After liver transplantation, any recurrence is, by definition, metastasis from the native liver. The culprit is either the presence of undiagnosed distant metastasis before transplantation or spillage of tumour cells during transplantation. Even an isolated recurrence implicates solitary metastasis and represents a local phenomenon of the systemic event, which highlights the importance of systemic therapy and the input of oncology as a critical component of the therapeutic strategy.

Immunity is the primary defence against cancer[7]. The adaptive immune system recognizes and eliminates tumour cells based on their expression of tumour-specific antigens[8]. Termed concomitant immunity, the immune response induced by the primary tumour inhibits the growth of secondaries[9]. However, after liver transplantation, concomitant immunity is suppressed pharmacologically. Any microscopic tumour in vitro can progress without immune surveillance. It was observed that post-transplant HCC recurrence progresses significantly faster than in patients treated with hepatic resection[10].

Calcineurin inhibitors, e.g., tacrolimus and cyclosporine, form the cornerstones of maintenance immunosuppression in liver transplantation. In addition to host immune suppression, they promote tumour progression via non-immune-mediated pathways related to augmented transforming growth factor expression[11,12]. From a retrospective series of 70 HCC patients treated with liver transplantation, quantified cyclosporine exposure was identified as an independent risk factor for HCC recurrence[13]. Subsequently, Vivarelli et al[14] confirmed that high tacrolimus exposure independently predicted HCC recurrence. Immunosuppressive therapy affects the course of tumours in transplant patients and must be fully addressed in the comprehensive management of a recurrence.

Throughout the course of treatment, the liver graft must be maintained. Reduction of immunosuppression increases the risk of graft rejection. Medical therapies potentially affect liver function. The use of immunotherapy is particularly concerned with immune-mediated graft injury. While formulating the treatment strategy, the benefits of the treatment must be balanced with the potential toxicities towards the liver graft.

Mammalian target of rapamycin (mTOR) is a protein involved in a signalling pathway that controls cellular growth and proliferation[15]. Rapamycin, more commonly known as sirolimus, inhibits the mTOR pathway to restrain regulatory T-cell proliferation[16]. Apart from immune modulation, mTOR is also involved in HCC pathogenesis and is associated with poor tumour biology[17-19]. Sirolimus has been investigated in a phase II trial showing promising efficacy against advanced HCC[20]. With the theoretical advantage over tumour control, sirolimus has been extensively investigated as immunosuppression therapy for patients engrafted for HCC[21-27] (Table 1).

The highest level of evidence came from a prospective trial conducted by Geissler et al[21], where 525 patients were randomized to receive either a sirolimus-based or an mTOR inhibitor-free regimen. In the study group, sirolimus was incorporated 4-6 wk after transplantation, with or without a concomitant calcineurin inhibitor. The overall and recurrence free survival rates were improved up to 5 years (overall survival: 79.4% vs 70.3%; P = 0.048) and 3 years (disease-free survival: 80.6% vs 72.3%; P = 0.0499). The proportion of patients with acute rejection appeared to be higher in the sirolimus group (23.4% vs 17.0%, respectively; P = 0.07), but the difference did not reach statistical significance. The results were in concordance with an updated meta-analysis that demonstrated a survival benefit in patients receiving sirolimus-based immunosuppression therapy (OR = 1.68; CI = 1.21-2.33)[23]. From the pooled results of 11 studies, the risk of graft rejection or hepatic artery thrombosis was not increased. Sirolimus was generally well tolerated. In a small proportion of patients (0-8.3%), sirolimus was discontinued for drug toxicity, mostly due to oral ulcers[28].

Everolimus is a derivative of sirolimus with a shorter elimination half-life (30 h vs 60 h) and a quicker time to steady state (4 d vs 6 d)[29,30]. The clinical advantage is easier dose adjustment. Everolimus received evaluation in a phase III trial for its role in advanced primary HCC that progressed despite sorafenib therapy[31]. However, no further survival benefit was observed upon switching to everolimus (overall survival: 7.6 mo vs 7.3 mo). Everolimus has been evaluated in prospective trials for its efficacy in liver transplantation, although they were not focused on oncological outcomes. In a prospective multicentre study, everolimus with a reduced dose of tacrolimus was associated with better preserved renal function (estimated glomerular filtration rate decline over 36 mo: 7.0 mL/min/1.73 m²vs 15.5 mL/min/1.73 m²; P = 0.005) compared with the standard dose of tacrolimus[32]. A similar regimen was studied in another prospective trial with a composite primary endpoint comprising rejection and graft loss[33]. Notably, in patients transplanted for HCC, recurrence was only observed in the control arm with a standard dose of tacrolimus (5/62 vs 0/62) after 12 mo of follow up. A direct comparison between everolimus and sirolimus was made in a meta-analysis[34]. Patients on everolimus had significantly fewer recurrences than those on sirolimus or calcineurin inhibitors (4.1% vs 10.5% vs 13.8%, respectively; P < 0.05). However, everolimus-treated recipients had a shorter follow-up time (13 mo vs 30 mo vs 43.2 mo, respectively) and fewer advanced tumours (HCC within Milan criteria: 84% vs 60.5% vs 74%, respectively; P < 0.05). The study did not compare survival, and no definite conclusions were drawn.

