Reassessing the boundaries of liver transplantation for hepatocellular carcinoma: Where do we stand with tumor down-staging?

Authors

  • Francis Y. Yao,

    Corresponding author
    1. Division of Gastroenterology, Department of Medicine, University of California, San Francisco, CA
    • Address reprint requests to: Francis Y. Yao, M.D., University of California, San Francisco, 513 Parnassus Avenue, Room S-357, San Francisco, CA 94143-0538. E-mail: francis.yao@ucsf.edu; tel: +1-415-476-2777; fax: +1-415-476-0659.

    Search for more papers by this author
  • Nicholas Fidelman

    1. Division of Interventional Radiology, Department of Radiology, University of California, San Francisco, CA
    Search for more papers by this author

  • Potential conflict of interest: Dr. Fidelman received grants from BTG.

Abstract

Down-staging of hepatocellular carcinoma prior to liver transplantation (LT) has generated a lot of interest in recent years and has been identified in two recent national conferences on hepatocellular carcinoma as one of the priorities for research. Down-staging is defined as reduction in the tumor burden using local regional therapy specifically to meet acceptable criteria for LT. The rationale behind down-staging of tumors initially exceeding conventional criteria for LT is to select a subgroup of tumors with favorable biology and prognosis for LT as assessed by their response to local regional therapy. The expectation is to achieve comparable posttransplant survival between patients who achieve successful tumor down-staging before LT and those whose tumors meet LT criteria at the outset without needing down-staging. The application of tumor down-staging requires a highly structured approach using a treatment protocol that includes five essential components: eligibility criteria, down-staging endpoints, selection of the type of local regional therapy, minimal observation period from successful tumor down-staging to LT, and criteria for treatment failure and exclusion from LT. This review article summarizes published data on down-staging and addresses key questions related to each of the components of the down-staging protocol as well as treatment efficacy. Conclusion: Based on a review of published data and recommendations from recent national and international conferences on hepatocellular carcinoma and LT, a standardized down-staging protocol is proposed to further evaluate the feasibility and efficacy of applying tumor down-staging on a broader scale. (Hepatology 2016;63:1014–1025)

Abbreviations
AFP

alpha-fetoprotein

ART

Assessment for Retreatment with TACE (score)

CT

computed tomography

DEB

drug-eluting beads

HCC

hepatocellular carcinoma

LI-RADS

Liver Imaging Reporting and Data System

LRT

local regional therapy

LT

liver transplantation

MELD

Model for End-Stage Liver Disease

mRECIST

modified Response Evaluation Criteria in Solid Tumors

MRI

magnetic resonance imaging

PEI

percutaneous ethanol ablation

RFA

radiofrequency ablation

TACE

transarterial chemoembolization

UCSF

University of California, San Francisco

UNOS

United Network for Organ Sharing

The incidence of hepatocellular carcinoma (HCC) has been rising in the United States and much of the developed world for more than two decades.[1] In 1996, the Milan criteria (one lesion ≤5 cm, two or three lesions ≤3 cm)[2] identified a subset of patients with early-stage HCC who could achieve excellent survival following liver transplantation (LT) with a low risk of posttransplant tumor recurrence. Since that time, the demand for LT as a definitive therapy for HCC has continued to increase. After the Model for End-Stage Liver Disease (MELD) allocation scheme for HCC within the Milan criteria was implemented in the United States in 2002, the proportion of patients receiving priority listing with HCC-MELD exception nearly doubled from 10.5% to 19.4% by 2008.[3, 4] As of 2012, malignancy (the majority being HCC) was the second most common indication for LT (22.4%), behind only hepatitis C (24.5%).[5]

The success of LT for HCC has also fueled widespread debate on modest expansion of the tumor size limits to offer LT as a curative therapy to more patients with HCC.[6-8] In recent years, however, there has been a shift from relying solely on morphologic criteria of size and number to incorporating markers of tumor biology including alpha-fetoprotein (AFP) and response to local regional therapy (LRT) in the selection scheme.[9, 10] In this context, high AFP and tumor progression despite LRT identify aggressive tumors with a substantially greater risk for tumor recurrence after LT.[11-15] Tumor down-staging, which encompasses both expanded criteria as well as response to LRT as a measure of favorable tumor biology, has been gaining momentum in recent years. Down-staging has been identified as one of the priorities for research in two national conferences on HCC.[16, 17] In this article, we review the rationale and goals of down-staging, essential components of the down-staging protocol, and published outcome data. We also discuss future directions and propose a standardized down-staging protocol to further evaluate the feasibility and efficacy of applying tumor down-staging on a broader scale.

