Immune‑mediated liver injury caused by checkpoint inhibitors: A case series and review of the literature

Introduction

Immune checkpoint inhibitors (ICIs) function by blocking the interaction between immune checkpoints and their corresponding partner proteins. This action hinders the transmission of the inhibitory signal enabling T cells to effectively eliminate cancer cells. Immune-related adverse events (irAEs) resulting from ICIs occur due to disrupted self-tolerance caused by the absence of T-cell suppression (1). Amongst the most common organs affected by ICIs are the endocrine glands, the gastrointestinal tract, the liver, the lungs and the skin. Hepatitis represents the predominant manifestation of hepatotoxicity linked to ICIs. This condition is characterized by hepatocellular damage, as indicated by increases in serum aminotransferases (ALT and AST), occasionally accompanied by rises in serum bilirubin levels (2).

‘Immune-mediated liver injury caused by checkpoint inhibitors (ILICI)’ is the preferred terminology characterizing ICI-induced liver injury which represents a different entity regarding the incidence, mechanism of action, and therapeutic approach (3,4). Typically, it manifests within the first months of treatment initiation, although it may appear as soon as after the first dose or even years later. It is also known that several irAEs, present following completion of treatment (1,2). Usually, patients are asymptomatic and ILICI is found incidentally in routine liver function tests (LFTs). Establishing the diagnosis is a challenging process, as it is made by excluding other possible causes of hepatic disfunction, such as infectious, metabolic, vascular, autoimmune, tumor-related or other drug-induced liver injury. Liver biopsy is not routinely recommended and based on ESMO guidelines for management of toxicities from immunotherapy should be considered for grades 3 and 4 (5,6).

Management of ILICI is dependent on the extent of liver injury, which is determined by evaluating the AST and ALT levels. Immunotherapy may be continued, temporarily withheld with subsequent resumption, rechallenged under an individualized treatment plan or permanently discontinued. Corticosteroid administration is the gold standard treatment, followed by immunosuppressive agents, such as mycophenolate mofetil (MMF), azathioprine, cyclosporine etc. in refractory cases. This report aims to raise awareness of ILICI in patients receiving immunotherapy, underscoring its unpredictable onset and the importance of timely recognition and appropriate management in routine clinical practice.

Case series

Study population

We retrospectively analyzed all consecutive patients treated at Attikon University Hospital, Chaidari, Greece, between 2020 and 2024 who developed ILICI, comprising a total of five cases. The median patient age was 69 years (range 50-85 years) and all patients were male. All cases had metastatic malignancies of different primary origins. Subjects enrolled in this study received at least one anti-PD1 agent, and two had additionally received an anti-CTLA4 agent, following the recommendations outlined in the National Comprehensive Cancer Network (NCCN) and ESMO guidelines. All presented with fatigue, jaundice, and abnormal liver function tests (LFTs). Liver function was evaluated by monitoring of i) LFTs including alanine transaminase (ALT) and aspartate transaminase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), serum bilirubin, the international normalized ratio (INR), total protein and albumin and ii) imaging (computed tomography-CT). The adverse effects' severity was assessed using the grading system from version 5.0 of the Common Terminology Criteria for Adverse Events (CTCAE) (7). The diagnosis was confirmed by liver biopsy in three cases, while in two patients, immune-related hepatitis was diagnosis by exclusion. Patients received between one and forty doses of immunotherapy before developing immune-related adverse events (irAEs). In all cases, glucocorticoids were used as first line treatment; mycophenolate mofetil was used as a second line treatment in one case. Table I displays clinical information of these patients, along with their demographics, treatment, adverse event grading, management, and outcomes. The authors actively participated in the clinical care of the patients and collected data through chart reviews.

Table I Baseline characteristics of our five patients.

Table I

Baseline characteristics of our five patients.

Patient no.	Sex	Age, years	Cancer type	Immunotherapy	Doses received	CTCAE grading	Liver biopsy	Treatment	Current status	Disease resolution
1	M	85	Lung cancer	Pembrolizumab	1	3	Yes	mPSL, MMF	Deceased	No
2	M	50	Oropharyngeal cancer	Nivolumab	2	3	Yes	mPSL	Deceased	No
3	M	81	Cholangiocarcinoma	Pembrolizumab	40	2	Yes	mPSL	Deceased	Yes (5 months)
4	M	80	Renal cell carcinoma	Nivolumab/ipilimumab	34/4	3	No	mPSL	Deceased	Yes (1 month)
5	M	51	Renal cell carcinoma	Nivolumab/ipilimumab	3/3	4	No	mPSL, PSL	Alive	Yes (1 month)

[i] CTCAE, Common Terminology Criteria for Adverse Events; M, male; MMF, mycophenolate mofetil; mPSL, methylprednisolone; PSL, prednisolone.