The data on mTOR inhibitor therapy for established recurrence after liver transplantation remain scarce. However, a combination of either sirolimus or everolimus with reduced-dose tacrolimus has been proven to be safe and effective in reducing recurrence[24,28,33]. There is inadequate evidence to recommend the optimal serum level of tacrolimus in this combination. In our experience, a sub-therapeutic level of tacrolimus might suffix. From Geissler’s prospective trial it appears that Siroliums monotherapy might be adequate for some patients[21]. From a registry database comprising 2491 patients transplanted for HCC, sirolimus was the only maintenance immunosuppressant affecting survival (5-year survival: 83.1% vs 68.7%, P < 0.05)[27]. Based on these findings, it appears sensible to incorporate an mTOR inhibitor with a reduced calcineurin inhibitor upon the diagnosis of recurrence. Overall, immunosuppression should be individualized and tapered to spare the remaining anti-tumour immunity. Patients should be closely monitored for liver function throughout the course of cancer treatment.

Sorafenib is a multi-tyrosine kinase inhibitor with activity against vascular endothelial growth factor-2 and -3, platelet-derived growth factor receptor and Ras ligand[40]. It inhibits tumour signalling and angiogenesis pathways involved in HCC pathogenesis. In a randomised controlled trial, sorafenib was shown to prolong the median survival of patients with advanced HCC for 3 mo (10.7 mo vs 7.9 mo, P < 0.001)[41]. The major drawback was a poorly tolerated side effects profile. Hand-foot skin reaction and gastrointestinal disturbances were reported in 21% and 39% of the patients, respectively. Although mostly graded as 1 and 2 in severity, drug-related adverse events have led to discontinuation of sorafenib in 29% of the patients.

The efficacy of sorafenib in post-transplant HCC recurrence has been studied in numerous retrospective series, mostly in combination with an mTOR inhibitor (Table 2)[42-52]. Sorafenib and mTOR inhibition had synergistic effects on tumour growth in xenograft mice[53]. Ras blockade silenced the feedback signalling of mTOR inhibition, leading to upregulation of its anti-tumour activity[54]. A retrospective cohort reported by De’Angelis et al[45] provided insights into the use of sorafenib in patients with advanced recurrence. The outcomes of 15 patients treated with sorafenib were compared with those of 24 patients receiving best supportive care. Sorafenib was stared at 400 mg twice daily. More patients in the sorafenib group received an mTOR inhibitor due to the time effect, but the difference did not reach statistical significance (46.7% vs 16.7%, P = 0.13). Sorafenib conferred disease control (partial response or stable disease) in 11 of the 15 patients (73.4%), translating into a survival benefit (median OS: 41.4 mo vs 19.1 mo; P = 0.013). Notably, there was a high proportion of patients requiring dose reduction (53.3%) or discontinuation of treatment (13.3%) due to drug toxicity.

Gomez-Martin et al[50] addressed the safety of combining sorafenib with an mTOR inhibitor in a post-transplant setting. In the multicentre cohort consisting of 31 patients with recurrent HCC, the immunosuppression was shifted to mTOR inhibitor therapy with initiation of sorafenib as systemic treatment. Most toxicities were grade 1 or 2. However, 2 episodes of gastric bleeding and 1 episode of cerebral haemorrhage were reported. The gastric bleedings were diffuse mucosal oozing unrelated to portal hypertension or ulcer disease. Thus, sorafenib appears to be effective to prolong survival after recurrence but at the cost of significant toxicity. Combination treatment with an mTOR inhibitor should be avoided in patients with potential bleeding complications.

Immunotherapy directs the host immunity towards the tumour[55]. The physiological immune response is regulated by immune checkpoints[56]. Immunotherapy consists of antibodies directed against these immune checkpoints on the T-cell surface to prompt reactions against tumour antigens. Examples include ipilimumab that targets cytotoxic T-lymphocyte-associated antigen 4 (CTLA-4) and nivolumab and pembrolizumab that target programmed cell death protein 1 (PD-1). Nivolumab has been validated in a large phase II trial for its safety and efficacy against primary HCC[57]. Nivolumab 3 mg/kg was given every 2 wk to 214 patients with advanced HCC. Disease control was achieved in 64% of all patients, including 61% of patients who had previously failed sorafenib treatment. The overall survival was 83% at 6 mo. There was a favourable side effect profile compared with that of sorafenib. Only 2%-4% of patients discontinued nivolumab due to drug toxicity.

