Refining the approach to down-staging of HCC prior to liver transplantation: Patient selection, loco-regional treatments, and systemic therapies

LIVER TRANSPLANTATION FOR HCC

The incidence of HCC has been steadily rising in the United States and much of the world for over 2 decades.¹ Liver transplantation (LT) remains the optimal treatment strategy for patients with early-stage HCC by providing the best oncologic resection while also restoring normal hepatic function. Since 1996, the Milan criteria (1 lesion ≤ 5 cm and 2-3 lesions ≤ 3 cm)² have been the benchmark for the selection of candidates with HCC for LT who can achieve excellent post-LT survival, with a relatively low risk of posttransplant tumor recurrence. After the implementation of the MELD allocation scheme with priority listing for HCC meeting Milan criteria in 2002, HCC has become the second most common indication for LT, accounting for more than 25% of all LTs performed in the United States.³

In recent years, there have been several proposals to modestly expand tumor size limits to offer LT as a curative therapy to more patients with HCC.^4,5 Among the proposed expanded criteria based on tumor size and number, the University of California, San Francisco criteria (1 lesion ≤ 6.5 cm and 2-3 lesions ≤ 4.5 cm with total tumor diameter ≤ 8 cm)⁶ and the Up-to-7 criteria have been independently validated.⁷ Examples of other expanded selection criteria include total tumor volume ≤ 115 cm³,⁸ the Kyoto criteria,⁹ extended Toronto criteria,¹⁰ and National Cancer Center Korea criteria,¹¹ most of which attempt to include laboratory or imaging surrogates for tumor biology such as biomarkers [eg, AFP and des-gamma-carboxy prothrombin] and ¹⁸F-FDG PET scan rather than rely on tumor burden alone. Expansion of the tumor size criteria alone without incorporating other biomarkers or strategies to further refine selection criteria appears to result in incrementally decreased post-LT survival with increasing tumor burden.^12–15

HCC DOWN-STAGING RATIONALE AND OUTCOMES

In the setting of worldwide organ shortages, expanding LT indications for HCC to meet the growing demand while ensuring acceptable post-LT outcomes has been challenging. One of the limitations of the expansion of tumor size limits alone is that it does not account for the effects of local-regional therapy (LRT). Given increasing wait times, neoadjuvant LRT is typically used as a bridge to control tumor growth and reduce the risk of waitlist dropout.¹⁶ In patients with HCC initially exceeding conventional LT criteria, down-staging is defined as a reduction in tumor burden using LRT to meet acceptable LT criteria (most commonly Milan criteria). Thus, patients with HCC being considered for down-staging can include those with Barcelona Liver Clinic Cancer (BCLC) stage A solitary tumor > 5 cm as well as those with BCLC intermediate-stage B with multinodular disease with at least 1 tumor > 3 cm. As the evidence supporting the role and efficacy of down-staging has strengthened in the past decade, the concept of down-staging was incorporated into the latest BCLC staging update.¹⁷

Tumor response to down-staging treatment(s) should be based on multiphase contrast-enhanced CT or MRI measurement of the size of only viable tumors, and the measurements should not include the area of necrosis resulting from LRT.¹⁸ While tumor progression despite LRT is associated with worse post-LT outcome,^19–21 successful down-staging serves as an effective tool to select a subgroup of patients with more favorable tumor biology who will likely do well after LT.¹⁸ According to an international consensus conference on LT for HCC, LT, after successful down-staging, should achieve a 5-year survival rate comparable to that of patients with HCC initially meeting conventional LT criteria and not requiring tumor downstaging.²²

Post-LT outcomes for those successfully down-staged using pre-LT LRT have not been significantly different from those always within Milan, especially when prespecified upper limits of tumor burden are employed. In a systematic review of nearly 1000 patients across 13 eligible studies not limited to tumor size or number, down-staging had a pooled success rate of 0.48 and a pooled post-LT HCC recurrence rate of 16%.²³ In the largest single-center experience²⁴ and in a follow-up multicenter study from region 5,²⁵ prespecified down-staging criteria included those with a single lesion ≤ 8 cm, 2–3 lesions < 5 cm, or 4–5 nodules all < 3 cm, with total tumor diameter < 8 cm. In these studies, >60% were able to be successfully down-staged to within Milan criteria with excellent 5-year posttransplant survival approaching 80% with no observed center-specific differences. In 2017, the United Network for Organ Sharing (UNOS) implemented these same down-staging tumor criteria nationally with patients eligible for automatic priority LT listing with HCC exception after successful down-staging to within Milan criteria after the mandated 6-month wait period (Table 1). A recent study using the UNOS national database²⁶ confirmed this policy whereby 3-year posttransplant survival was 79% for those initially meeting United Network for Organ Sharing down-staging (UNOS-DS) criteria and down-staged before LT compared to 83% for those always within Milan (p = 0.17). In addition, a large retrospective study ²⁷ including 20 LT centers and nearly 800 patients initially beyond Milan criteria demonstrated a 5-year recurrence rate of <20% in successfully down-staged patients (n = 465) with inferior overall and recurrence-free survival in the subset of patients who were not down-staged prior to LT (n = 324). Most recently, a large, multicenter analysis of 2645 patients with HCC,including 341 who were successfully down-staged into Milan on pre-LT imaging and 182 patients who were not down-staged, was able to evaluate long-term post-LT survival and recurrence outcomes.²⁸ The authors found 10-year post-LT survival of 61% in those always within Milan, 52% in the down-staged group, and only 43% in the not down-staged cohort. Similarly, 10-year post-LT recurrence estimates were 13% in the Milan group and 21% in the down-staged group, whereas >40% in the non-down-staged group experienced tumor recurrence.

