Immune Checkpoint Inhibitors for Recurrent Hepatocellular Carcinoma After Liver Transplantation: Safety Under an Immunosuppression-Preserving Strategy

Introduction

Hepatocellular carcinoma (HCC) is the sixth most common malignancy worldwide [1,2]. At the time of diagnosis, approximately 90% of patients present with liver cirrhosis when HCC is diagnosed and only 20% to 30% of patients with HCC are considered eligible for liver resection [2–5]. Liver transplantation (LT) is another potential curative option for patients with co-existing liver cirrhosis and HCC. As transplant criteria have expanded over time, the proportion of LT recipients who later develop posttransplant HCC recurrence has increased, with reported rates ranging from 10% to 20% [2,6,7]. Treatment options in this setting are extremely limited. While systemic therapies such as tyrosine kinase inhibitors and immunotherapy have reshaped the management of nontransplant HCC, their safety and efficacy among LT recipients remain poorly defined.

Immune checkpoint inhibitors (ICIs) are now widely recommended as first-line therapy for advanced HCC in the general population [8–10]. However, their application in solid-organ transplant recipients is highly controversial because ICIs may trigger allograft rejection through T-cell activation. Reported rejection rates in prior case reports and systematic reviews range from 30% to 40%, with rejection often leading to graft failure and death [11–17]. The absence of standardized guidance on how to adjust immunosuppressive regimens further complicates clinical decision-making, and most published data describe strategies that reduce or minimize immunosuppression that may augment ICI activity but substantially increase the risk of rejection.

Given the lack of evidence, many LT recipients with recurrent HCC are considered unsuitable for ICIs and are left with few alternative systemic options. At our institution, we adopted an immunosuppression-preserving strategy, maintaining or slightly increasing baseline immunosuppressant doses and using low-dose corticosteroids selectively in patients with impaired renal function. Whether this conservative approach can mitigate rejection while still allowing potential antitumor activity has not been established. Therefore, the primary objective of this study was to evaluate the safety of administering ICIs under an immunosuppression-preserving strategy in LT recipients with recurrent HCC. The secondary objective was to provide exploratory, descriptive observations on treatment response and progression-free survival in this highly heterogeneous, clinically challenging population. Because of the small sample size and variable treatment histories, all oncologic findings should be interpreted as hypothesis-generating.

Material and Methods

STUDY DESIGN AND PATIENT SELECTION:

This was a retrospective study of LT recipients who received ICIs to treat HCC recurrence at China Medical University Hospital, Taichung City, Taiwan, from August 1, 2017, to December 31, 2024. The last enrolled patient received treatment for more than 1 month from the deadline. Data were censored on June 30, 2025. Comprehensive examinations before LT revealed that these patients did not have distant metastasis. We collected the following data from patients: sex, age, LT type, date of LT, date of tumor relapse, immunosuppressant type before and after ICI therapy, immunosuppressant type, immunosuppressant dose, immunosuppressant drug level, duration of ICI treatment, tumor response, ICI-related adverse effects, systemic therapy before ICI treatment and combination treatment.

The Research Ethics Committee of the China Medical University Hospital, Taichung City, Taiwan, approved the study design (reference CMUH110-REC1-235). All patients were aware of the potential risks associated with ICIs in patients receiving organ transplants and previously provided written informed consent.

IMMUNOSUPPRESSION STRATEGIES AND TYPES OF ICIS:

Calcineurin inhibitors constitute the standard maintenance immunosuppressant at our center. For patients with HCC, mammalian target of rapamycin (mTOR) inhibitors are routinely incorporated unless contraindicated by significant proteinuria or an estimated glomerular filtration rate (eGFR) below 40 mL/min. During ICI therapy, our institutional strategy is to preserve immunosuppression by maintaining or slightly increasing baseline calcineurin inhibitor or mTOR inhibitor doses, rather than reducing them. In patients with chronic renal insufficiency (eGFR 40–60 mL/min), low-dose corticosteroids (prednisolone 10 mg/day) were added to avoid further escalation of alcineurin inhibitor or mTOR inhibitor doses, while providing additional graft protection.