Patient characteristics and outcomes

A total of five male patients were included in this case series. Underlying malignancies were lung cancer (n=1), oropharyngeal cancer (n=1), cholangiocarcinoma (n=1), and renal cell carcinoma (n=2). Immunotherapy regimens consisted of pembrolizumab (n=2), nivolumab (n=1), or the combination nivolumab/ipilimumab (n=2). The median interval between immunotherapy initiation and ILICI was 15.5 months (range 1-30 months). According to CTCAE grading system three of the patients exhibited grade 3 hepatotoxicity, and the other two grade 2 and grade 4 respectively and the pattern was both hepatocellular and cholestatic. Patients exhibited symptoms after the first dose of immunotherapy (n=1), after two to three doses (n=2), and after thirty-four and forty doses of ICIs (n=2). Liver biopsy was performed in three cases. All biopsies revealed lobular inflammatory infiltrate, consisting mainly of CD3+, CD8+ T-lymphocytes and abundant CD68+ histiocytes, and centrilobular cholestasis along with periportal fibrosis and interface hepatitis, which is characterized by the extension of the mononuclear inflammatory infiltrate to hepatocytes located at the limiting plate resulting in injury or necrosis (‘piecemeal necrosis’) (Table II; Fig. 1). These pathological findings could be attributed to ILICI based on current evidence (8,9). All patients were treated with corticosteroids, and only one required additionally the administration of mycophenolate mofetil. With regards to the patients' outcomes, four of our five patients died after the onset of ILICI: Two deaths were attributed to the immune-mediated liver injury, whereas the remaining two were related to complications arising during hospitalization and not directly caused by ILICI. Only one patient remained alive at follow-up. Disease resolution was achieved in three cases (60%), with an interval between the initiation of treatment and recovery that ranged from 1 to 5 months. Two patients (40%) showed no evidence of disease resolution.

Figure 1 Representative images of the main morphological and immunohistochemical features of our cases, derived from different patients. (Aa) Portal tract and (Ba) hepatic lobule hematoxylin and eosin staining (magnification, x200) reveals mild to moderate hepatic injury. (Ab) Portal tract and (Bb) hepatic lobule CD3 immunostaining (magnification, x200), and (Ac) portal tract and (Bc) hepatic lobule CD8 immunostaining (magnification, x200). The inflammatory infiltrate consists mainly of CD3+, CD8+ T-lymphocytes and accumulates in the portal tracts with occasional extension to the hepatic lobules (interface hepatitis). In the portal tracts the T-lymphocytes invade the bile ductule epithelium (arrowheads in insets; magnification, x400). (Ad) Portal tract and (Bd) hepatic lobule CD68 immunostaining (magnification, x200). CD68+ histiocytes are increased predominately around the portal tracts and within the hepatic sinusoids.(Ca) Masson trichrome staining (magnification, x100) showing mild portal-periportal fibrosis (Cb) Masson trichrome staining (magnification, x100) showing occasional bridging fibrosis between portal tracts and branches of the central vein (arrowheads).

Table II Ishak modified histological activity index and histological findings in our cases.

Table II

Ishak modified histological activity index and histological findings in our cases.

Patient	Immunotherapy	Ishak score	Histological findings
Case 1	Anti-PD1	Grade: 4/18; stage: 1/6	• Moderate periportal and mild lobular inflammatory infiltrate consisting mainly of CD3 (+), CD8 (+) T-lymphocytes and abundant CD68 (+) histiocytes.
			• Interface hepatitis.
			• Cholestasis.
			• Few foci of necrotic debris in the hepatic lobules.
			• Lymphocytic invasion of the portal tracts.
			• Mild portal/periportal fibrosis.
			• Rare PD(L)-1 expression.
Case 2	Anti-PD1	Grade: 6/18; stage: 1/6	• Mild periportal and moderate lobular inflammatory infiltrate consisting mainly of CD3 (+), CD8 (+) T-lymphocytes and CD68 (+) histiocytes, especially in the hepatic lobules.
			• Interface hepatitis.
			• Cholestasis.
			• Numerous foci of necrotic debris in the hepatic lobules.
			• Mild periportal fibrosis.
			• Rare PD(L)-1 expression.
Case 3	Anti-PD1	Grade: 4/18; stage: 1/6	• Moderate periportal and mild lobular inflammatory infiltrate consisting mainly of CD3 (+), CD8 (+)
			• T-lymphocytes and abundant CD68 (+) histiocytes.
			• Interface hepatitis.
			• Cholestasis.
			• Mild periportal, focally ‘bridging’ fibrosis.
			• Rare PD(L)-1 expression.