HCC Diagnosis

The diagnosis of HCC may be established radiographically if typical imaging features are present on a multiphase dynamic contrast-enhanced computed tomography (CT) or magnetic resonance imaging (MRI). The presence of arterial hyperenhancement followed by washout on delayed phases of imaging is highly specific for HCC ≥10 mm in diameter, and these diagnostic parameters have been widely adopted for HCC diagnosis.[18, 19]

The Liver Imaging Reporting and Data System (LI-RADS) was launched in March 2011 to improve standardization and consensus regarding performance, interpretation, and reporting CT and MRI of the liver in patients with cirrhosis or other risk factors.[20] According to the LI-RADS classification, definitely and probably benign lesions are labeled as LI-RADS-1 and LI-RADS-2, respectively. Lesions suspicious for HCC are assigned a category based on the likelihood that a lesion is HCC: intermediate probability for HCC (LI-RADS-3), probably HCC (LI-RADS-4), and definitely HCC (LI-RADS-5). LI-RADS category assignment depends on four imaging criteria: arterial phase hyperenhancement, washout, capsule, and threshold growth (defined as >50% diameter increase in <6 months). For an arterially enhancing lesion ≥2 cm, LI-RADS-5 observations are based on one additional imaging criterion, equivalent to class 5 using the United Network for Organ Sharing (UNOS)/Organ Procurement and Transplantation Network guidelines.[17, 21] For an arterially enhancing lesion 1-1.9 cm, two additional criteria are needed for the lesion to be classified as LI-RADS-5. One of the goals of LI-RADS is to render false-positive imaging diagnoses of HCC exceedingly rare. Studies demonstrating the actual proportion of LI-RADS-3, LI-RADS-4, and LI-RADS-5 observations that subsequently are proven to be HCC are still needed for validation of these imaging categories.[20]

Definition of Down-Staging of HCC for LT

Down-staging is defined as a reduction in the tumor burden using LRT specifically to meet acceptable criteria for LT.[22] Radiographic evaluation of treatment response should adhere to the LI-RADS principles and should be based on measurements of the maximum diameter of only viable tumors by multiphase CT or MRI. The measurements should not include the area of necrosis resulting from LRT (Fig. 1). This follows the principles of the European Association for the Study of Liver Disease[23] and modified Response Evaluation Criteria in Solid Tumors (mRECIST)[24] guidelines. The evaluation of response to down-staging is not always straightforward, as illustrated in some examples (Fig. 2A,B). If there is more than one area of enhancement in an otherwise necrotic lesion, then the diameter of the entire lesion should be counted toward the overall tumor burden and additional treatments are needed for down-staging.

Figure 1.

Contrast-enhanced CT images in the arterial phase for a patient with a solitary large HCC undergoing down-staging. (A) A LI-RADS-5 lesion measuring 8 cm in maximum diameter before treatment. (B) After two DEB-TACE treatments, there was residual tumor enhancement within a partially necrotic tumor. The maximal diameter of the enhancing portion should be measured (arrow) for the purpose of tumor size determination and whether successful down-staging is achieved. (C) Complete tumor necrosis after an additional DEB-TACE treatment.

Figure 2.

Contrast-enhanced CT images in the arterial phase illustrating two cases in which multiple areas of enhancement are seen within the tumor following local regional therapy. (A) Following DEB-TACE of a lesion measuring 5.5 cm in maximum diameter, there are two areas of residual tumor enhancement in an otherwise necrotic lesion. The entire lesion should be counted toward the overall tumor burden, and additional down-staging treatment is needed. (B) Following DEB-TACE of a lesion measuring 6 cm in maximum diameter, there are multiple areas of heterogeneous tumor enhancement and partial necrosis within the tumor. The entire lesion should be counted toward the overall tumor burden, and additional down-staging treatment is needed.

The term “down-staging” should not be used loosely to describe tumor “down-sizing” or the effects of LRT, without adhering to HCC staging classifications. In the UNOS tumor–node–metastasis classification and staging system (Table 1),[25] the Milan criteria have been incorporated as T1 and T2 stages, and only patients with HCC meeting T2 criteria are currently eligible for receiving priority listing for LT. This staging system provides a good framework for the standardized application of down-staging.

Table 1. UNOS Tumor–Node–Metastasis Classification and Staging System for HCC
  1. a

    The staging system incorporated the Milan criteria into T1 and T2.

