Liver transplantation as therapy for hepatocellular carcinoma

Liver transplantation can provide curative therapy in selected patients with hepatocellular carcinoma. Well‐established criteria include tumours that are within the Milan criteria and without evidence of vascular or extrahepatic involvement. Modest expansion of the original Milan criteria has been shown to achieve similar recurrence‐free survival rates. Overall, HCC recurrence occurs in about 10%‐15% of LT recipients, most within the first 2 years. Predictors of post‐transplant recurrence include high alpha‐foetoprotein, macrovascular invasion, as well as tumour size and number. Once HCC recurs after transplantation, prognosis is poor, though better if detected early. There is no established role for systemic prophylactic post‐transplant chemotherapy.


| INTRODUC TI ON
Liver transplantation (LT) was established as an effective therapy for small, hepatocellular carcinoma (HCC) in a landmark study by Mazzaferro and colleagues in 1996. Candidates had a single lesion no larger than 5 cm, or up to three lesions, no larger than 3 cm, without macrovascular invasion or spreading outside of the liver and survival was similar to that in patients transplanted without HCC. 1 Although the "Milan criteria" are extensively applied in transplant centres, many have expanded the criteria to provide more patients with treatment options while maintaining acceptable recurrencefree transplant survival. HCC is the indication for approximately 30% of the LT performed in the US and Europe.

| S ELEC TI ON OF B E S T C AND IDATE S FOR LIVER TR ANS PL ANTATION
The goal of allocation strategies is to minimize wait-list deaths while achieving acceptable post-LT graft and patient survival. With the well-recognized shortage of donated livers worldwide, prioritizing patients with HCC, many with compensated cirrhosis, vs those with decompensated cirrhosis, without creating a disparity in access is challenging. Policies continue to evolve. The Model of End Stage Liver Disease (MELD) score prioritizes patients on the LT list.
Patients requiring LT for HCC may be underserved on the transplant list by their natural MELD score, as they often do not have underlying decompensated liver disease. Patients with stage T2 lesions, defined as falling within the Milan criteria (Table 1) based upon lesion size and or growth characteristics (on cross-sectional, contrast imaging performed on equipment that meets minimal technical specifications), are eligible for MELD exception points in the US. The scans must be interpreted by a radiologist at a liver transplant centre. If the lesion is atypical, a biopsy must be positive for HCC. Once a patient is listed for LT with a MELD exception, there is a 6-month waiting period for implementation of the exception, partly to decrease the risk of being transplanted too early with an HCC that exhibits aggressive tumour biology resulting in a poorer post-transplant prognosis.  (Table 1). Of note, the model predicts a worse outcome for patients with hepatitis C virus (HCV) infection, but the study was performed in patients transplanted before direct-acting antivirals (DAAs).
Alpha-foetoprotein (AFP) has repeatedly been shown to identify patients with HCC with a high risk of recurrence after LT. 4,5 Patients with HCC within the Milan Criteria but an AFP >1000 ng/mL have a 5-year recurrence-free survival of only 20%. 5 Thus, these patients must demonstrate an AFP response to LRT to be considered for LT.
In a recent analysis of UNOS data evaluating patients undergoing downstaging of HCC, an AFP ≥100 ng/mL was an independent predictor of post-LT HCC recurrence. 4 Since patients may receive multiple treatments for HCC over an extended period of time, dynamic models could improve prediction of post-LT HCC recurrence. The Hazard Associated with Liver Transplantation for Hepatocellular Carcinoma (HALTHCC) score was developed to predict overall post-LT survival 6 and was recently validated in an international cohort. 7 It uses pre-LT AFP, MELD-sodium, and tumour burden as a continuous risk metric. There was a strong correlation between pre-LT risk and explant pathology predicting recurrence, but further prospective studies are needed to test this score.

| MANAG EMENT OF PATIENTS WITH H CC ON THE WAITING LIS T
Most patients on the waiting list with HCC will undergo LRT as a bridge to LT. Modalities include trans-arterial chemoembolization, radio-embolization, radiofrequency ablation, microwave ablation, and stereotactic body radiation therapy. LRT prevents wait-list dropoff because of tumour progression. In a propensity matched analysis from the UK, one session of LRT was associated with a 49% lower risk of delisting for progression of HCC and/or post-LT recurrence. 8 This advantage was lost if more than 3 LRTs were needed, which might suggest that poorer tumour biology and/or longer waiting times are risks of a worse outcome. The goal of these bridging therapies is to achieve a complete response, defined by the absence of viable tumour on cross-sectional imaging and AFP <100 ng/mL. Patients with a complete response had the best outcomes in a study evaluating the Modified Response Evaluation Criteria in Solid Tumours (mRECIST) classification immediately following the first LRT. 9

