Immune checkpoint inhibitors (ICIs) enhance the anti-tumor immune response by effectively relieving tumor-mediated immune suppression (1). Camrelizumab is a humanized monoclonal antibody targeting programmed cell death protein 1 (PD-1) that has demonstrated notable anti-tumor activity and clinical efficacy in advanced hepatocellular carcinoma and gastrointestinal malignancies, with a favorable safety profile (2, 3). However, irAEs induced by ICIs may affect multiple organs and systems, resulting in varying degrees of tissue damage. Immune-mediated hepatotoxicity (IMH) is a distinct form of hepatotoxicity specifically associated with ICI therapy.

The diagnosis of IMH primarily relies on abnormal liver function tests, with exclusion of other potential etiologies of liver injury, such as active viral hepatitis, autoimmune hepatitis, and hepatic metastases. The severity of IMH is classified into grades 1 to 5 according to the CTCAE v5.0 (4). During IMH, ALT and/or AST are typically markedly elevated, with or without concomitant increases in bilirubin levels. Most cases lack specific clinical symptoms. Elevated bilirubin is often indicative of an increased risk of liver failure, warranting close clinical monitoring and timely implementation of more aggressive intervention strategies (5).

Currently, the management of IMH remains limited and primarily relies on glucocorticoids and certain immunosuppressive agents. According to the “Guidelines for the diagnosis and treatment of Drug-induced liver Injury (2023 Edition)” issued by the Chinese Society of Clinical Oncology (CSCO), foundational hepatoprotective therapies mainly encompass anti-inflammatory agents, liver cell membrane stabilizers, and drugs that promote biliary excretion (6). Given the complex and heterogeneous pathogenesis of IMH, combination pharmacotherapy is commonly employed in clinical practice based on individual patient profiles, aiming to achieve multi-target intervention and synergistic effects. The use of glucocorticoids should be tailored according to the severity of liver injury. For glucocorticoid-refractory cases, mycophenolate mofetil (MMF) has been established as an effective second-line therapeutic option; however, there remains a lack of consensus regarding optimal third-line treatment strategies (6).

The primary treatment modalities of non-biological artificial liver support systems include plasma exchange (PE), double plasma molecular adsorption system (DPMAS), and continuous renal replacement therapy (CRRT). Among these, PE is currently recognized as the most effective liver support therapy and is the only intervention explicitly recommended by existing clinical guidelines. The European Association for the Study of the Liver (EASL) Clinical Practice Guidelines on Acute Liver Failure and the American Association for the Study of Liver Diseases (AASLD) guidelines both list artificial liver support therapy as a key management strategy for acute liver failure, clearly indicating its clinical efficacy in specific patient populations (7). Based on several retrospective controlled studies by Zhou, Zhong, Guo et al. (8–10), it has been demonstrated that the efficacy of DPMAS combined with PE is comparable to that of full-volume PE in the treatment of liver failure, with potential to further improve short-term outcomes in patients with acute liver failure. When plasma supply is limited, therapeutic efficacy can be maintained even with reduced plasma dosage (11). This case reports a patient with severe IMH who received multimodal therapy comprising hepatoprotective agents, glucocorticoids, and immunosuppressants, in combination with an artificial liver support system, and achieved a favorable clinical outcome.

With the expanding clinical application of ICIs, irAEs have become an increasingly important concern. IMH is a clinically significant hepatic irAE, with a reported incidence of approximately 1%–15% (12, 13). Evidence suggests that ICI therapy increases the overall risk of liver injury (relative risk (RR) = 1.40), particularly severe events (RR = 2.55) (14). A meta-analysis of 613 irAE-related deaths revealed that 124 cases (20.2%) were attributed to IMH (15), and an international multicenter melanoma analysis further indicated that IMH carried one of the highest mortality rates among irAEs (15). Against this background, the present case illustrates an unusually rapid-onset, high-grade IMH complicated by subacute liver failure (SALF), underscoring the need for timely recognition and escalation of care.