However, immune checkpoint modulation of cell-mediated immunity is implicated in transplant organ tolerance[58,59]. Downregulation of these pathways may inadvertently lead to transplant rejection[60]. In fact, clinical trials for immune checkpoint inhibitors often exclude solid organ transplant recipients due to the fear of graft injury[61,62]. Current experience in immunotherapy after liver transplantation is confined to case reports and small series[63-66] (Table 3). Limited survival (0.3 mo to 3 mo) was observed among the 10 patients treated with anti-PD-1. The salvage nature of immunotherapy must be considered while interpreting the results. Most patients had developed disease progression with sorafenib. Moreover, the clinical decision to employ immunotherapy for transplant patients is usually much delayed until treatment failure is evident. Although the therapeutic effect is considered rapid for immunotherapy, a 3-m interval is usually necessary before the treatment response can be evaluated⁵⁷. In the reports, the rather limited survival interval after immunotherapy might not allow the efficacy of immunotherapy to be assessed.

Acute rejection occurred in 3 of the 10 reported cases receiving anti-PD-1 treatment. Although a limited number of events precludes risk factor analysis, a hypothesis could be proposed. Two patients with rejection were relatively young, aged 14 and 20 years, respectively. A young age is a recognized risk factor for acute rejection after liver transplantation, and more aggressive immunosuppression is usually employed[67]. A long duration after transplantation is usually protective of acute rejection. However, the trend is not obvious from this series of observations. Practically, most recurrence occurs early after transplantation as well.

The differential effect of PD-1 and CTLA-4 blockade on rejection may also have implications[68]. Among the 5 reported liver transplant patients treated with immunotherapy for melanoma, rejection was observed only in patients receiving a PD-1 inhibitor[66,69-71] (Table 3). These clinical observations concurred with the findings from in vitro studies. Using a murine model it was demonstrated that the PD-1 pathway may play a stronger role in allograft tolerance than CTLA-4 and that PD-1 blockade could be associated with a higher risk of transplant rejection[72]. However, the effect of CTLA-4 blockade on HCC control has not been systematically investigated. The role of immunotherapy in treating HCC recurrence after liver transplantation remains largely unknown. The potential efficacy should not be overlooked but has to be balanced with its safety[73]. Further study in a large patient cohort is warranted to elucidate optimal patient selection.

The results for surgical resection have been retrospectively reported for patients with intra-hepatic or extrahepatic oligo-recurrence (Table 4)[4,39,76-79]. Patients eligible for surgical treatment ranged from 25% to 50%. The lung and liver were common sites for resection (Table 4). Survival benefits have been consistently demonstrated in patients treated with surgery, with a median survival of 28 mo to 65 mo observed for patients receiving surgery, compared with 5 mo to 15 mo in those receiving systemic treatment only[4,39,76-79].

Selection bias was inevitable because surgical candidates were invariably patients with localized disease and a better prognosis. In the most recent series, patient selection was further refined with an additional criterion being the absence of progression while on systemic treatment[79]. The genuine benefit conveyed by surgery could be questioned because the selected patients had a limited disease burden and more favourable tumour biology. However, a prospective randomized trial is unlikely under the current setting to be given ethical concern. A matched retrospective comparison is also difficult due to the intrinsic differences between the patients with oligo- and disseminated recurrence.

Reviewing the current literature, long-term survival after post-transplant recurrence has been achieved with surgical resection. Across numerous reported series, surgical treatment remained an independent predictor of superior survival after recurrence[4,76,77,79]. Surgery is supported as the treatment of choice in patients with resectable recurrence, especially when the tumour biology is favourable.

This review focuses on genuine intra-hepatic recurrence. In the setting of primary HCC, surgical resection with partial hepatectomy confers favourable oncological outcomes[80]. Surgical resection is therefore given full consideration here when macroscopic disease is confined within the liver. Because recurrence usually occurs early, graft function is preserved and rarely precludes hepatectomy. Resectability is more determined by disease burden. A tumour-free future remnant with adequate volume is the prerequisite for graft hepatectomy. The main concern for graft resection is morbidity. Graft resection is technically challenging due to extensive hilar adhesions. Immunosuppressed patients are susceptible to infective complications. Graft resection for recurrence has been reported in several retrospective series in small numbers (Table 5)[81-83]. High morbidity (60%-80%), but no mortality, was reported. In the first series, infective complications occurred in 63% patients, including 5 Gram-negative bacteraemia requiring intravenous antibiotics[83]. Two series reported oncological outcomes[82,84], with 3-year overall survival rates ranging from 50% to 70%. Graft resection appears to be a feasible treatment for recurrent HCC, offering the chance of long-term survival. However, high morbidity, especially infection complications, is expected.

