor down-staging for LT are summarized in Tables 1 and 2; the data are classified according to the endpoint of down-staging. The concept of applying locoregional therapy to reduce the size of HCC tumors and thus facilitate resection or LT was first tested by Majno et al.13 (Hôpital Paul-Brousse, Villejuif, France). Two subsequent studies have suggested that a response to locoregional therapy with transarterial chemoembolization (TACE) for HCC meeting or exceeding the Milan criteria could serve as a prognostic marker for improved posttransplant outcomes and for the selection of LT candidates.10, 11 A number of more recent studies have formally evaluated the intention-to-treat outcomes of HCC down-staging to the Milan criteria for LT eligibility2–8 (Table 1). The endpoints of these studies include posttransplant survival, HCC recurrence, and dropout from the waiting list due to tumor progression or other causes. These studies are difficult to compare because of differences in their inclusion criteria (the upper limits for the tumor size and number) and in the types of locoregional therapies. Some of the studies were small (20 or fewer patients undergoing LT after down-staging6–8) and had short follow-up.
|Authors/Location||Design||Eligibility Criteria||Treatment||Down-Staging Success Rate||Posttransplant Outcome*||Level of Evidence|
|Yao et al.3 (2008)/San Francisco, CA (n = 61)||Prospective study with consecutive patients||Beyond the Milan criteria: 1 lesion ≤ 8 cm, 2 to 3 lesions ≤ 5 cm with a total diameter ≤ 8 cm, or 4 to 5 lesions ≤ 3 cm with a total diameter ≤ 8 cm||TACE, RFA, PEI, resection||70%||4-year survival, 92%; 4-year intention-to-treat survival, 69%; HCC recurrence, 0% (35)||2b|
|Ravaioli et al.5 (2008)/ Bologna, Italy (n = 48)||Prospective study||Beyond the Milan criteria: 1 lesion ≤ 6 cm, 2 lesions ≤ 5 cm, or 3 to 5 lesions ≤ 4 cm with a total diameter ≤ 12 cm||TACE, RFA, PEI, resection||69%||3-year tumor-free survival, 71%; HCC recurrence, 18% (32)||2b|
|Lewandowski et al.7 (2009)/ Chicago, IL (n = 43)||Retrospective cohort study||Beyond the Milan criteria with no upper limits for the size of up to 3 lesions (United Network for Organ Sharing T3)||TARE (yttrium 90)||58%†||1-year survival, 89%; 3-year intention-to-treat survival, 59%; HCC recurrence, 22% (9)||3b|
|De Luna et al.6 (2009)/ Stanford, CA (n = 27)||Retrospective cohort study||Beyond the Milan criteria with no upper limits for the tumor size or number||TACI||63%||3-year survival, 79%; HCC recurrence, 7% (15)||3b|
|Barakat et al.8 (2010)/ Houston, TX (n = 32)||Retrospective cohort study||Beyond the Milan criteria with no upper limits for the tumor size or number||TACE, RFA, TARE||56%||2-year survival, 75%; HCC recurrence, 14% (14)||3b|
|Authors/Location||Design||Eligibility Criteria||Treatment||Criteria for Response/ Success Rate||Posttransplant Outcome*||Level of Evidence|
|Majno et al.13 (1997)/Villejuif, France (n = 35)||Retrospective cohort study||Any number of tumors with the largest diameter > 3 cm||TACE||World Health Organization (50% reduction of the product of the perpendicular diameters of the largest lesion)/54%||5-year tumor-free survival, 71% with down-staging (19) versus 29% without a response to TACE (16)||2b|
|Graziadei et al.12 (2003)/Innsbruck, Austria (n = 15)||Prospective study||Beyond the Milan criteria with no upper limits for the tumor size or number||TACE||>50% reduction in the tumor size/67%||4-year survival, 41%; 5-year intention-to-treat survival, 31%; HCC recurrence, 30% (10)||4|
|Otto et al.10 (2006)/Mainz, Germany (n = 62)||Retrospective study with consecutive patients||Beyond the Milan criteria with no upper limits for the tumor size or number||TACE||RECIST/44% underwent LT [stable disease (18) or minimally progressive disease (9) before LT]||5-year freedom from recurrence, 75% (27)||2b|
|Millonig et al.11 (2007)/Innsbruck, Austria (n = 33)||Prospective study||Beyond the Milan criteria but within the UCSF criteria†||TACE||RECIST/85% with a complete or partial response||5-year survival, 67% with a complete response (6), 64% with a partial response (22), and 25% with stable or progressive disease (2)‡||2b|
|Chapman et al.9 (2008)/St. Louis, MO (n = 76)||Retrospective cohort study||Beyond the Milan criteria with no upper limits for the tumor size or number||TACE||RECIST/22%||5-year survival, 94%; HCC recurrence, 6% (17)||3b|