RECENT MULTICENTER DOWN-STAGING PROSPECTIVE TRIALS AND LOCO-REGIONAL TREATMENT CONSIDERATIONS

The most commonly employed LRT options for down-staging are performed trans-arterially, including Y-90 radio-embolization (TARE) and chemo-embolization (TACE). TARE uses glass or resin microspheres to induce profound radiation and anti-tumor effect with little ischemic damage, whereas TACE can be delivered with cytotoxic agent(s) mixed with lipiodol and gelfoam particles (conventional or cTACE) or with doxorubicin drug-eluting beads-TACE). Advantages of these 2 modalities include the ability to target large (>5 cm) lesions and/or multiple lesions in a single territory with coverage of satellite nodules. The PREMIERE trial²⁹ was a small, single-center, randomized controlled trial comparing conventional TACE with TARE, primarily in patients who were potentially eligible for bridging or down-staging to LT. Median time to progression was significantly longer with TARE at >26 months compared to 7 months with TACE, but there was no significant difference in radiographic response rate or overall survival. A recent multicenter prospective study from the Multicenter Evaluation of Reduction in Liver Size before Liver Transplantation (MERITS-LT) consortium³⁰ enrolled 209 consecutive patients from 7 transplant centers in 4 UNOS regions who met UNOS-DS criteria and underwent initial down-staging treatment from 2015 to 2019. Probability of successful down-staging to Milan criteria within 2 years from the initial down-staging procedure exceeded 80%, with no differences seen in mRECIST radiographic response, probability of or time to successful down-staging, waitlist dropout, or LT when comparing TACE (n = 132) and TARE (n = 62) as initial treatment. Further, there were no significant differences in explant histology based on the type of the first LRT received, though TARE patients had a higher proportion with completely necrotic tumor(s) (31% vs. 21%) and a lower proportion with both tumor beyond Milan criteria (23% vs. 43%) and microvascular invasion (8% vs. 21%) (all p > 0.25). A second recent multicenter down-staging prospective trial³¹ showed the clear benefit of LT after down-staging in patients initially exceeding Milan criteria. In this study, 74 patients were down-staged and then subsequently randomized to LT versus non-LT therapies. The study demonstrated a 5-year overall survival of nearly 80% in the LT arm versus just over 30% for those randomized to non-LT therapy (HR 0.32, 95% CI 0.11–0.92; p = 0.035).

The risk of hepatic decompensation due to LRT should be considered when selecting patients for down-staging since a decrement in residual liver function can be treatment limiting and shorten survival, especially if successful down-staging is not achieved. It has been proposed that only patients with adequate hepatic function (eg, Child’s class A/B and bilirubin ≤ 3 mg/dL) should undergo attempted down-staging.¹⁸ Several studies have assessed the use of TACE as a bridge to LT in patients with HCC, with mildly decompensated cirrhosis^32,33 with patients found to be at an increased risk for treatment-related hepatic decompensation, though safe and successful treatment can be accomplished using a super-selective approach. Impaired liver function after receiving multiple TACE procedures for intermediate-stage HCC has been shown to limit subsequent systemic treatment options, thus decreasing overall survival.^34,35 Studies for those with a limited hepatic reserve are lacking for TARE in which the typical bilirubin cutoff is 2–3 mg/dL. Trans-arterial modalities may continue to be employed in the re-treatment of residual or recurrent disease following the initial round of LRT if underlying liver function remains optimal for further treatment with the goal of maintaining patients within LT criteria.¹⁸ Assessing treatment response after radiation-based therapy is particularly challenging as arterial enhancement and washout can persist for several months with increasing rates of necrosis over time.^36,37 Alternative approaches for the treatment of residual and/or recurrent tumor often include thermal ablation and stereotactic body radiotherapy, depending on the size and location of the tumor(s) and adequacy of hepatic function for additional LRT.

While tumor ablation (including radiofrequency, microwave, and cryoablation) loses its efficacy with larger tumors,³⁸ several Asian studies have shown that combination therapy with TACE followed by radiofrequency ablation in BCLC early/intermediate-stage patients with >3 cm HCC is generally safe and leads to superior outcomes compared to TACE alone.^39–41 For example, in a single-center retrospective analysis of 128 patients undergoing TACE plus radiofrequency ablation compared to 271 receiving TACE alone, Ren et al³⁹ found improved overall and progression-free survival at 5 years not only for patients with a tumor diameter of 3.1–5 cm but also for those with tumors > 5 cm. For this reason, combination TACE + ablation has been a common down-staging strategy in previous US multicenter studies used in ∼30%–40% of patients,^25,30 though this may decrease with increasing use of Y-90 radioembolization.