Literature review

Additionally, we conducted a review of published case reports regarding ILICI. A comprehensive search strategy was implemented for case reports published from 2016 until 2024 in PubMed database (https://pubmed.ncbi.nlm.nih.gov/) following the terms: [checkpoint inhibitors] AND [hepatitis]. After reviewing the titles and the abstracts of all 185 publications, we excluded 91 articles due to irrelevance. Apart from the PubMed search, four more cases were identified from the broader literature, including reference lists of relevant articles. We reviewed from patients' records demographic data and disease characteristics such as gender, age, primary tumor type, type and dosage of ICIs, toxicity grading, concomitant liver disease and immune related adverse events, performance of liver biopsy, treatment received and disease outcome. Case reports and case series were included if they described patients with an unambiguous diagnosis of ILICI and provided sufficient clinical detail regarding treatment exposure, liver toxicity, management, and outcomes. Articles providing inadequate clinical information or not clearly attributing hepatotoxicity to ICIs were excluded.

Discussion

ICIs are monoclonal antibodies that target immune checkpoints, such as PD-1/PD-L1 and CTLA-4, that block. This leads to T-cell activation against cancer cells and an imbalance within the immune system, that often results in immune-related adverse events (irAEs). There are three categories of ICI approved for treatment of different types of cancer: PD-1 inhibitors, PD-L1 inhibitors, and CTLA-4(10). Although several predictive biomarkers, such as PD-L1 expression, tumor mutation burden, microsatellite instability, microbial diversity, hypoxia, interferon-γ and extracellular matrix, are utilized to predict response to immunotherapy, the data regarding biomarkers for irAEs are scarce (10,11).

ILICI is a severe irAE that cannot be predicted nor confirmed, as there are not well-established specific biomarkers for the prediction or diagnosis of immune associated liver injury (12,13). Atallah et al (14) describe four risk factors associated with higher risk of liver damage: dual inhibition, female sex, low baseline ALP and high baseline ALT. Higher rates of liver damage are observed in individuals receiving anti-CTLA-4 agents combined with an anti-PD-1/PD-L1; indeed, up to 23% of patients receiving combined anti-PD-1/anti-CTLA-4 treatment have grade 3 LFT abnormalities (12,14-16). It is also observed that the coexistence of liver metastases is not related to the incidence or ILICI severity (14,16). At the time of ICI initiation, all patients had stage IV metastatic disease. Given the small sample size and heterogeneity of prior systemic treatments in our cohort, definitive conclusions cannot be drawn. However, published data suggest that prior treatment exposure, including history of ICI therapy and targeted therapies, may potentially influence the risk of ILICI, whereas its impact on prognosis remains unclear and appears to depend primarily on the severity and clinical course of hepatic involvement (17-19).

Data on the incidence of multiple irAEs is limited and originates from heterogeneous populations and retrospective studies. These studies support that the frequency of irAEs affecting multiple sites ranges from 13 to 53%, with higher prevalence amongst patients treated with combination therapy (16,20). It has also been noted that the occurrence of these irAEs is associated with response to treatment and a favorable prognosis (21-23). ILICI occurs in 5-10% of patients receiving monotherapy and in 25-30% of patients treated with combination therapy (6). In addition to our cohort, we identified 110 cancer patients from the literature who developed ILICI. The demographic and clinical characteristics of these patients, together with treatment details and outcomes, are summarized in the supplementary material (Table SI). Typically, it manifests within the first months of treatment initiation, although it may appear as soon as after the first dose or even years later. It is known that several irAEs, present following completion of treatment (1,2). The onset of ILICI in our patients was after 1 to 40 cycles of immunotherapy treatment, in accordance with the published literature, in which 25.4% of patients developed liver damage after the administration of a single dose of ICIs; nevertheless, there were also cases in which ILICI appeared after 28 to 51 doses. Furthermore, assessment of concomitant liver disease status, including hepatic metastases, revealed no correlation with the occurrence of hepatotoxicity, as 82 patients (74.5%) had no prior liver disease. Additionally, nearly half of the patients (48.2%) experienced at least one more immune-related adverse event.