ICI classification and dosing were based on clinical evidence, patient conditions, and financial considerations. ICIs were administered in 2- or 3-week cycles depending on the agent used. Patients underwent regular laboratory monitoring, including complete blood counts, liver biochemistry, renal function, endocrine profiles, and tumor markers, approximately every 1 to 3 weeks. Radiologic assessments (computed tomography or magnetic resonance imaging) were performed every 2 to 3 months to evaluate treatment response according to modified Response Evaluation Criteria in Solid Tumors (mRECIST) criteria.

STATISTICAL ANALYSIS:

The primary endpoints were safety outcomes, including adverse events leading to ICI discontinuation, immune-related hepatitis, biopsy-proven or clinically suspected graft rejection, graft failure, and death. Hepatitis or rejection was diagnosed based on 1 or more of the following criteria: (a) histologic confirmation on liver biopsy; (b) elevation of alanine aminotransferase (ALT) or aspartate aminotransferase (AST) to more than 3 times the upper limit of normal; or (c) total bilirubin exceeding twice the upper limit of normal. These biochemical thresholds correspond to Grade ≥2 hepatic toxicity according to the Common Terminology Criteria for Adverse Events (CTCAE) version 5.0 [18], and were selected to capture clinically significant events warranting medical attention. Patients with hepatitis attributable to viral infection or to concomitant locoregional therapies, such as transarterial chemoembolization or ablation, were excluded. Other adverse events were also graded according to the CTCAE.

The secondary endpoint was oncologic response among patients with recurrent HCC, evaluated using mRECIST, version 1.1. Because patients initiated ICIs at different points in their clinical course and frequently received heterogeneous subsequent systemic or locoregional treatments, overall survival was not compared; such variability would introduce substantial confounding and limit interpretability. Therefore, progression-free survival (PFS) was selected as the primary time-to-event measure of treatment effectiveness.

Descriptive statistics were summarized as frequencies with percentages or as medians with ranges. Categorical variables, including sex, age, immunosuppressant class, steroid use, and the interval between LT and recurrence, were compared using the chi-square test or Fisher exact test, as appropriate. Continuous variables were analyzed using the Mann-Whitney U test or Kruskal-Wallis test. PFS was estimated using the Kaplan-Meier method. Given the small sample size and limited number of events, multivariable Cox regression models were restricted to a maximum of 1 to 2 covariates to reduce overfitting and ensure model stability.

All analyses were performed using IBM SPSS Statistics, version 19, and statistical significance was defined as a 2-sided P value <0.05.

Results

PATIENT CHARACTERISTICS:

A total of 25 LT recipients who received ICIs for recurrent HCC were included. Baseline characteristics are presented in Tables 1 and 2. Two patients who received LT at another hospital were later referred for systemic treatment; therefore, some data were missing for these 2 patients. Regarding ICI regimens, 6 patients received nivolumab monotherapy, 3 received pembrolizumab monotherapy, 10 received pembrolizumab plus lenvatinib, and 6 received atezolizumab plus bevacizumab (Table 3).

SAFETY IN LT RECIPIENTS:

Three patients (12%) developed grade 2 or higher immune-related adverse events (irAEs), and 2 (8%) had grade 3 or higher events. One patient developed marked elevation of liver enzymes (AST/ALT from 25/22 mg/dL to 81/142 mg/dL) approximately 1.6 months after first ICI therapy, which resolved following adjustment of immunosuppressants and corticosteroid administration. Two patients developed clinically significant gastrointestinal bleeding requiring hospitalization and endoscopic evaluation. One had diffuse gastric and duodenal ulcers with active bleeding, while the other had multiple bleeding esophageal ulcers. These events occurred 1.0 and 2.4 months after ICI initiation, respectively, and both patients discontinued ICIs thereafter. These 3 patients had received nivolumab, pembrolizumab, and pembrolizumab plus lenvatinib, respectively. The incidence of irAEs was higher in patients without steroids than in those receiving concurrent low-dose steroids (3/10 [30%] vs 0/15 [0%], P=0.052). However, irAE risk did not differ significantly according to the number or type of immunosuppressant classes used.