PREDICTORS OF INFERIOR DOWN-STAGING-RELATED OUTCOMES

Although the incorporation of standardized down-staging criteria (i.e., UNOS-DS) into national policy has largely been a success story, there are several important scenarios for which down-staging has shown less promise. Most notably, upper limits in tumor burden probably exist beyond which attempted tumor down-staging is less likely to be successful. The term “All-comers” has been defined as those exceeding UNOS-DS criteria with any number of tumors, with a total tumor diameter > 8 cm but without extra-hepatic disease or macrovascular invasion. In a single-center analysis of the “all-comers” criteria, the probability of tumor down-staging to Milan criteria decreased with increasing tumor burden such that < 50% could be successfully down-staged if the sum of the largest lesion plus number of lesions was >12.⁴² In a follow-up analysis of the prospective MERITS-LT consortium,⁴³ a significantly lower probability of successful downstaging of 67% at 1 year (from initial treatment) was observed in the “all-comers” group (n = 82) compared to 83% in those initially meeting UNOS-DS criteria (n = 229). In this study, for every 1 unit increase in the size of the largest lesion (cm) plus number of lesions, the probability of successful downstaging dropped by 14%. Not surprisingly, patients with HCC initially beyond Milan who were successfully down-staged have a higher risk of waitlist dropout compared to those within Milan from the outset. In a national analysis,⁴⁴ the probability of dropout within 2 years of listing was 26% for UNOS-DS and 32% for “all-comers” compared to < 20% in Milan patients (p < 0.001). Further, only 15% of “all-comers” ultimately received LT in a single-center analysis due to high rates of tumor progression leading to waitlist dropout.⁴² For this reason, intention-to-treat survival at 5 years from the initial down-staging procedure was much worse in the “all-comers” cohort at 20% compared to nearly 60% of patients meeting UNOS-DS criteria.⁴²

Not only does the likelihood of successful down-staging and receipt of LT decrease with increasing initial tumor burden, both explant pathology and post-LT survival can be compromised as well. In the MERITS-LT consortium³⁰ and in the UNOS database,^26,45 at least 30%–40% of those requiring tumor down-staging prior to LT were found to have tumor under-staging to beyond Milan criteria in the explant. This is particularly important as multiple explant-based prognostic models have demonstrated worse post-LT survival related to tumor under-staging in the explant.^46–48 In patients who require tumor down-staging, the higher the tumor burden on the last imaging study prior to LT, the greater the risk of under-staging on explant pathology.^26,30 Therefore, it has been proposed that down-staging to within Milan criteria should be the minimal requirement for LT, with a recommendation to perform additional LRT to further reduce the viable tumor burden and ideally to achieve complete tumor necrosis prior to LT.¹⁸ Regarding post-LT outcomes, 3-year post-LT survival in the UNOS database was similar for Milan and UNOS-DS patients, as noted above, but post-LT HCC recurrence was higher in the UNOS-DS groups (13% vs. 7%).²⁶ In addition, 3-year post-LT in the UNOS database was significantly worse for “all-comers” patients at 70% compared to ~80% in the Milan and UNOS-DS groups.²⁶ Given that liberalizing down-staging inclusion criteria results in a lower rate of successful downstaging and a higher rate of waitlist dropout⁴² as well as inferior post-LT survival,²⁶ patients initially exceeding UNOS-DS criteria are considered for MELD exception after successful down-staging only on a case-by-case basis by the National Liver Review Board in the United States.

In addition to radiographic or explant tumor burden, AFP is an important predictor of tumor-specific survival. Higher AFP levels have been consistently identified as a risk factor for waitlist dropout^49–51 and as a negative predictor of post-LT outcome.^48,52–54 In the down-staging population in particular, AFP > 100 ng/mL has been identified as the threshold associated with a higher risk of HCC recurrence and death, with 3-year post-LT survival of 60% compared with >80% among patients with HCC with AFP < 20 ng/mL at the time of LT.²⁶ In the MERITS-LT down-staging consortium,³⁰ pre-treatment AFP-L3 ≥ 10% (HR 3.7, p value 0.02) has been associated with increased dropout risk. Whether more stringent AFP criteria should be applied to down-staging candidates is not yet clear, but AFP and potentially other biomarkers may be helpful for further risk stratification in the down-staging population. In adition, previous studies have identified decompensated cirrhosis as a significant predictor of unsuccessful down-staging^24,25 presumably due to receipt of less aggressive treatments, given concerns for further decompensation following LRT and increased risk of liver-related death regardless of whether they receive LRT. In the multicenter prospective study from UNOS region 5,²⁵ down-staging treatment failure was observed in all Child-Pugh B/C patients with pre-treatment AFP ≥ 1000 ng/mL; thus, these patients should typically be excluded from attempted down-staging with LRT. It is unknown whether high AFP predicts systemic therapy downstaging failure.

ADVANCES IN SYSTEMIC THERAPIES FOR INTERMEDIATE-TO ADVANCED STAGES OF HCC

For patients with more advanced stages of HCC not eligible for liver transplant, systemic treatment options have expanded rapidly over the past 5 years with the advent of molecularly targeted therapies, which inhibit tumorigenic pathways, including angiogenesis, as well as immunotherapies that promote an immune response against the cancer.^17,55 Targets of immune checkpoint inhibitors (ICI) include PD-1 (nivolumab and pembrolizumab), PD-L1 (atezolizumab and durvalumab), and CTLA-4 (ipilimumab and tremelimumab). Recently, the combination of targeted therapies with immunotherapy^56–59 as well as immunotherapy agents alone^60,61 have demonstrated the ability to achieve deep and durable tumor shrinkage in a subset of patients with advanced disease.

For nearly 10 years, the multikinase inhibitor sorafenib⁶² was the only approved therapy for advanced-stage HCC despite showing only modest survival benefits. In recent years, multiple new systemic treatments targeting angiogenesis have been approved for intermediate-to advanced-stage HCC based on survival improvements in phase 3 trials in the first-line and second-line systemic therapy contexts, including the multi-targeted tyrosine kinase inhibitor (TKI), lenvatinib⁶³ as a first-line therapy option, and the TKI regorafenib⁶⁴ and cabozantinib⁶⁵ and as well as the VEGFR2-targeted monoclonal antibody, ramucirumab⁶⁶ as second-line or later-line options after progression on sorafenib.