Liver damage is characterized by the elevation of serum aminotransferases and/or total bilirubin, the severity of which is assessed by grading the elevation of the values according to CTCAEv5.0 guidelines (6,7). This adverse event usually occurs as an incidental finding, but in more severe cases, it manifests with jaundice, pain, fever and bruising, while encephalopathy and coagulation disorders indicating acute liver failure are rare at this point. The liver enzyme pattern is predominantly hepatocellular, with elevation of serum aminotransferases (AST, ALT). A cholestatic or mixed pattern of liver function test abnormalities occurs at rare occasions and elevated total bilirubin designates extended liver damage (5,9,24). The lack of diagnostic biomarkers makes ILICI a diagnosis by exclusion. This requires the exclusion of other possible causes of liver damage, such as viral or systemic infections, hepatic diseases (NAFLD, autoimmune hepatitis, alcoholic hepatitis, hepatic metastases), genetic liver diseases (Wilson's disease, hemochromatosis, alpha-1 antitrypsin deficiency), biliary diseases (obstruction, cholecystitis, cholangitis), musculoskeletal and cardiovascular disorders, and drugs other than ICIs (5,24). Furthermore, there are substantial differences in laboratory findings between ILICI and autoimmune hepatitis (AIH). Patients with AIH demonstrate higher serum levels of bilirubin, gammaglobulin and immunoglobulin G (IgG) compared to those with ILICI. Circulating autoantibodies, such as ANA, ASMA, pANCA are positive in cases of AIH and normally absent or present in low titers in cases of ILICI (9,12,25).

Liver biopsy is not routinely recommended in the diagnostic work-up of ILICI, since the histopathological findings are non-specific and not pathognomonic, thereby limiting its utility in guiding clinical decision-making. That is reflected in current practice, as only 64 (58.2%) of the reviewed patients underwent liver biopsy. Nevertheless, histopathological findings between AIH and ILICI demonstrate several distinguishing differences. Active AIH is characterized by a predominately plasma cell rich portal infiltrate extending to the periportal parenchyma of the liver often causing necrosis (interface hepatitis). ILICI demonstrates a mainly lobular pattern of hepatic injury, which in the case of anti-PD(L)-1 agents, is heterogenous and may involve the periportal liver parenchyma, whereas in the case of anti-CTLA4 agents it is characterized by a multinucleated giant cell reaction with fibrin deposition and central vein endotheliitis (6,26,27). Additionally, ILICI is associated with CD3+ and CD8+ lymphocytic infiltration and lesser CD20+ or CD4+ lymphocytes, which are commonly observed in AIH (8,28,29). Variations in lymphocyte subtypes are anticipated due to the predominant expression of PD-1 and CTLA4 in CD8+ cytotoxic T lymphocytes (8). The histopathological evaluation of our cases demonstrated a predominately CD3+, CD8+ T-lymphocytic infiltrate mainly distributed periportally with frequent extension to the hepatic lobules. CD3+, CD4+ T-lymphocytes were present to a lesser extent, mainly periportally. B-lymphocytes and plasma cells were infrequent. CD68+ histiocytes were increased in the hepatic sinusoids in two of our cases (Fig. 1). In the cohort of the published clinical cases, liver biopsies revealed heterogeneous but recurring patterns, most notably lobular hepatitis with CD8+ and CD3+ T-cell-predominant infiltrates, which distinguish ILICI from classical AIH as described above.

Assessment of the toxicity grading is pivotal concerning the management of ILICI according to the ESMO and NCCN guidelines. For grades 1 and 2, a close monitoring of LFTs and differential diagnosis to exclude other causes of liver failure should be performed. Immunotherapy should be continued for grade 1 (AST/ALT elevated up to three times the upper limit of normal-ULN) with close monitoring of liver enzymes, while for grade 2 (AST/ALT three to five times the ULN) additional laboratory and imaging testing are required, and therapy should be temporarily withheld until grade 1 or full recovery of LFTs. If there is no improvement in laboratory findings in spite of ICIs discontinuation, administration of corticosteroids 0.5 to 1 mg/kg/day prednisone or equivalent is recommended, and the dose should be readjusted to 1-2 mg/kg/day if there is no response within 72 h after initiation. Immunotherapy shall be resumed after improvement of liver function. For grades 3 (AST/ALT up to five to twenty times ULN) and 4 (ASL/ALT more than twenty times the ULN), ICIs should be permanently discontinued and treatment with corticosteroids 1 to 2 mg/kg/day prednisone or equivalent should start promptly. In refractory cases, when corticosteroids do not yield a response within 2-3 days, it is advisable to contemplate alternative immunosuppressive agents, such as mycophenolate mofetil (MMF), azathioprine, cyclosporine etc.