Two larger studies3, 4 prospectively applied a down-staging protocol that defined the upper tumor size and number limits for inclusion, and the Milan criteria were used as the endpoint for successful down-staging. The University of California San Francisco (UCSF) group reported the outcomes of 61 patients who underwent down-staging3; 43 patients (70%) underwent successful down-staging, and 35 patients (57%) underwent LT. The 4-year posttransplant survival rate was 92%, and the overall 4-year intention-to-treat survival rate was 69%. In a recent update from the UCSF group involving 88 patients,4 the dropout rate was 27%. The 5-year posttransplant survival rate was 84%, and the overall 5-year intention-to-treat survival rate was 61%. There were only 2 patients with HCC recurrence after LT. Consequently, successful down-staging apparently can be used to select a subgroup of patients with good tumor biology who have a low risk of tumor recurrence after LT. In a study of 48 patients by a group at the University of Bologna,5 the inclusion criteria were more liberal than those of the UCSF study in terms of the upper size limits for patients with 3 to 5 lesions. This study also incorporated a serum alpha-fetoprotein (AFP) level ≤ 400 ng/mL as a criterion for LT. The dropout rate was 31%, and the intention-to-treat survival rate was 60% at 3 years. The 3-year disease-free survival rate after LT was 71% for 32 patients, and 18% experienced HCC recurrence. The intention-to-treat outcomes were not significantly different from those for patients with HCC who met the Milan criteria without down-staging.
Using the Milan criteria as the endpoint for down-staging appears to be a reasonably conservative starting point, but more data about long-term outcomes are still needed. A standardized radiological assessment of the response based on the viable tumor is crucial; this is evidenced by the discrepancies in the treatment response criteria of the published reports.7 The down-staging process likely plays a major role in the selection of patients whose tumors have a more favorable biology and who respond well to treatment and also do well after LT.
WHAT ARE THE ELIGIBILITY CRITERIA FOR DOWN-STAGING?
Patients with extrahepatic HCC metastases and patients with radiological evidence of vascular invasion should be excluded from down-staging. Vascular invasion is well established as a strong predictor of postoperative tumor recurrence.
With respect to the tumor size and number, the majority of the published reports have not provided upper limits before down-staging.6–13 Only a few reports have clearly stated the upper limits of the tumor burden before down-staging.3–5 In the UCSF experience,3, 4 the upper limits for down-staging were a single tumor with a diameter up to 8 cm, 2 to 3 tumors with individual diameters up to 5 cm and with a total diameter up to 8 cm, or 4 to 5 tumors with individual diameters up to 3 cm and with a total diameter up to 8 cm. However, there were very few patients in the UCSF experience with 4 or 5 tumors who underwent down-staging. In the Bologna study,5 the upper limits included a single lesion with a diameter up to 6 cm, 2 tumors with individual diameters up to 5 cm, or 3 to 5 tumors with individual diameters up to 4 cm with a total diameter up to 12 cm. The upper size limits for 3 to 5 lesions in the Bologna study were, therefore, more liberal than those in the UCSF study. At a recent national conference on HCC in the United States, Pomfret et al.16 reported that patients with 4 or 5 tumors would be excluded by their proposed criteria for down-staging (a modification of the UCSF down-staging protocol). The proposed criteria for eligibility for down-staging included patients with HCC exceeding the Milan criteria who have a single tumor with a diameter up to 8 cm or 2 to 3 tumors with individual diameters up to 5 cm and with a total diameter up to 8 cm.16
A total tumor volume of 250 cm3 has been proposed as a cutoff for down-staging.17 With the tumor volume used as a criterion, the tumor size is considered to have a more significant impact on tumor recurrence than the tumor number. The usefulness of the tumor volume as a selection criterion needs to be further evaluated. Only Barakat et al.8 evaluated the growth pattern of tumors as a predictor of successful tumor down-staging. The response rate was only 33% for patients with infiltrative tumors who underwent down-staging but was 100% for patients whose tumors were not infiltrative. An AFP level > 1000 ng/mL was found to be a predictor of treatment failure in the UCSF study.3 An AFP level > 1000 ng/mL was included as an additional exclusion criterion at the US national conference (unless the AFP level decreases to <500 ng/mL with locoregional therapy).16 In the Bologna study,5 an AFP level ≥ 400 ng/mL was an exclusion criterion for LT.