With growing evidence for the activity of ICI immunotherapy across a range of solid tumors, clinical trials of ICI in HCC were undertaken, which demonstrated that ICI as monotherapy could elicit robust and durable responses in a subset of ∼15%–20% of patients with advanced stages of HCC.^61,67–69 In phase 3 trials, the anti-PD-1 ICI, pembrolizumab, improved overall survival compared with placebo in an Asian population after progression on prior sorafenib,⁷⁰ and the anti-PD-L1 and anti-PD-1 ICI, durvalumab and tislelizumab, respectively, both showed non-inferiority for overall survival outcome compared to sorafenib in the first-line setting.^61,71

The addition of an anti-angiogenic agent to ICI can promote vascular normalization and inhibit immune-suppressive elements of the tumor microenvironment, augmenting anti-tumor immunity.^72,73 Preclinical evidence for increased activity in combination with ICI has been validated in clinical trials demonstrating higher rates of objective radiographic response with the combination of ICI with bevacizumab, a monoclonal antibody targeting VEGF, or with anti-angiogenic TKI including lenvatinib⁵⁸ or rivoceranib.⁵⁷ The combination of atezolizumab plus bevacizumab is the first ICI combination regimen to improve survival over sorafenib in HCC, achieving median overall survival of 19.2 months and objective radiographic responses in 30% of patients, including 8% with a complete radiographic response.^56,74

Combinations of anti-CTLA-4 monoclonal antibodies with anti-PD-1 or anti-PD-L1 monoclonal antibodies have also shown potential for augmented rates of immune response in HCC, including the combination of the anti-PD-1 targeted monoclonal antibody, nivolumab, with the anti-CTLA-4 targeted monoclonal antibody, ipilimumab, which demonstrated objective responses in 32% of patients in a phase 2 trial.⁶⁰ A randomized, phase 2 trial of the CTLA-4 inhibitor, tremelimumab, combined with durvalumab showed higher rates of radiographic response for the combination when compared with monotherapy arms, with objective responses occurring in 24% of patients treated with the combination;⁷⁵ the ensuing, randomized, phase 3 trial of this combination demonstrated a significant survival benefit for the combination over sorafenib, with a median of 16.4 versus 13.8 months, respectively (HR, 0.78), leading to regulatory approval.⁶¹

EFFICACY OF ICI COMBINATIONS IN EARLY-STAGE HCC: STUDIES IN NEOADJUVANT AND ADJUVANT SETTINGS

The substantial progress achieved by ICI combinations in systemic therapy for advanced stages of HCC has led to the re-examination of the role of systemic therapy in earlier stages,⁷⁶ including as neoadjuvant therapy before surgery and in the adjuvant setting after definitive resection or ablation. Response outcomes and translational analyses of ICI in other tumor types suggest a potential for greater efficacy in earlier lines of therapy, perhaps due to the presence of a more robust immune response before receipt of cytotoxic therapies, further supporting the study of ICI in earlier stages of HCC, as well.

A phase 1b study evaluated the combination of nivolumab plus cabozantinib for 8 weeks in 15 patients with high-risk HCC being evaluated for resection.⁷⁷ Twelve of 15 patients (80%) underwent successful resection, and 5 of 12 (42%) experienced major pathologic responses. Immune profiling of resected tumors demonstrated enrichment for effector T cells and tertiary lymphoid clusters as well as an increase in CD138 ⁺ plasma cells and a distinct B-cell distribution in responders compared to nonresponders, suggesting the involvement of B cells as well as T cells in the immune response. A randomized, phase 2 study examined 6 weeks of neoadjuvant nivolumab given every 2 weeks or the combination of nivolumab every 2 weeks plus ipilimumab for 1 dose before surgery in 27 patients with resectable HCC.⁷⁸ Twenty patients underwent successful surgery, while 7 were canceled due to the progression or lack of surgical candidacy for other reasons; no patients had surgery canceled due to complications of therapy. Of the 20 patients who underwent resection, major pathologic response was achieved in 3/10 (33%) of patients treated with neoadjuvant nivolumab, and in 3/11 (27%) of patients treated with neoadjuvant nivolumab plus ipilimumab. One patient had a complete pathologic response despite tumor enlargement on preoperative MRI, which was determined to be due to immune infiltration. Postoperatively, patients received adjuvant immunotherapy according to their original treatment assignment for up to 2 years. Median progression-free survival was 9.4 months in the nivolumab group and 19.5 months in the combination group. Immune profiling of resected tumors demonstrated increased immune infiltration in patients achieving a major pathologic response.

In sum, these studies provide early evidence for the activity and potential utility of perioperative ICI combination therapy in HCC. Studies are ongoing to determine the efficacy of ICI as monotherapy or in combinations in the adjuvant setting as well as neoadjuvant.