Class 
T1aOne nodule ≤1.9 cm
T2aOne nodule 2.0–5.0 cm; two or three nodules, all <3.0 cm
T3One nodule >5.0 cm; two or three nodules, at least one >3.0 cm
T4aFour or more nodules, any size
T4bT2, T3, or T4a plus gross intrahepatic portal or hepatic vein involvement
N1Regional (portal hepatitis) node involvement
M1Metastatic disease, including extrahepatic portal or hepatic vein involvement
Stage 
IT1
IIT2
IIIT3
IVA1T4a
IVA2T4b
IVBAny N1, any M1

Rationale and Goals of Down-Staging

The rationale behind down-staging of tumors initially exceeding conventional criteria for LT is to select those with favorable biology and prognosis for LT based on their response to LRT.[22, 26] The concept of applying LRT for the purpose of reducing the size of HCC to facilitate resection or LT was first tested by Majno and colleagues from Hospital Paul Brousse, Villejuif, France.[27] In this study, down-staging (for tumor diameters >3 cm) was defined as a 50% reduction of the product of perpendicular diameters of the largest lesion. The 5-year recurrence-free survival after LT was 71% in the 19 patients who achieved successful down-staging with transarterial chemoembolization (TACE) compared to only 29% in the subgroup of 16 patients who did not achieve down-staging with TACE and 49% in the other 22 patients who had no treatment. Subsequent studies by Otto et al. from Mainz, Germany,[13] and Millonig et al. from Innsbruck, Austria,[14] have suggested that response to TACE for HCC within or beyond Milan criteria may serve as a prognostic marker for improved posttransplant outcome and for the selection of candidates for LT. These results support the rationale behind down-staging.

Despite the small number of patients in the study by Majno et al., their findings were intriguing and provided the impetus for formal investigations on down-staging prior to LT.[28-37] According to the International Consensus Conference on LT for HCC, LT after successful down-staging should achieve a 5-year survival rate comparable to that of HCC patients who meet the criteria for LT without requiring down-staging.[38] Furthermore, the goal of LT for HCC (with or without down-staging) is to achieve survival comparable to nonmalignant indications. The outcome following LT for HCC treated with down-staging should not be compared to other treatments such as resection when determining who should be candidates for LT after down-staging.

Down-Staging Protocol

The application of tumor down-staging requires a highly structured approach that includes five essential components: eligibility criteria, down-staging endpoints, selection of the type of LRT, minimal observation period from successful tumor down-staging to LT, and criteria for treatment failure and exclusion from LT (Fig. 3). Published studies on tumor down-staging are summarized based on the endpoints of down-staging to within Milan criteria (Table 2) or other criteria for response to LRT (Table 3). The problems in many of these studies include the lack of a well-defined protocol with all the necessary components and the short duration of post-transplant follow-up.

Table 2. Summary of Results of Down-Staging to Within Milan Criteria
Authors, Institution (n)Study DesignEntry CriteriaTreatmentDown-Staging Success RatePosttransplant Outcome (n = Transplanted)
  1. a

    Only the most recent series on down-staging from the same institution is reported.

  2. b

    Radiographic assessment of response was based on measurement of the entire tumor including the treated or necrotic portion and not just the viable tumor.

  3. Abbreviations: TACI, transarterial chemoinfusion; TARE, transarterial radioembolization.

Yao et al.,[37] San Francisco, CA (n = 118)aProspective study

Beyond Milan,

1 lesion ≤8 cm,

2–3 lesions ≤5 cm and total diameter ≤8 cm, 4–5 lesions ≤3 cm and total diameter ≤8 cm

TACE, RFA, PEI, resection65%5-yr survival 78%, 5-yr intention-to-treat survival 56%, 8% HCC recurrence (n = 64)

Green et al.,[36] Colorado

(n = 22)

Retrospective cohort study

Beyond Milan,

UNOS T3; no upper size limits for up to 3 lesions

TACE (DEB)50%29% HCC recurrence (n = 7)

Barakat et al.,[35]

Houston, TX (n = 32)

Retrospective cohort studyBeyond Milan, no upper limits in tumor size and numberTACE, RFA, TARE56%2-yr survival 75%, 14% HCC recurrence (n = 14)

Jang et al.,[34]

Soul, Korea (n = 386)

Retrospective cohort studyBeyond Milan, no upper limits in tumor size and numberTACE42%5-yr survival 55%, 30% HCC recurrence (n = 37)
Lewandowski et al.,[33] Northwestern, Chicago, IL (n = 43)Retrospective cohort study

Beyond Milan,

UNOS T3; no upper size limits for up to 3 lesions

TARE58%b1-yr survival 89%, 3-yr intention-to-treat survival 59%, 22% HCC recurrence (n = 9)

De Luna et al.,[32]

Stanford, CA

(n = 27)

Retrospective cohort studyBeyond Milan, no upper limits in tumor size and numberTACI63%3-yr survival 79%, 7% HCC recurrence (n = 15)

Ravaioli et al.,[31]

Bologna, Italy (n = 48)

Prospective study

Beyond Milan

1 lesion ≤6 cm,

2 lesions ≤5 cm,

3–5 lesions ≤4 cm and total diameter ≤12 cm

TACE, RFA, PEI, resection69%3-yr tumor-free survival 71%, 3-yr intention-to-treat survival 60%, 18% HCC recurrence (n = 32)
Table 3. Summary of Results of Down-Staging Based on RECIST and Other Criteria
Authors, Institution (n)Study DesignEntry CriteriaTreatmentCriteria for Response; Success RatePosttransplant Outcome (n = Transplanted)
  1. a

    UCSF criteria: 1 lesion ≤6.5 cm, 2–3 lesions ≤4.5 cm with total diameter ≤8 cm.