The severity or stage of the underlying liver disease should be taken into consideration when locoregional therapy is being applied for down-staging. This issue should be approached differently in patients undergoing down-staging for LT and in patients who are not considered candidates for LT. For patients who are not candidates for LT, locoregional therapies have an impact on survival only if the liver disease is not too advanced (eg, patients with Child class A cirrhosis). In the context of tumor down-staging before LT, it may be argued that the restriction of therapy to patients with good liver function is not necessary because LT is available in the event of hepatic decompensation. Nevertheless, there is a limit to the severity of liver disease beyond which down-staging should not be applied because of the high risk of rapid liver function deterioration and mortality before patients meet the criteria for successful down-staging and become eligible for LT. The published literature includes only limited data to guide our decision about restrictions for down-staging treatments based on the severity of the underlying liver disease.
BY WHAT TECHNIQUE SHOULD DOWN-STAGING BE ACHIEVED?
The review of the literature was structured with the starting point that all treatments that can be used for HCC outside the transplant setting can also be used for down-staging HCC. The literature on efficacy and safety was reviewed. Efficacy was defined as the ability to achieve HCC down-staging according to the authors' criteria and by the rate of recurrence after LT. The techniques described in the literature include radio frequency ablation (RFA),3–5, 8 TACE,3–13 arterial chemoinfusion,6 alcohol injection,3–5 resection,3–5 and radioembolization.7, 8 With the exception of alcohol injection, it is clear that all these techniques are able to reduce the tumor burden to achieve the author-defined endpoints. Unfortunately, there is little unanimity about what the appropriate endpoints should be. The evaluated endpoints include down-staging to meet the Milan criteria and RECIST (Tables 1 and 2). The starting points have also been very heterogeneous. A number of studies have included several treatment modalities.3–5, 8 The sample sizes have mostly been small, and this has led to a lack of confidence in the results.
Nonetheless, patients who undergo down-staging with TACE have better posttransplant survival rates than those who do not; however, it is not certain that the 2 groups are identical, and the reasons that patients are assigned to one treatment group or the other have not been provided.13 Furthermore, patients for whom down-staging is successful have better survival rates than those for whom down-staging is attempted but is unsuccessful.10, 11, 13 The rate of successful down-staging varies from 22% to 70%. Posttransplant HCC recurrence rates vary considerably from approximately 0% to 18% in larger studies with at least 3 years of posttransplant follow-up3–6 (Table 1). This is no doubt a result of different selection criteria, but it is difficult to pinpoint why the recurrence rate is better in one study versus another. The posttransplant recurrence rate is very important because in most programs the shortage of livers means that any expansion of the listing criteria will lead to additional deaths on the waiting list. Thus, for every HCC patient who develops recurrent HCC and dies, another patient on the waiting list also dies because the liver that he or she might have received went to the HCC patient. None of the studies included an assessment of the effects that the expansion of the waiting list to include patients with down-staged HCC might have on the overall mortality rate of patients on the waiting list. Only a single uncontrolled study has compared the efficacy of down-staging with transarterial radioembolization (TARE) versus down-staging with TACE, and it found TARE to be the superior treatment modality in achieving the desired response before LT.7 How the treatments were chosen for these patients is unclear, and this raises the concern of a selection bias in the comparison of these 2 treatments.