RISKS OF ICI COMBINATIONS IN PERIOPERATIVE HCC MANAGEMENT

A paramount consideration in the perioperative context is the toxicity of individual systemic therapies and combination regimens, including the risk of incurring hepatic decompensation or other complications. Anti-angiogenic TKI and bevacizumab confer risk of toxicity, including bleeding, impaired wound healing, and varying degrees of hepatotoxicity. In phase 3 trials of the available TKI agents sorafenib, lenvatinib, cabozantinib, and regorafenib in HCC, grade 3 or higher hepatotoxicity occurs in up to around 12% of patients, though comparisons between trials are limited by heterogeneity in protocol and reporting.^62–65 Bleeding rates for TKI are relatively low, with the highest rate of grade 3 or higher hemorrhage reported for regorafenib in 6%, though this frequency was lower than that of the placebo arm, in which 8% of patients had grade 3 or higher bleeding events.⁶⁴ In the atezolizumab plus bevacizumab regimen now established as the preferred first-line therapy in HCC,^56,74 concern for bleeding risk led to the eligibility requirement for patients to undergo an upper endoscopy within 6 months before starting treatment to exclude patients with high-risk varices or other bleeding risk. In this carefully selected population, upper gastrointestinal bleeding occurred in 7% of patients treated with the combination regimen versus 4.5% for sorafenib alone, and 30% of patients had bleeding events of any grade in the combination group compared to 18% for sorafenib. Rates of toxicity from TKI and bevacizumab may be higher and overall survival lower in patients with greater degrees of baseline hepatic dysfunction, with trends toward higher rates of ascites, encephalopathy, and bleeding seen in Child-Pugh B subgroups; prospective data are limited owing to the exclusion of patients with Child-Pugh B or worse liver dysfunction from most phase 3 trials.^79–82

Immunotherapy with ICI confers the distinct risk of immune-mediated adverse events (IMAE) attributed to immune activation and inflammation occurring in normal tissues, including the risk of immune-mediated hepatitis. For ICI as monotherapy, the risk of IMAE requiring steroid immunosuppression has been reported to be 10% or less.^58,61 ICI combinations that include anti-CTLA-4 monoclonal antibodies, generally incur higher rates of immune-related toxicity, which is dose related.⁸³ For the combination of durvalumab given with a single high dose of tremelimumab, the rate of IMAE requiring systemic steroids was 20%, including grade 3-4 transaminitis in around 5% of patients.⁶¹ For the combination of nivolumab 1 mg/kg combined with ipilimumab with a dosage of 3 mg/kg dose for 4 doses every 3 weeks followed by monthly nivolumab, the steroid requirement was over 50%, with grade 3 or 4 immune-mediated hepatic events occurring in 20% of patients.⁶⁰ These combinations of ICI regimens incorporating CTLA-4 inhibition have not been studied prospectively in patients with greater degrees of hepatic dysfunction to date.

In the neoadjuvant setting, the risk of adverse events that compromise surgical candidacy is of paramount importance. In the single-center neoadjuvant studies cited above, there were no adverse events that precluded proceeding to resection after neoadjuvant treatment with cabozantinib plus nivolumab⁷⁷ or with nivolumab, with or without ipilimumab.⁷⁸

SPECIFIC RISKS OF PREOPERATIVE SYSTEMIC THERAPY IN LT CANDIDATES: PRE-TRANSPLANT HEPATIC DYSFUNCTION

While the improvements in the efficacy of modern systemic therapies, including ICI combinations in HCC, prompt the consideration of their use in earlier stages of disease, including in patients undergoing attempted down-staging,⁷⁶ there are several specific concerns related to its use in this population. First, most randomized trials of systemic HCC treatments have predominantly enrolled patients with compensated liver disease (ie, Child-Pugh A), given the concerns of hepatic decompensation with experimental therapies in patients with a greater degree of hepatic dysfunction at baseline. Owing to the highly selected patient population as well as the challenges in distinguishing between hepatic decompensation due to tumor burden versus drug-related toxicity, the risk of hepatic decompensation is not a clinical end point routinely assessed in systemic therapy trials.⁸⁴ This is unfortunate as several studies have shown the importance of hepatic decompensation as a driver of increased mortality in patients with HCC, ^85–87 including patients with advanced HCC being treated with systemic therapy.⁸⁸ Not surprisingly, rates of ascites and encephalopathy are higher in Child-Pugh B patients treated with sorafenib compared to Child-Pugh A patients.⁸⁹ Specific treatment regimens that may be appropriate as the first line of systemic therapy for unresectable Child-Pugh B patients who are determined to be not candidates for ICI combination regimens due to hepatic impairment may include single agent PD-1 or PD-L1 blockade (such as with durvalumab, nivolumab, or pembrolizumab)⁹⁰ or sorafenib or lenvantinib, with consideration of empiric starting dose reduction depending on the degree of hepatic dysfunction.^79,91,92

ICI monoclonal antibodies have demonstrated acceptable safety in the setting of varying types and degrees of organ dysfunction, including hepatic dysfunction. A phase 1/2 trial of nivolumab monotherapy in 49 Child-Pugh B7/8 patients demonstrated comparable safety to Child-Pugh A patients, with objective response seen in 12% and tumor control observed in 55%, suggesting suitability in certain Child-Pugh B patients.⁹³ A retrospective case series of 18 patients with Child-Pugh B hepatic dysfunction treated with nivolumab for advanced stages of HCC demonstrated similar rates of immune-related toxicity though higher rates of all-cause adverse events, with grade 3 or higher adverse events (predominantly unrelated to treatment) occurring in over 90% of patients; the objective response rate was 17% in this small series.⁹⁴ There are no prospective studies reporting on the safety of durvalumab, tislelizumab, or pembrolizumab in patients with Child-Pugh B hepatic dysfunction to date, though no dosage adjustments are required for mild-moderate hepatic dysfunction based on no significant impact in pharmacokinetic studies.