  2. b

    An additional 68 patients had tumors within Milan criteria showing significantly better survival with CR or PR than SD or PD.

  3. Abbreviations: CR, complete response; PD, progressive disease; PR, partial response; SD, stable disease.

Chapman et al.,[29] St. Louis, MO (n = 76)Retrospective cohort studyBeyond Milan, no upper limits in tumor size and numberTACERECIST; 22%5-yr survival 94%, 6% HCC recurrence (n = 17)
Otto et al.,[13] Mainz, Germany (n = 62)Retrospective consecutive patientsBeyond Milan, no upper limits in tumor size and numberTACERECIST; 44% underwent LT (18 SD, 9 minimal PD before LT)5-yr freedom from recurrence 75% (n = 27)
Millonig et al.,[14] Innsbruck, Austria (n = 33)Prospective studyBeyond Milan, within UCSF criteriaaTACERECIST; 85% with CR or PR5-yr survival 67% for CR (n = 6), 64% for PR (n = 22), and 25% for SD/PD (n = 2)b
Graziadei et al.,[28] Innsbruck, Austria (n = 15)Prospective studyBeyond Milan, no upper limits in tumor size and numberTACE>50% reduction in tumor size; 67%4-yr survival 41%, 5-yr intention-to-treat survival 31%, 30% HCC recurrence (n = 10)
Majno et al.,[27] Villejuif, France (n = 35)Retrospective cohort studyAny tumor number >3 cm in largest diameterTACEWHO (50% reduction of the product of the perpendicular diameters of largest lesion); 54%5-yr tumor-free survival 71% with down-staging (n = 19) compared to 29% without response to TACE (n = 16)
Figure 3.

Schematic diagram for down-staging of HCC for LT.

An intention-to-treat principle was applied in most studies, which factored in waiting list outcome including mortality and removal from the waiting list due to tumor progression or other causes. In one study by Jang et al. from Seoul, Korea,[34] the intention-to-treat outcome is difficult to interpret because of the low probability for LT despite successful tumor down-staging as a result of the scarcity of deceased donors. Consequently, less than 10% of the cohort ultimately underwent LT. Case series with fewer than 10 patients are excluded. Despite the limitations in many published studies, more robust data with long-term follow-up are emerging to gain insight into the efficacy of tumor down-staging before LT.

Eligibility Criteria

Patients with radiographic evidence of tumor vascular invasion should not be considered for down-staging.[22] Only a few studies have clearly stated the upper limits in tumor burden prior to down-staging.[30, 31, 37] In the University of California, Francisco (UCSF) down-staging protocol,[30, 37] the upper limits included one tumor ≤8 cm, two or three tumors each ≤5 cm and the sum of the maximal tumor diameters ≤8 cm, and four or five tumors each ≤3 cm and the sum of the maximal tumor diameters ≤8 cm. In another study, by Ravaioli et al. from Bologna, Italy,[31] the upper limits included one tumor ≤6 cm, two tumors each ≤5 cm, and three to five tumors each ≤4 cm with the sum of maximal diameters ≤12 cm. Studies by Lewandowski et al. from Northwestern[33] and Green et al. from Colorado[36] used UNOS T3 criteria (no upper limits in tumor diameter) as entry criteria. The eligibility criteria proposed in the US national conference on HCC represent a modification of the criteria used in the UCSF down-staging protocol after excluding those with four or five nodules.[17]

Patients considered for down-staging should have preserved liver function and good performance status to allow for LRT to be performed safely. TACE is typically offered to patients with bilirubin up to 3 mg/dL.[19, 39] Selection criteria for yttrium-90 transarterial radioembolization are more stringent than those for TACE. Patients with bilirubin >2 mg/dL or aminotransferase levels greater than five times the upper limits of normal are typically not offered radioembolization.[40]