| Downstaging to within the Milan criteria
To be eligible for LT and the MELD exception, patients with a tumour burden outside of the Milan criteria must undergo LRT to downstage the tumour burden to within the Milan criteria. Moreover, downstaging may allow time to determine whether the biology of a patient's HCC is favourable for LT. A multicenter study evaluated 187 consecutive patients with HCC and initial tumours outside the Milan criteria (single lesion ≤8 cm, 2-3 lesions each ≤5 cm or 4-5 each ≤3 cm with an aggregate ≤8 cm), who were enrolled in a downstaging protocol. Successful downstaging to within the Milan criteria followed by LT was achieved in 58% of patients. 10 Tumour progression occurred in 32% of patients and 5% had liver-related deaths without LT. On multivariate analysis, a pretreatment AFP >1000 ng/mL (HR 3.3) and Childs Pugh Class B or C (HR 1.6) were associated with treatment failure, suggesting that these patients were unlikely to benefit from downstaging. 10 In a US analysis of 3819 patients with HCC from 2012-2015 in the UNOS database, the 3-year post-LT HCC recurrence was 6.9% in those within the Milan criteria, 12.8% in those downstaged through UNOS downstaging criteria and 16.7% in those downstaged from beyond the UNOS downstaging criteria. 4 Independent predictors of post-LT deaths in the downstaged groups were an AFP ≥100 at the time of LT (HR 2.4) and being in short (<3 months) or medium wait-time (3-9 months) vs long waittime (>9 months) regions (HR 3.1), 4 suggesting that wait-time can help define "low-risk" tumour biology.

| Donor selection in patients with HCC
Timely access to LT can affect LT outcomes in patients with HCC. One study has suggested that to minimize the risk of post-LT recurrence in patients with HCC the optimal waiting time was at least 6 months and no more than 18 months. 13 The use of extended criteria donors is an attractive way to increase access to donors in areas with longer LT   (5) Pre-treatment AFP >1000 ng/mL and/or Childs Pugh Class B or C predict downstaging failure (8) AFP ≥100 ng/mL at LT predicts recurrence; Increased risk of post-LT death in short and medium wait time regions (5) 3-y post-OLT survival 79.1% 3-y recurrence

| Immunosuppression and prophylactic systemic chemotherapy
Patients at high risk of recurrent HCC may benefit from adjustment of their post-LT immunosuppression. Immunosuppression has been shown to be involved in accelerated growth of recurrent post-LT HCC.
The use of higher levels of calcineurin inhibitors (CNI) earlier in the post-LT phase has been associated with an increased risk of HCC recurrence. 16 The mammalian target of rapamycin (mTor) inhibitors, sirolimus and everolimus, has been associated with antineoplastic effects on HCC, 17 and their use has been associated with lower HCC recurrence rates compared to CNI. 18  Thus, early adjustment of immunosuppression limiting the use of CNI and using m-TOR inhibitors could help reduce HCC recurrence rates.
The outcomes of preemptive sorafenib, a multitargeted, orally active tyrosine kinase inhibitor, in LT recipients at high risk of HCC recurrence have been mixed, 20,21 and systemic chemotherapy is not routinely recommended to prevent recurrent HCC, even in high-risk patients.

HCC recurrence occurs despite careful selection of LT recipients
and is because of tumour cell dissemination from circulating cancer cells and micrometastases before or during total hepatectomy. 22 HCC recurrence usually occurs early, a median of 12.3 months after LT. Late recurrence more than 2 years after LT is rare. 23 Recurrent HCC developed in 29 of a series of 132 LT patients for HCC in Italy (15.9%). Fifteen (71%) of these occurred within the first 18 months, and only 2 (7%) more than 2 years later. 23 The RETREAT score was developed with these factors and assigning 0-3 points for increasing AFP levels, 2 points for the presence of MVI, and 0-3 points for the increasing size of the largest viable tumour diameter plus the number of viable tumours for a total score of 8 (Table 2) (Table 3). 24

| Treatment of recurrent HCC after liver transplantation
Overall survival if recurrent HCC is detected after LT is poor. Multifocal recurrence and early recurrence after LT (≤6 months) are associated with worsened survival. 32,33 At present there is no consensus algorithm for the treatment of recurrent HCC. An earlier diagnosis of recurrence could help improve outcomes as several case series have shown improved survival in patients who underwent resection for localized recurrence compared to patients with multifocal recurrence. 23,32,33 Early use of sorafenib and other mTOR inhibitors may also be options as well as resection or LRT, though more data are needed. 34 In a metaanalysis of 1021 LT recipients with HCC recurrence, the median survival after diagnosis of recurrence was 13 months. A variety of treatments were used including surgical resection, systemic therapy with sorafenib, LRT, and the best supportive care. Surgical resection of localized recurrent HCC (42 months) and sorafenib use (18 months) were associated with higher survival rates, although, sorafenib use was associated with increased adverse effects. 35 The best supportive care was associated with the lowest survival rate (3.3 months). 35 If feasible, resection or LRT should be attempted in combination with the use of sorafenib and mTOR inhibitors as any form of therapy appears to have better overall outcomes than supportive care.
The use of checkpoint inhibitor immunotherapy has increased for the treatment of HCC with promising results as second-line therapy for HCC if sorafenib fails. 36 At the time of this writing, 2 checkpoint inhibitors, nivolumab and pembrolizumab, have been approved for the treatment of HCC by the US Food and Drug Administration. Both are anti-PD1 monoclonal antibodies that release the "brakes" of the immune system to control HCC. Although pre-LT treatment data have been promising for HCC, no current data exists to show a benefit for post-LT HCC recurrence. In fact, the use of checkpoint immunotherapy must be carefully considered because of potential uncontrolled immune reactivation. Several case reports of severe acute T-cell-mediated rejection and antibody-mediated rejection have been described in the post-LT setting with graft loss and death. 37 Further efficacy and safety studies are needed before routine use of these medications.

CO N FLI C T O F I NTE R E S T
NT has received institutional grant support from Gilead Sciences and consulting from Intercept. BK and JK have no declared conflicts.