The combination of atezolizumab plus bevacizumab introduces additional risks of anti-angiogenic therapy associated with bevacizumab, as well as ICI risk. A recent multicenter trial of consecutive patients treated with atezolizumab plus bevacizumab⁹⁵ showed comparable rates of treatment-related adverse events in the 48 Child-Pugh B patients compared to the 154 Child-Pugh A patients with similar objective response rates and disease control rates across Child-Pugh class as well. In patients at high risk of gastrointestinal bleeding, however, such as those with untreated varices or recent bleeding events, or patients with portal hypertension who have not received recent upper endoscopy, the combination of atezolizumab plus bevacizumab should not be initiated given the risk of exacerbating bleeding risk with VEGF inhibition. This concern is based on prior phase 2 studies of bevacizumab in advanced HCC, including varying proportions of Child-Pugh A and B patients without consistent screening endoscopy requirements, in which grade 3 or higher hemorrhagic complications, including upper gastrointestinal bleeding occurred in 7-19% of patients.^96,97

SPECIFIC RISKS OF SYSTEMIC THERAPY IN LT CANDIDATES: POSTTRANSPLANT REJECTION

In the context of LT candidates, another significant safety concern for ICI-based therapies is the risk of post-LT acute rejection and graft failure. The inhibition of the PD-1 or PD-L1 checkpoint on the tumor, liver, and immune cells can result in acute T-cell–mediated allograft rejection.⁹⁸ In patients who have already undergone liver transplantation for HCC or other indications, case series suggest that acute T-cell–mediated rejection occurs in at least 30% of cases treated with ICI for new or recurrent cancers and that rejection risk may be higher in patients whose liver allograft is positive for PD-L1 expression.^99,100 Due to the high risk of fatal rejection, prior LT is generally regarded as an absolute contraindication to the implementation of subsequent ICI therapy.

In the pre-LT setting, the risk of rejection may be influenced by a variety of factors, including the time elapsed since the last dose of ICI therapy before the introduction of foreign allograft antigens. A recent case series of 5 patients with HCC receiving pre-LT immunotherapy, with nivolumab at UCSD found that both patients who received nivolumab within 3 months of LT developed acute cellular rejection and severe hepatic necrosis, one of whom required repeat LT. None of the patients who underwent LT > 3 months from the last dose of immunotherapy developed rejection or graft loss.¹⁰¹ A recent report from Mt. Sinai¹⁰² described 9 patients with HCC who underwent LT after receiving nivolumab as an element of pre-LT tumor treatment with nearly all patients receiving their last dose within 4 weeks of LT. Four of the 9 patients were down-staged into Milan criteria and 8 of the 9 also received LRT. After a median follow‐up of 16 months after LT, no severe allograft rejections/loss, tumor recurrence, or deaths occurred. Based on these reports, UNOS issued a recent guidance update that the use of immunotherapy should not exclude patients with HCC from receiving exception points and undergoing LT. However, it does appear that a washout period of ∼12 weeks prior to LT should be considered, especially given the relatively long half-lives, which can approach 4 weeks for commonly used anti-PD-1 and anti-PD-L1 ICI monoclonal antibodies, including atezolizumab, durvalumab, nivolumab, and pembrolizumab; T-cell activation and proliferation can persist for longer periods, leading to a risk for late toxicity.^83,103 For this reason, immunotherapy would not be recommended if a short wait time to transplant is expected, such as in the case of live donor LT or if a patients with HCC has a high laboratory MELD score. Further, given the higher risk for IMAE associated with ICI combinations, including anti-CTLA-4 agents and the paucity of data in the pre-LT context, anti-CTLA-4-containing combinations would not be recommended in the down-staging context at this time, except through clinical trials.

COMBINATION SYSTEMIC THERAPY WITH LRT: BIOLOGIC RATIONALE, DATA, AND ONGOING TRIALS

Systemic therapy is typically reserved for patients with HCC who either are not suitable for LRT or who have tumor progression despite LRT. However, improved survival with recently approved combination systemic therapies coupled with inferior tumor-related outcomes in certain patients with intermediate-stage HCC being considered for LT suggests that rethinking this dynamic may allow for curative therapy (ie, LT) in those who otherwise would have a median expected survival of only 2–3 years.¹⁷ There is certainly a biologic rationale for combining LRT with systemic therapy. For example, enhanced antigen presentation from LRT-induced tumor cell necrosis can stimulate peripheral immune response, including the activation of dendritic cells.^55,104 This appears to enhance the effect of ICIs such as tremelimumab, leading to the accumulation of intratumoral CD8⁺ T cells in patients with clinical response.¹⁰⁵ In addition, increased hypoxia after LRT-induced ischemia leads to the upregulation of angiogenic factors¹⁰⁶ that could render VEGF inhibitors (eg, bevacizumab) more effective. Finally, TARE appears to induce adaptive and innate immune activation of the local HCC microenvironment, particularly in both T and NK cells that could provide an ideal setting to follow TARE with ICI therapy to improve local tumor control.^107,108

Several previous studies have evaluated combination LRT with systemic therapy with mixed results to date. The phase 2 prospective, randomized trial SORAMIC trial¹⁰⁹ compared TARE and sorafenib (n = 216) versus sorafenib alone (n = 208) and showed similar median overall survival between groups (12.1 vs. 11.4 months, respectively). Similarly, the SPACE¹¹⁰ and TACE-2¹¹¹ prospective randomized trials, which included over 300 patients each with intermediate-stage HCC, both failed to show an improvement in overall survival in those treated with drug-eluting beads-TACE plus sorafenib compared to those treated with drug-eluting beads-TACE plus placebo. More recently, the prospective phase 2 TACTICS trial¹¹² compared TACE with or without sorafenib in 156 patients with HCC without extra-hepatic disease or vascular invasion, with 88% meeting BCLC stage A or B. In contrast to most prior studies, sorafenib was initiated 2–3 weeks before initial TACE, and the study design allowed for on-demand/as-needed additional TACE sessions throughout the study period. Progression-free survival, defined as tumor progression that was “unTACEable,”, was significantly longer in the TACE plus sorafenib arm (25.2 mo) compared to the TACE alone group (13.5 mo) with improvement in 2-year overall survival as well (83% vs. 65%). The authors concluded that TACE, in combination with sorafenib, should not be terminated at the point of intrahepatic tumor progression as TACE appears to still be effective in this setting. Recent retrospective studies have also suggested a potential benefit of TACE plus lenvatinib compared to either treatment as monotherapy^113–116 though patients offered TACE presumably had less advanced-stage HCC while those offered combination therapy perhaps had improved liver function.