Inclusion of patients with impaired hepatic function (Child's C cirrhosis, serum bilirubin >3 mg/dL) into down-staging programs is subject to debate. On the one hand, LT provides a potential safety net for patients who develop hepatic decompensation after LRT. On the other hand, treatment-related decline in liver function in the face of suboptimal tumor response may hasten a patient's demise. There are limited data on the safety of TACE in patients with hepatic impairment. One study in high-risk patients undergoing TACE has shown a significantly increased risk for procedure-related mortality or an urgent need for LT in patients with bilirubin >4 mg/dL, international normalized ratio for prothrombin time >1.5, creatinine >2 mg/dL, albumin <2 g/L, Child's C cirrhosis, MELD >20, and/or hepatic encephalopathy prior to TACE.[41] Sieghart et al.[42] have proposed the Assessment for Retreatment with TACE (ART) score, designed to predict overall patient survival after TACE or bland embolization. The components of the ART score are Child-Pugh-Turcotte score increase by 1 or >2 points (1.5 and 3 points, respectively), aspartate aminotransferase increase >25% (4 points), and radiologic tumor response to first TACE procedure (1 point). Patients with an ART score of ≤1.5 had significantly longer overall survival than patients with a score of ≥2.5. Clinical utility of the ART score has yet to be validated by other investigators.[43]

Endpoint of Down-Staging

Most of the published reports on down-staging used the Milan criteria as the endpoint of down-staging (Table 2). Other studies regarded down-staging as a response to LRT based on the World Health Organization or mRECIST criteria (Table 3). It seems logical to use the Milan criteria as the endpoint of down-staging as they have been widely considered the benchmark against which the outcome using other criteria for LT as well as tumor down-staging should be measured.[38] Furthermore, the Milan criteria have been incorporated into the UNOS staging classification (as T1 and T2; Table 1) in the United States to determine eligibility for priority listing for LT (MELD exception) in HCC.[25]

Down-Staging Strategies

Due to extensive disease burden (large lesions and/or multifocal disease) in patients undergoing tumor down-staging, TACE has been employed as a first-line therapy in the majority of published studies. Multiple TACE sessions are usually required before all active disease foci have been targeted. Many institutions have transitioned from conventional TACE to drug-eluting bead (DEB) TACE. The efficacy of DEB-TACE is probably equivalent to or slightly better than that of conventional TACE.[39]

For residual or recurrent disease, additional TACE should be performed if a patient was a responder to prior TACE and if liver function remains adequate (bilirubin <3 mg/dL). Subsegmental highly selective TACE may be performed safely even for some patients with poor liver function.[37, 41] Patients with preserved liver function (bilirubin <2 mg/dL) who have large or multifocal tumors refractory to TACE may be offered radioembolization.[35]

There are angiographic and other imaging features that predict poor response to TACE.[44] Among nonresponders to TACE, local ablative techniques may be considered if there are no more than three foci of residual/recurrent disease and if tumor location is feasible for treatment. Percutaneous radiofrequency ablation (RFA) is the most commonly used technique, and it achieves the best results in lesions measuring ≤3 cm in maximal diameter.[18, 19] Larger lesions may be treated with a laparoscopic or open approach. Microwave ablation may be employed as an alternative to RFA. The relative efficacy of achieving local control using RFA and microwave ablation is probably similar. However, lesions abutting large vessels may respond better to microwave ablation due to the lack of a “heat-sink” effect associated with the use of microwave technology. The use of cryoablation for treatment of HCC has been limited in comparison to RFA and microwave ablation.

Percutaneous ethanol injection (PEI) is inferior to RFA with respect to local tumor control.[18, 19] PEI should therefore be limited to treatment of small lesions ≤3 cm not suitable for thermal ablation due to their anatomic location, such as lesions close to the gallbladder or bowel.[31, 37] PEI can be performed with needles as small as 25 gauge. Therefore, this technique may be preferable for patients with ascites, severe thrombocytopenia, or severe coagulopathy, who may be at a higher risk of bleeding associated with the use of larger ablation probes (≥18 gauge).

In our approach, even if the tumors have been successfully down-staged to meet LT criteria, we perform additional LRT intended to induce complete tumor necrosis while the patient is still awaiting LT.[37] Due to high initial tumor burden, it is unlikely that patients who achieve complete tumor necrosis following down-staging are cured of their tumor and no longer need LT. Long-term response is seen primarily in very early-stage HCC with a single lesion <2 cm treated with RFA.[19] Those with larger or multiple tumors still face a significant risk of tumor recurrence or progression while on the waiting list for LT, despite initial complete response to LRT.[45]

Minimum Observation Period

A minimal observation period from successful down-staging to LT is required to observe for stability of the tumors (maintaining successful down-staging) prior to LT. This minimum observation period was 3 months in the UCSF protocol,[30, 37] although the median time from first down-staging treatment to LT in the UCSF study was much longer at 9.8 months.[37] The optimal duration of this observation period is unknown. A minimum observation period of 3 months should be based on two consecutive cross-sectional imaging studies performed at least 3 months apart showing disease stability with successful tumor down-staging being maintained according to imaging criteria.