Regarding our patients only one had grade 2 hepatotoxicity, while the rest exhibited grades 3 and 4. With respect to the published literature, 101 patients (91.9%) of the presented with grade 3 and grade 4 hepatitis (45.5 and 46.4% respectively). The severity of the disease is evaluated using CTCAE grading system. We treated all five patients with corticosteroids, and one patient with persistent ILICI received mycophenolate mofetil five days after corticosteroid failure. Concerning the cohort of the published case reports, 108 patients were treated with glucocorticoids (98.2%), while of the two remaining patients, one was treated solely with budesonide and for the other one there is no information provided concerning the treatment. Out of 108 patients, 38 patients (34.5%) received only glucocorticoids, and 71 patients (64.5%) required more than one additional agent.

The administration of infliximab is contraindicated in immune mediated hepatitis as it predisposes in acute liver failure (6,8,30,31). Nonetheless, recent publications present evidence suggesting that the use of infliximab in cases of steroid-refractory ILICI could provide a positive and durable response, but more sufficient data are needed to establish its use as standard treatment (32). After patient's response, corticosteroid tapering should have a duration of 4-6 weeks, with the possibility of re-escalation if necessary. Of note, for grade 3 or 4 liver injury monotherapy with anti-PD(L)-1 agent may be resumed in patients previously treated with a combination regimen (6,33).

In our cases immunotherapy was permanently discontinued. Treatment continuation was not consistently reported across the published case reports, therefore a reliable conclusion cannot be reached. ICI resumption or rechallenge should follow an individualized strategy. Current clinical guidelines recommend permanent cessation of ICIs for most grade 4 toxicities, particularly in the context of severe or life threatening irAEs (e.g. myocarditis, nephritis) (6,34). Retreatment or rechallenge with ICIs is feasible and is typically reserved for patients whose toxicities have been completely resolved and were non-life-threatening (e.g. skin, endocrine, gastrointestinal) (35,36). Nevertheless, emerging evidence suggests that re-administration of ICIs is reasonable even after severe irAEs in carefully selected patients. This approach requires strict multidisciplinary supervision and close monitoring, along with the incorporation of risk-mitigation strategies, such as ICI class switching and organ-specific prophylaxis to reduce the risk of recurrent toxicity (37,38). Published data show that restarting ICIs after irAEs is associated with recurrent toxicity and increased safety concerns, while efficacy appears comparable to initial treatment (39,40).

The mean time between the first laboratory confirmation of hepatic disfunction in our cohort and corticosteroids initiation was seven days. Resolution of ILICI occurred in three patients (60%) and normalization of AST, ALT and bilirubin values was observed on average within three months of treatment, ranging from one to five months. In relation to literature-derived cohort, ILICI was resolved in 85.5% of the patients (94 patients), and median time to resolution was 2 months, ranging from 3 days to 24 months. Furthermore, of the five patients, four died subsequent to ILICI onset. Two fatalities were directly related to ILICI, whilst the other two were associated with complications during hospitalization unrelated to the immune-mediated adverse event. Regarding the cohort of the published case reports, death occurred in 14.5% (16 patients out of 110) highlighting that ILICI can be a fatal complication.

In conclusion, we present five cases of patients attended in our hospital after developing ILICI. All patients were treated with ICIs for different primary tumors and disease course. This small case series combined with an extensive literature review demonstrates that the time of occurrence of ILICI cannot be predicted, as it varies from a few days to several months after the first dose. Timely recognition and appropriate management are essential components of clinical care, although their prognostic impact was not formally assessed in the present study. Additionally, it needs to be noted that even if the majority of patients recover, there is a critical proportion that will succumb to the disease.

Supplementary Material

Published case reports.

Acknowledgements

Not applicable.

Funding

Funding: No funding was received.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

MK, AK and APs substantially contributed to the conception and design of the study. GK, AB, IK and RZ contributed to data acquisition and curation. PE, APa, MA, MM and NG contributed to data analysis and interpretation. PF and MS performed pathological analysis and validation, while KP contributed to radiological analysis. IK and AB conducted the literature search and data visualization. MK, IK and PE drafted the original manuscript. All authors contributed to critical revision of the manuscript for important intellectual content. AK and APs provided supervision, and AK acquired funding. MK and AK confirm the authenticity of all the raw data. All authors read and approved the final manuscript.

Ethics approval and consent to participate

Not applicable.

Patient consent for publication

Written informed consent was obtained from all patients (or their legal guardians/next of kin, where applicable) for the publication of anonymized clinical data and images. All identifying details have been removed to protect patient privacy. In cases where patients were deceased, informed consent for publication was obtained from the next of kin.

Competing interests

The authors declare that they have no competing interests.

References