Given the exciting prospect of improving outcomes for patients with HCC with intermediate-stage HCC, there are multiple ongoing randomized clinical trials evaluating the combination of TACE plus systemic therapy in this population (Table 2). On the other hand, studies assessing the combination of LRT with systemic therapy in the pre-transplant population are limited.⁷⁶ Hoffman et al¹¹⁷ performed a randomized control trial of TACE plus sorafenib versus TACE alone and showed no difference in time to tumor progression, but this trial was limited to those meeting the Milan criteria. Ongoing studies in patients with intermediate-stage HCC exceeding the Milan criteria who are listed for LT include PLENTY202001, an unblinded randomized controlled trial that will evaluate pembrolizumab plus lenvatinib combination therapy, as well as ESR-20-21010, a single-arm, open-label phase 2 multicenter trial that will evaluate durvalumab plus tremelimumab, with a mandated washout period of at least 10 weeks before LT. In addition, DULECT2020-1, a small trial, will evaluate durvalumab plus lenvatinib in patients with HCC, with locally advanced HCC before liver transplant. Certainly, patients with HCC with intermediate-stage HCC should be a priority group for recruitment in prospective trials to determine whether a combination of LRT with systemic therapy can improve waitlist and potentially post-LT outcomes.

PROPOSED INCORPORATION OF SYSTEMIC THERAPY INTO LT DOWN-STAGING AFTER LRT IN HIGH-RISK PATIENTS

Given inferior outcomes with LRT alone in certain high-risk patients initially exceeding Milan criteria, the incorporation of systemic therapy in down-staged patients with residual HCC tumors could improve transplant-related or advanced disease-related outcomes by delaying time to progression or achieving further tumor shrinkage while reducing the risk of cumulative hepatotoxicity from serial LRT procedures. Further, in the subset of patients with HCC who are not able to be down-staged with LRT, transition to systemic therapy followed by objective response may result in successful tumor down-staging to facilitate wait-listing. In patients who are deemed not transplant candidates, earlier initiation of active systemic therapies may improve survival. Therefore, we propose the following protocol outlining an approach to down-staging using both LRT and systemic therapy with a focus on these high-risk patients.

As outlined in Figure 1, participants initially beyond UNOS-DS criteria (ie, within “all-comers”; Table 1) who have undergone successful down-staging of HCC to within Milan criteria with Child-Pugh class A or B liver function being considered for LT listing are appropriate for this protocol. In addition, select patients within UNOS-DS criteria who are successfully down-staged but thought to have a high risk of subsequent waitlist dropout (eg, elevated AFP and/or multi-focal HCC) or those not successfully down-staged to Milan criteria with LRT(s) but who do not have extra-hepatic disease or vascular invasion could also be included on a case-by-case basis. Of note, only patients with residual/recurrent disease should be included, whereas patients with suspected complete tumor necrosis after LRT (ie, mRECIST complete response; LR-TR non-viable disease) should be monitored closely with multiphase imaging performed at least every 3 months. Patients would be listed for LT once successful down-staging into Milan criteria was achieved with a standard, mandatory 6-month waiting period before granting of HCC MELD exception points (ie, MMAT-3).

In patients with residual tumor burden after LRT warranting transition to systemic therapy, the choice of systemic therapy should be determined based on standard-of-care eligibility for systemic therapy based on the diagnosis of unresectable HCC assessed as not appropriate for further LRT due to size, location, or hepatic dysfunction. Compensated, Child-Pugh A patients would be eligible to receive atezolizumab plus bevacizumab, assuming no recent GI bleeding and either treated varices or without varices on recent upper endoscopy; alternate options for first-line therapy would include durvalumab with or without a single dose of tremelimumab, lenvatinib, or sorafenib in patients not eligible for atezolizumab plus bevacizumab due to liver dysfunction, performance status, or other considerations.⁹⁰ Child-Pugh B patients could receive approved ICI monotherapy such as pembrolizumab or durvalumab;⁹⁰ alternate options would be with sorafenib or lenvantinib with consideration of empiric starting dose reductions in patients with high MELD and short predicted waiting list time. Eligible patients would receive systemic treatment until LT is predicted to occur within about 12 weeks for patients receiving ICI, based on the case series data detailed above, or until the development of unacceptable toxicity; the duration of therapy in patients with complete radiographic response can be determined on a case-by-case basis. Systemic treatment with bevacizumab should be discontinued at least 4 weeks before predicted LT, and at least 7 to 10 days (approximating 5 drug half-lives) prior to the predicted LT date for TKI therapies, or until the development of unacceptable toxicity, whichever comes first. Tumor progression beyond Milan criteria and leading to systemic therapy discontinuation would presumably result in waitlist removal. All on-treatment laboratory assessments and imaging would be performed as part of routine standard-of-care. At a minimum, cross-sectional imaging, MELD labs and liver enzymes, CBC, and AFP testing should be performed every three months. All imaging should be reviewed at the transplant center’s multidisciplinary tumor board. Patients would be eligible for additional LRT during the period in which they are on systemic therapy. This allowance for as-needed LRT mirrors the previously referenced TACTICS trial design,¹¹² which was one of the few studies to date to show an improvement in overall survival with TACE combined with sorafenib compared to TACE alone. Decisions on if/when to hold systemic therapy in patients scheduled for LRT would need to be made through multidisciplinary discussion. In the TACTICS trial,¹¹² sorafenib was discontinued for 2 days before and 2 days after each TACE session.