Criteria for Down-Staging Failure and Removal From Waiting List

Macrovascular tumor invasion and extrahepatic tumor spread are obvious exclusion criteria for LT. What should we do if after initial successful down-staging there is evidence of tumor progression but the active tumor burden remains within LT criteria? Recent studies have suggested that mRECIST progression (defined as a 20% increase in diameter or development of one or more new tumors) after LRT predicts poor posttransplant outcome for HCC within and beyond Milan criteria[13-15] or a high risk of waiting list dropout in a region with prolonged waiting list time.[9, 45] An argument could be made for excluding these patients from LT. According to expert opinion in the International Consensus Conference on LT for HCC,[22] if tumor progression occurs but the active tumor burden remains within down-staging entry criteria, LT should be put on hold and repeat LRT should be undertaken until the endpoint of down-staging is achieved for the patients to be eligible again for LT. Similarly, after successful down-staging, if tumor progression occurs (by mRECIST criteria) but the tumor burden remains within Milan criteria, we recommend deferring LT and repeating LRT until disease stability is achieved on repeat imaging. On the other hand, progression of tumors to beyond entry criteria for down-staging constitutes an exclusion criterion for LT.[30, 37]

Additional exclusion criteria have been considered. In the study by Barakat et al. from Houston,[35] patients with infiltrative tumors undergoing down-staging had a response rate of only 33% versus 100% if the tumors were not of the infiltrative type. The Bologna group excluded patients with AFP ≥400 ng/mL from LT.[31] In the UCSF study, AFP >1000 ng/mL was a significant predictor of treatment failure.[37] According to the proposed criteria from the US national conference, LT should not be performed in those with AFP >1000 ng/mL unless it decreases to <500 ng/mL with LRT.[17] These AFP levels are somewhat arbitrary and require validation.

Outcome of Down-Staging of HCC Prior to LT

Can Down-Staging Achieve the Same Posttransplant Outcome as That in Patients Meeting LT Without Down-Staging?

Among studies using the Milan criteria as the endpoint of down-staging, only three included a control group (within Milan criteria) and had adequate follow-up of ≥3 years after LT.[31, 32, 37] All three studies demonstrated comparable survival rates between the down-staged group and the control group (Table 4).

Table 4. Comparison of Survival and HCC Recurrence Between the Down-Staged Group and the Milan Group
Author, Institution (n)Intention-to-Treat Survival (Kaplan-Meier Method)a

Posttransplant Survival (Kaplan-Meier Method)

Recurrence-Free Probability (Kaplan-Meier Method)

  1. a

    In analysis of intention-to-treat survival, time 0 represents first down-staging treatment in the down-staging group and the time of listing for liver transplant in the Milan group.

  2. b

    Survival without HCC recurrence was reported.

  3. Abbreviations: DS, down-staging; NR, not reported.

Yao et al.,[37]

San Francisco, CA

DS (n=118): 56% (5-yr)

Milan (n=488): 63% (5-yr)

P=0.29

DS (n=64): 78% (5-yr)

Milan (n=332): 81% (5-yr)

P=0.69

DS (n=64): 91% (5-yr)

Milan (n=332): 88% (5-yr)

P=0.66

Ravaioli et al.,[31]

Bologna, Italy

NR; only proportion of patients alive were reported: DS 56%;

Milan 63%

NR

DS (n=32): 71% (3-yr)b

Milan (n=88): 71% (3-yr)b

P=not significant

De Luna et al.,[32]

Stanford, CA

NR

DS (n=15): 79% (3-yr)

Milan (n=68): 82% (3-yr)

P=0.51

NR

In the latest published series from UCSF,[37] 77 of 118 patients (65%) enrolled into the down-staging protocol had successful down-staging and 64 patients (54%) had received LT. The 5-year posttransplant survival was 78%, and the 5-year intention-to-treat survival was 56%. Tumor recurrence developed in 8% after a median posttransplant follow-up of 3.8 years. There were no significant differences in the 5-year intention-to-treat survival, posttransplant survival, or recurrence-free probabilities when compared to a control group of 488 patients listed for LT with HCC initially meeting Milan criteria over the same period.

In the study by Ravaioli et al. from Bologna, Italy, 48 patients were included in the down-staging protocol and 32 patients underwent LT with a 3-year disease-free survival of 71%.[31] Among them, 18% had developed HCC recurrence after LT. In the other group of 129 patients with HCC meeting Milan criteria without requiring down-staging, 88 patients underwent LT. In the down-staging cohort, the 3-year disease-free survival after LT was 71% and 14% developed HCC recurrence. The difference in the disease-free survival between the down-staging and Milan groups was not statistically significant.

In a study by De Luna et al. from Stanford involving 28 patients undergoing down-staging, 15 had received LT, with 3-year posttransplant survival of 79%.[32] Among 95 patients meeting Milan criteria at the outset, 68 patients had undergone LT and their 3-year posttransplant survival was 82%. The difference in the posttransplant survival was not significantly different between the two groups (P = 0.51).