The safety of down-staging is very difficult to assess. Only 1 study has reported safety data in any detail, and it described a 2.2% mortality rate for patients undergoing TACE.9 This is higher than the rate expected in a nontransplant setting. Another report described a mortality rate of 3.3% after resection or laparoscopic RFA for down-staging.3
The overall assessment of the literature is that there is no clear evidence for the superiority of one technique versus others in achieving successful down-staging. The studies have been mostly retrospective, and the prospective studies have been uncontrolled. The selection criteria used to determine the exclusion of patients from down-staging procedures are not clear. There are sufficient reports of the risks and mortality associated with down-staging that we should not accept down-staging as a procedure that can be only beneficial. Concerns should be raised about the risks of down-staging patients with advanced cirrhosis (ie, patients with Child class B cirrhosis and especially patients with Child class C cirrhosis).
WHAT CRITERIA SHOULD BE USED TO DELIST DOWN-STAGED PATIENTS AND TO DEFINE DOWN-STAGING FAILURES?
An established consensus statement based on past experience suggests that the best curative treatment for HCC is LT within established selection criteria: the Milan criteria2 or the UCSF criteria.15 In the process of down-staging tumors, the tumor burden according to a radiological assessment should be reduced to meet the maximum acceptable criteria for LT listing, regardless of the original tumor size and number. In addition, extensive data in the literature also support the incorporation of a reduction in the AFP level before LT. AFP cutoffs of 300, 400, and 1000 ng/mL have been suggested in several studies,3, 5, 16 but the best AFP level for predicting outcomes after LT has not been fully elucidated.
Once the tumor has been successfully down-staged and meets acceptable criteria, a minimal observation period of 3 months has been recommended before listing for LT.3, 4, 16 If the tumor progresses after listing beyond the acceptable criteria for LT, then temporary removal from the waiting list (delisting) is necessary so that the patient can undergo down-staging again at the same starting point. If repeat down-staging is successful, then the patient will become eligible for LT listing under the same criteria. For patients who respond to the initial down-staging process, the further treatment of residual tumors by TACE or other ablative techniques should be considered when the waiting time is expected to be at least 6 months after down-staging.
The definition of treatment failure for down-staging should be uniform in future studies evaluating the efficacy of down-staging. The proposed criteria include the following:
- 1Evidence of tumor progression reaching one or more of the endpoints resulting in permanent removal from the waiting list for LT (delisting). These endpoints include radiological evidence of vascular invasion and extrahepatic tumor spread. The UCSF group has also defined treatment failure as tumor progression beyond its inclusion criteria (based on the tumor size and number) for down-staging.3, 4
- 2Death due to any cause before LT.
- 3Recurrence of HCC after LT.
Our review of the literature suggests that uniform criteria are needed for future studies investigating further the efficacy and safety of HCC down-staging before LT. Our recommendations are as follows:
- 1We define HCC down-staging as reducing the tumor size with locoregional therapy specifically to meet acceptable criteria for LT (grade D).
- 2The tumor response to down-staging treatments should be based on radiological measurements of the sizes of viable tumors, and the measurement should not include the areas of necrosis resulting from locoregional therapy (grade D).
- 3We propose a goal of achieving a 5-year posttransplant survival rate of 60% to 70% after tumor down-staging. Whether intention-to-treat survival rates should be compared for those meeting the Milan criteria and those undergoing down-staging remains a debatable issue (grade D).
- 4We propose using the Milan criteria as the endpoint for down-staging, and more data about long-term outcomes are needed (grade C).
- 5Well-defined upper limits for the tumor size and number should be used to determine the eligibility for down-staging before LT (grade D). Other factors to be considered include an upper limit for the AFP level; beyond this limit, down-staging treatments are not likely to be successful, and posttransplant outcomes may be significantly worse.
- 6TACE is the most commonly used treatment modality, but there is no evidence that one treatment is superior to others in achieving down-staging (grade D). More data on the safety of down-staging are needed.
- 7A minimal observation period of 3 months between successful down-staging and LT is recommended (grade D).
- 8We propose the following criteria for treatment failure for down-staging: (1) tumor progression to one or more of the endpoints resulting in permanent removal from the waiting list (vascular invasion, extrahepatic spread, or tumor sizes and numbers beyond the inclusion criteria), (2) death due to any cause before LT, and (3) the recurrence of HCC after LT (grade D).