We anticipate participants’ typical wait time from listing to LT to be over 12 months in many regions, which includes the 6-month wait rule following tumor down-staging applied by UNOS. The donor age range should be adjusted (eg, 98–99 years of age) to ensure that listed patients do not receive LT offers while on systemic therapy or for the 12 weeks after treatment is stopped. Before LT, patients should have at least 1 contrast-enhanced CT or MRI abdominal scan and chest CT after at least 6–12 weeks of being off systemic therapy to ensure no rapid disease progression or extra-hepatic disease that would make a patient ineligible for LT. A major challenge with such a protocol is that the timing of deceased donor LT is somewhat difficult to predict but would need to be considered on a center-by-center basis, especially so as to not delay time to LT given the high risk of waitlist dropout in these patients.

In considering the utility of such a down-staging protocol that combines the effects of both local-regional and systemic therapy, there are many important end points and objectives worth studying. It will be particularly useful to assess the safety and tolerability of systemic therapy in this particular HCC patient population. Certainly, dose interruptions or reductions should be allowed for those on systemic therapy who experience treatment-related toxicity. In addition, while the safety of systemic therapy combining anti-VEGF/TKI and immunotherapy in Child-Pugh A patients is well established, the optimal regimen for those with Child-Pugh B disease is mostly unknown. This is even more relevant given that many of these patients will be receiving concurrent loco-regional treatment(s) that may further increase their risk of decompensation.

In terms of tumor-related outcomes, treatment response assessment using Liver Imaging Reporting and Data System treatment response algorithm after LRT or mRECIST¹¹⁸ in those receiving LRT or systemic therapy can help classify objective response as well as tumor progression. Notably, the objective response has been shown to predict improved patient survival in both patients receiving LRT^119,120 as well as those on systemic therapy.^121–125 Serum biomarkers, particularly AFP, are often used as an adjunct to radiographic assessment to monitor treatment response but optimal biomarker thresholds still need to be determined. It would also be important to evaluate the number (and type) of LRT required after initial successful tumor down-staging and, in the relevant subset, how often systemic therapy can achieve tumor down-staging in patients who are not able to be down-staged with LRT. Additionally, the evaluation of rates of waitlist dropout and successful LT, as well as intention-to-treat survival, is crucial. The reason for dropout (eg, tumor progression, liver-related death, and treatment-related toxicities including IMAE) should be noted, especially since 1 of the main limitations of the UNOS HCC database is that the reason for waitlist removal is not currently captured. In those who receive LT, explant pathology should be assessed to understand the proportion with high-risk pathology, including tumor under-staging to beyond T2 criteria, microvascular invasion, and poorly differentiated tumor grade, as well as features associated with pathologic response to ICI including immune infiltrate. In addition to the usual assessment of post-LT HCC recurrence and overall survival, relevant post-LT outcomes include rates of acute cellular rejection and graft failure, especially in those receiving immunotherapy within 3-6 months prior to LT.

CONCLUSIONS

Down-staging of HCC represents a structured approach that combines the expansion of tumor size limits beyond conventional criteria with the objective and sustained response to tumor-directed therapy as a risk stratification tool suggesting favorable tumor biology. Given the excellent post-LT outcomes in those successfully down-staged into conventional LT criteria, down-staging has gained broad acceptance in clinical practice.^17,126–128 However, the expansion of active systemic therapies, including anti-VEGF targeted therapies and TKI as well as immunotherapy with ICI alone or in combination, represent a potential lifeline for certain patients with intermediate-stage HCC initially exceeding LT criteria who have high rates of unsuccessful down-staging, waitlist dropout, explant under-staging, and post-LT recurrence. We are eagerly awaiting results from several ongoing randomized clinical trials to determine if the combination of LRT plus systemic therapy can improve outcomes for patients with intermediate-stage HCC, including patients receiving downstaging or bridging therapy before LT, compared to LRT alone.

In this review, we have proposed an expansion of the conventional down-staging approach that includes certain UNOS-DS patients with a high risk of dropout (e.g., multi-focal disease and elevated AFP) as well as “all-comers” who exceed UNOS-DS criteria. Introducing systemic therapy after initial successful down-staging (or if LRT alone does not achieve down-staging) may allow for curative therapy (ie, LT) in this relatively small subset of patients with HCC who have significant urgency for LT but otherwise would have poor expected survival. Characteristics related to the underlying liver disease (eg, Child-Pugh score), tumor biology (eg, size/number of lesions, treatment response, and AFP), and center-specific LT wait time should be taken into account when deciding on the type and length of systemic therapy administration and whether additional LRT should be offered during this period. Finally, given previously reported inferior post-LT outcomes in patients initially exceeding UNOS-DS criteria²⁶ coupled with safety concerns surrounding the use of pre-LT immunotherapy, it will be critical to ensure both acceptable short-term and long-term post-LT graft and overall survival. If so, including systemic therapies as an option after failure of LRT could allow for successful LT in a select subgroup of patients with HCC who are the most in need and would derive significant transplant survival benefits.