What Is the Likelihood of Failure of Down-Staging and Removal From the LT Waiting List?

The rates of successful down-staging in published reports using Milan criteria as the endpoint range from 42% to 69% (Table 2); but these studies differed according to entry criteria, the type of LRT, and how successful down-staging was defined, making generalization of the results difficult. Due to more advanced HCC, the rate of removal from the LT waiting list as a result of tumor progression is expected to be higher in patients undergoing down-staging when compared to those with HCC meeting acceptable LT criteria at the outset. In the study from UCSF,[37] the 1-year and 2-year cumulative probabilities for dropout by competing risk analysis were 24% and 34% in the down-staging group, which were significantly higher when compared to the respective 1-year and 2-year cumulative probabilities of dropout of 20% and 26% in the Milan group (P = 0.04). No other studies performed detailed analysis of the cumulative probabilities of dropout from the waiting list. In the Bologna study,[31] the overall crude unadjusted dropout rate was 31% in the down-staging group versus 21% in the Milan group. The median time from listing to LT was 6.2 months in the down-staging group and 4.8 months in the Milan group. In the Stanford study,[32] the proportion of waiting list dropout was 26% in the down-staged group and 18% in the Milan group. The median waiting list time to LT was similar in the down-staged group and the Milan group (11.4 months versus 10.9 months).

What Factors Predict Down-Staging Failure or Removal From the LT Waiting List?

According to the study from Houston, patients with infiltrative tumors undergoing down-staging had a response rate of only 33% compared to 100% if the tumors were not of the infiltrative type.[35] High AFP has been shown to be a marker of poor prognosis after LT[11, 12, 46] as well as a predictor of waiting list dropout.[45] In the study from UCSF,[37] a baseline AFP of ≥1000 ng/mL was associated with a 2.4-fold increased risk of waiting list dropout after down-staging compared to AFP <1000 ng/mL.

The severity of liver failure is an important determinant of the feasibility and safety of applying LRT for down-staging, but there have been no reported upper limits in the MELD score or Child-Turcotte-Pugh score for exclusion of patients from down-staging. In the UCSF study,[37] patients with Child's B cirrhosis had a 1.7-fold greater risk of waiting list dropout compared to those with Child's A cirrhosis, but Child's C cirrhosis was not found to be a risk factor for waiting list dropout. This discrepancy may be related to the small numbers with Child's C cirrhosis and referral bias. There is obviously a limit to the severity of liver disease beyond which down-staging should not be applied due to a high risk of hepatic decompensation or death before meeting criteria for successful down-staging. This also highlights the importance of counseling patients about the potential risks of down-staging.

Is One LRT Better Than the Other in Achieving Down-Staging?

TACE is the mainstay of down-staging treatment, and most studies have used conventional TACE with thiodized oil; but DEB-TACE may be safer with a lower risk of liver toxicity when compared to conventional TACE.[19, 39] The Northwestern group compared the outcome of down-staging between 43 patients treated with TACE and 43 others receiving radioembolization.[43] The probability of down-staging (from UNOS T3 to T2) was 58% in the radioembolization group versus 31% in the TACE group (P = 0.023). The median index tumor diameter went from 5.6 cm to 3.4 cm after radioembolization and from 5.7 cm to 4.3 cm after TACE (P < 0.001). The caveats of this study include possible selection bias and the inclusion of both treated and necrotic portions of the tumor in the radiographic assessment of tumor down-staging. Nevertheless, as the experience with radioembolization continues to grow, it is a promising treatment modality for tumor down-staging.

Does Down-Staging Result in Selection of Tumors With Favorable Biology for LT?

If tumor down-staging serves as a tool for selecting tumors with favorable biology and good prognosis, then the rate of HCC recurrence and the incidence of unfavorable explant histologic characteristics (microvascular tumor invasion and poorly differentiated tumor grade) should be low or comparable to those meeting LT criteria without down-staging. In the UCSF study,[37] the incidence of microvascular or macrovascular invasion in the explant was 3.2% in the down-staged group versus 6% in the Milan group (P = 0.34). Poorly differentiated tumor grade was found in none of the patients in the down-staging group versus 8.5% in the Milan group (P = 0.08). In contrast, in the Bologna study,[31] 56% of patients who underwent LT after down-staging had microvascular tumor invasion and 75% had poorly differentiated tumor grade in the liver explant. Furthermore, 50% had tumors in the explant not seen on preoperative imaging. The reasons for the large discrepancies between these two studies are not entirely clear, but the more liberal entry criteria used in the Bologna study may have been a contributing factor.