CLINICAL OUTCOMES OF ICIS THERAPY:

Two patients (8%) achieved complete response. One patient developed intra-abdominal metastases 34.5 months after LT and initially showed no response to targeted therapy. Following palliative radiotherapy combined with nivolumab, the patient achieved radiologic disease disappearance and normalization of alpha-fetoprotein (AFP) level, with no recurrence by the end of follow-up (PFS 48.3 months). The second patient developed multiple lung metastases 45.1 months after LT and was treated with atezolizumab plus bevacizumab, achieving complete resolution of lesions and remaining disease-free at last follow-up (PFS 31.7 months). Four additional patients (16%) maintained stable disease for over 12 months. One patient, a 70-year-old woman with lenvatinib failure before, remained on atezolizumab plus bevacizumab without locoregional therapy and continued to exhibit radiologic and biochemical stability for 27.4 months. The disease control rate at 1 year or greater was 24% (6/25).

Median PFS for each regimen was as follows: atezolizumab plus bevacizumab: 4.83 months; pembrolizumab plus lenvatinib: 7.90 months; pembrolizumab monotherapy: 3.83 months; and nivolumab monotherapy: 7.48 months. No statistically significant differences were observed across regimens (Figure 1).

In univariable Cox analysis, recurrence within 1 year after LT (HR 3.47, 95% CI 1.24–9.72, P=0.018) and AFP levels over 100 ng/mL during ICIs therapy (HR 2.56, 95% CI 1.04–6.29, P=0.041) were associated with poorer PFS. Steroid use was associated with a lower incidence of irAEs but demonstrated a non-significant trend toward shorter PFS (HR 2.30, 95% CI 0.93–5.69, P=0.071). Neither the number of immunosuppressants nor everolimus use or its trough concentration showed a significant association with PFS. In the multivariable Cox model, limited to 2 variables to prevent overfitting, recurrence within 1 year after LT remained the only independent predictor of poorer PFS (HR 2.98, 95% CI 1.03–8.61, P=0.043; Table 4, Figure 2).

Discussion

The primary aim of this study was to evaluate whether ICIs can be safely administered to LT recipients when baseline immunosuppression is preserved rather than reduced. Our findings suggest that this immunosuppression-preserving approach may mitigate severe graft-related complications while still allowing some antitumor activity. However, given the small sample size and treatment heterogeneity, all efficacy signals observed in this cohort should be interpreted as exploratory and not comparative.

The use of ICIs in LT recipients remains challenging because immune activation can precipitate acute allograft rejection, a potentially fatal complication in liver grafts. Although ICIs are established first-line therapy for advanced HCC in the general population, their role after LT is limited by the scarcity of safety data and the vulnerability inherent to patients with immunosuppression. Most published evidence consists of case reports or case series, with reported rejection rates of approximately 30% to 40% among organ transplant recipients [15,16]. In this context, the present study provides real-world data regarding safety and clinical outcomes among carefully selected LT recipients treated with ICIs.

Graft rejection and immune-related hepatitis remain the greatest concerns in this setting. Previous systematic reviews have highlighted substantial mortality associated with rejection, especially in LT patients [11–17]. Previous systematic reviews and case reports have highlighted substantial variability in immunosuppressive strategies among solid organ transplant recipients receiving ICIs. In these reports, the approaches ranged widely, from continuing calcineurin inhibitors, switching to mTOR inhibitors, tapering immunosuppression, or even maintaining only low-dose corticosteroids, reflecting highly heterogeneous clinical practices across institutions [19–27]. Importantly, most published studies come from melanoma, which has a far greater intrinsic responsiveness to immunotherapy across the whole cancer spectrum. Therefore, the favorable responses reported in melanoma cannot be directly extrapolated to LT recipients with recurrent HCC, and the actual antitumor effectiveness of ICIs in this population remains uncertain. Given these limitations, no standardized or optimal immunosuppression-adjustment strategy can be derived from the existing literature. Our institutional practice follows a conservative, immunosuppression-preserving strategy, maintaining or slightly increasing baseline calcineurin or mTOR inhibitor doses and adding low-dose corticosteroids when needed. Under this approach, we observed relatively low rates of rejection-related hepatitis or severe irAEs, compared with rates of prior pooled analyses. Notably, steroid use was associated with fewer irAEs but also showed a non-significant trend toward shorter PFS, suggesting a potential trade-off between graft protection and antitumor activity, and an observation that warrants further investigation.