Due to the significant morbidity associated with graft resection, the role of ablative treatments has been investigated. Radiofrequency ablation (RFA) is a thermal ablative technique delivered via a needle electrode. For primary HCC, RFA offers equivalent local control as resection for small tumours less than 3 cm in size[85]. The advantage is that it can be performed percutaneously under radiological guidance. A favourable location would be away from the major vasculature and adjacent organs. RFA becomes more appealing when a small tumour is situated in the deep parenchyma, where major hepatectomy is necessitated for resection. In the setting of post-transplant recurrence, the additional benefit is the avoidance of morbidities associated with re-laparotomy in an immunocompromised patient.

One retrospective cohort compared RFA with resection for post-transplant HCC recurrence[84]. In the reported series, 15 patients were treated with surgery while 11 received RFA. The author reported similar 3-year (51% vs 51%, P = 0.88) and 5-year (35% vs 28%, P = 0.88) overall survivals in two groups. However, both hepatic and extra-hepatic recurrences were included, and the results represented the outcomes of heterogeneous procedures. Morbidity and mortality after graft resection were not reported.

Stereotactic body radiation therapy (SBRT) is a precise method of delivering ablative radiation by tomographic modulation. Intense and focused doses of radiation are given in a few or single fractions. SBRT for post-transplant HCC recurrence has several theoretical advantages. The radiation beam is focused on the tumour, sparing the adjacent normal liver parenchyma. A higher dose of radiation is delivered while the risk of collateral damage is minimized[86]. Moreover, SBRT is delivered over fewer treatment days than the 10-20 d for conventional radiotherapy, during which systemic therapy is usually deferred. SBRT is usually completed in 1-5 fractions, allowing systemic treatment to be commenced early.

Moreover, it is now established in pre-clinical models that stereotactic radiation upregulates anti-tumour immunity[87-89]. High-dose radiation stimulate antigen-presenting cells, leading to the activation and proliferation of tumour-specific cytotoxic T cells[89]. The abscopal response (from “ab scopus”, meaning away from target) denotes this systemic effect leading to the regression of metastatic lesions outside the irradiation field[90]. Interestingly, abscopal effect is synergistically enhanced when combined with immunotherapy-mediated PD-1 blockade[87], which potentially confers a further clinical advantage to SBRT for recurrent HCC because the role of systemic therapy is crucial.

SBRT has been investigated in several prospective studies for primary HCC[91-94]. In these series, the tumour size ranged from 2 cm to 7 cm. At 2 years after ablation, local control was achieved in 80% to 95% of patients. The figure compares favourably with that reported for RFA of small tumours[95,96]. In contrast to RFA, vascular invasion is not a contraindication[97,98]. In direct retrospective comparison, local control was found to be superior in the SBRT group for tumours more than 2 cm in size (HR: 3.35; P = 0.025). Grade III or above morbidity was similar (SBRT vs RFA: 5% vs 11%; P = 0.31). While RFA loses efficacy with increasing tumour size[99,100], SBRT seems to be as effective when treating larger tumours. To date, the role of SBRT for post-transplant HCC recurrence has yet to be systematically evaluated. While systemic control is of utmost importance, the potential of the SBRT and immunotherapy combination should be conscientiously explored.

In patients with multifocal intra-hepatic recurrence, trans-arterial chemoembolization (TACE) offers the opportunity of regional control. Ko et al[101] first reported the results of TACE for recurrent HCC after liver transplantation with 1- and 3-year survival rates of 47.9% and 6.0%, respectively. However, in their series, 64.3% of patients developed concomitant extra-hepatic metastasis, which could have affected the oncological outcome as well. Zhou et al[102] prospectively compared TACE versus systemic therapy in patients with unresectable intra-hepatic recurrence. Survival benefits were achieved in the TACE arm (P = 0.013), indicating that regional control could have contributed to the improvement in overall survival. Notably, both studies reported no major morbidity after graft liver TACE. In Zhou et al[102]’s series of 14 patients, no biliary complications were observed over a median follow up of 14.5 mo.

Intra-arterial irradiation with Yttrium-90 (Y-90) microspheres has gained popularity in recent years to treat unresectable HCC. Injected through the feeding vessels, these microspheres emit high-dose radiation after entrapment at the pre-capillary level. In a large-scale longitudinal cohort comprising 291 patients, Y-90 achieved a 40%-60% response rate[103]. The median survival was 17.2 mo in patients with Child’s A cirrhosis. In contrast to TACE, portal vein thrombosis is not a contraindication. Considering the potential synergistic effect of irradiation and immunotherapy, clinical studies are ongoing to investigate the benefit of combining Y-90 and anti-PD1 therapy for primary HCC. Their results will shed light on further applications concerning post-transplant HCC recurrence.