Future Directions

Although published results on tumor down-staging before LT are encouraging, they are based entirely on single-center experience. Multicenter studies in regions with different lengths of waiting list time are needed to determine if down-staging is applicable on a broader scale. In this setting, a standardized down-staging protocol is necessary to yield reproducible results.

Proposed guidelines for a standardized down-staging protocol are summarized in Table 5. The eligibility criteria for down-staging are based on recommendations in the US national conference.[17] They represent a modification of the UCSF down-staging criteria after excluding those with four or five lesions. The numbers with four or five lesions in the UCSF study were small, thus raising uncertainties regarding the efficacy in this subgroup.[30, 37] When considering the more liberal eligibility criteria proposed by the Bologna group,[31] supporting data are limited; and there are concerns about the higher rate of HCC recurrence, as well as the much higher proportion of patients with microvascular invasion and poorly differentiated tumor grade. Only patients with adequate hepatic function (Child's A and B, total bilirubin up to 3 mg/dL) should undergo down-staging based on recommended guidelines for TACE.[19] We recommend TACE as the primary down-staging treatment. The proposed endpoint of down-staging is to meet the Milan/UNOS T2 criteria. Criteria for removal from the waiting list for LT are largely based on expert opinion presented during the International Consensus Conference on LT for HCC.[22] High AFP ≥1000 ng/mL is an additional exclusion criterion unless there is a substantial drop in the level to <500 ng/mL with LRT, as recommended in the US national conference.[17] Infiltrative growth pattern is also considered an exclusion criterion[35] as infiltrative HCC is associated with a very poor overall prognosis.[47] A minimum observation period of 3 months from the time of successful down-staging to LT (two consecutive imaging studies 3 months apart) is recommended to insure disease stability prior to LT.

Table 5. Proposed Standardized Down-Staging Protocol
  1. a

    Recommendations based on the US National Conference.[17] Criteria based on cross-sectional imaging with multiphase CT or MRI of the abdomen.

  2. b

    Recommendations based on guidelines for a TACE procedure.[19]

  3. c

    Recommendations based on expert opinion presented during the international consensus conference on LT for HCC.[18] Criteria based on cross-sectional imaging with multiphase CT or MRI of the abdomen.

Eligibility criteria
1. UNOS T3 with upper limits in tumor diameter defined as followsa:
a. Single lesion ≤8 cm
b. 2 or 3 lesions each ≤5cm with the sum of the maximal tumor diameters of all lesions ≤8 cm
2. Absence of extrahepatic disease and vascular invasiona
3. Child's A and B cirrhosis with total bilirubin ≤3 mg/dLb
Criteria for successful down-stagingc
1. Residual tumor size and diameter within Milan criteria (UNOS T2)
a. Only viable tumor(s) considered; tumor diameter measurements should not include the area of necrosis from tumor-directed therapy
b. If there is more than one area of residual tumor enhancement, then the diameter of the entire lesion should be counted toward the overall tumor burden
Criteria for down-staging failure and exclusion from LT
1. Progression of tumor(s) to beyond inclusion/eligibility criteria for down-staging (as defined above)c
2. Tumor invasion of a major hepatic vessel based on cross-sectional imagingc
3. Lymph node involvement by tumor or extrahepatic spread of tumorc
4. Infiltrative tumor growth pattern
5. AFP ≥1000 ng/mL, unless the AFP level decreases to <500 ng/mL after LRTa
Additional guidelines
1. A minimal observation period of 3 months between successful tumor down-staging and LT is required. This should be based on two consecutive cross-sectional imaging studies at least 3 months apart showing disease stability with successful down-staging being maintained according to imaging criteria
2. A patient with acute hepatic decompensation after down-staging treatment is not eligible for LT unless criteria for successful down-staging and minimal observation period are met

In the future, there will be new technologies available to access tumor DNA in the circulating blood. HCC recurrence after LT is due either to the growth of occult metastasis or to the engraftment of circulating tumor cells.[48] Selecting LT recipients with low baseline levels of circulating HCC cells and adding biological markers to the accepted combination of morphological criteria may further limit the risk of HCC recurrence after LT.[48]

Conclusion

The results of recent published data on HCC down-staging support the notion that down-staging selects a subset of patients who could achieve posttransplant survival similar to those meeting LT criteria without down-staging. Multicenter studies using standardized criteria are needed to determine the feasibility of down-staging beyond the single-center experience. Based on review of published data and expert opinion in national and international conferences, proposed guidelines for a standardized down-staging protocol are presented (Table 5). Preliminary results of a multicenter study using a uniform down-staging protocol have shown favorable post-LT outcomes after tumor down-staging with no significant center effect,[49] thus supporting the feasibility of a broader application of down-staging before LT.

Acknowledgment

We thank Nancy Tran, Lynn Remedios and Crystal Yao for technical and editorial assistance in preparing the manuscript.

Ancillary