Regarding oncologic outcomes, 2 patients achieved complete remission, and 4 maintained stable disease for more than 1 year. While encouraging, these responses should be interpreted cautiously. Many patients received tyrosine kinase inhibitors in combination with ICIs, and 4 ICI regimens were used during the study period. This therapeutic heterogeneity makes it challenging to isolate the direct effect of ICIs. Our findings that late recurrence (>1 year after LT) and lower AFP levels were associated with better PFS align with the findings of previous literature [28–30], suggesting that tumor biology strongly influences post-transplant HCC behavior. However, these associations remain hypothesis-generating due to the small sample size and limited event numbers.

The role of immunosuppressive regimens, particularly mTOR inhibitors, remains of significant interest. Some studies have suggested potential synergistic interactions between mTOR inhibitors and ICIs or protective effects against rejection [20,31,32]. In 1 case report, Esfahani et al reported that the number of IFN-γ+ CD4+ T cells and the serum IFN-γ level increased with the addition of sirolimus, likely promoting ongoing anti-PD-1 efficacy [33]. They provide insights into a potential synergistic response between these 2 agents. Some articles suggest that mTOR inhibitors will be a better option than other immunosuppressants when ICI therapy is administered [20,34]. However, we did not observe clear associations between mTOR inhibitor use or trough levels and clinical outcomes. This may be due to the heterogeneous use of immunosuppressants, limited sample size, and the retrospective nature of our study. Nonetheless, our findings emphasize the feasibility of administering ICIs without reducing baseline immunosuppression, potentially supporting prolonged treatment exposure and mitigating severe rejection.

The role of immunosuppressive regimens, particularly mTOR inhibitors, continues to attract interest. Experimental and clinical observations suggest potential synergistic interactions between mTOR inhibition and PD-1 blockade, or protective effects against rejection [20,31–34]. Our findings support the feasibility of administering ICIs without reducing baseline immunosuppression, a strategy that may enable prolonged treatment exposure while mitigating severe rejection.

Taken together, these findings suggest that ICIs can be considered a salvage option for carefully selected LT recipients with recurrent HCC who have preserved performance status and limited systemic treatment alternatives. Preserving immunosuppression may reduce severe immune-related toxicity, although it may also attenuate antitumor activity. Therefore, treatment decisions should be individualized, ideally through multidisciplinary discussion, with explicit counseling regarding the risk of graft rejection.

This study has several limitations. First, the small sample size limits statistical power, precludes robust multivariable analyses, and restricts interpretation of efficacy outcomes. Second, substantial heterogeneity in systemic introduces confounding that cannot be fully adjusted in this retrospective design. Finally, it should be acknowledged that ICIs are generally not recommended for solid organ transplant recipients in most international guidelines because of the high risk of allograft rejection. However, in our institution, their use is considered in carefully selected cases after multidisciplinary discussion, particularly for malignancies known to be responsive to immunotherapy or in cases lacking other feasible systemic options. These decisions reflect individualized, last-resort clinical judgment rather than routine practice. This institutional context should be considered when interpreting the applicability of our findings to other centers. Nonetheless, our study provides meaningful real-world evidence supporting the cautious use of ICIs under an immunosuppression-preserving strategy in LT recipients with recurrent HCC and highlights the need for larger prospective studies to establish optimal management practices.

Conclusions

Our findings suggest that ICIs can be administered under an immunosuppression-preserving strategy without clearly increasing the risk of severe graft-related complications. However, the potential interaction between steroid use, graft protection, and antitumor activity remains uncertain. In our institution, ICIs are considered only in carefully selected cases after multidisciplinary discussion, typically when patients have ICI-responsive malignancies or no effective systemic alternatives. Given the small sample size and retrospective design, these results should be interpreted cautiously, and larger multicenter studies are needed to clarify the safety and appropriate clinical context for the use of ICIs.