IMH remains a diagnosis of exclusion and requires careful differentiation from autoimmune hepatitis (AIH) and other causes of acute liver dysfunction. Classically, AIH is characterized by hypergammaglobulinemia (typically elevated IgG), positive autoimmune serologies, and histologic interface hepatitis with a prominent plasma-cell component (16). In our patient, serum IgG was within the normal range, and autoimmune serologies were negative (including ANA and anti-ENA), making AIH less likely. According to European Society for Medical Oncology (ESMO) guidance, liver biopsy is not routinely required when the clinical course after ICI exposure is typical and alternative etiologies have been excluded; however, histopathological assessment is recommended for atypical presentations or inadequate response to initial therapy (7, 17). In this case, biopsy was pursued because etiological attribution was complicated by combination therapy and the patient rapidly progressed to SALF, prior to escalation to intensive immunosuppression and extracorporeal support. Histopathology supported an immune-mediated phenotype and increased confidence in the working diagnosis and helped deprioritize competing etiologies. In addition, applying the R-value framework recommended in the Chinese Guidelines for drug-induced liver injury (DILI) (6, 18), the R value at onset was 0.75 (ALT 200 U/L; ALP 554 U/L), indicating a cholestatic pattern consistent with the disproportionate elevation of alkaline phosphatase and bilirubin.

IMH most commonly occurs 4–12 weeks after ICI initiation or after 1–3 doses (19), whereas our patient developed severe liver injury approximately 20 days after starting camrelizumab and progressed rapidly to SALF, representing an uncommon, fulminant clinical course. A structured work-up was performed to exclude alternative etiologies. Infectious causes were ruled out by negative serologic testing for hepatitis viruses (A–E), cytomegalovirus (CMV), and Epstein–Barr virus (EBV). Mechanical obstruction and vascular causes were also excluded: despite the cholestatic presentation, abdominal CT and color Doppler ultrasound showed no intrahepatic or extrahepatic bile duct dilatation and demonstrated normal portal venous flow. The temporal relationship and clinicopathologic correlation were more consistent with an immune-mediated process than with chemotherapy-related hepatotoxicity. Liver function remained stable during the initial SOX regimen but deteriorated rapidly after camrelizumab was introduced. Together with biopsy findings of portal lymphocytic aggregates, this pattern supports IMH as the most likely driver of SALF in this patient.

From a mechanistic perspective, irAEs are complex and likely multifactorial. In the liver, PD-1/PD-L1 signaling contributes to immune tolerance by restraining excessive intrahepatic T-cell activation; blockade of this pathway may lower the threshold for immune-mediated hepatocyte and/or bile duct injury (20, 21). Other proposed contributors include off-target immune activation and antibody-mediated tissue injury, although their relative roles likely vary across patients. Additionally, non–mutually exclusive mechanisms have been proposed, including antigen cross-reactivity/epitope spreading and cytokine-associated amplification of inflammation (22). In our patient, the short latency (~20 days after exposure) and IMH CTCAE v5.0 grade 4 severity are compatible with a rapid, clinically meaningful disruption of hepatic immune tolerance after PD-1 pathway blockade; however, mechanistic interpretation remains hypothesis-generating in a single-case report, particularly in the absence of longitudinal immune profiling.

Therapeutically, the patient showed limited biochemical response to first- and second-line hepatoprotective therapy. When total bilirubin increased to 171 μmol/L, artificial liver support was initiated in parallel with continued hepatoprotective agents, corticosteroids with adjunct immunosuppression, and supportive care. Following this multimodal escalation, hepatic synthetic function improved substantially (prothrombin activity (PTA) 30.4% → 80% → 97%), which temporally coincided with clinical stabilization and subsequent recovery. Prior reports suggest that PE can remove circulating bilirubin-bound toxins and inflammatory mediators and has been associated with immunologic shifts (e.g., changes in regulatory T-cell/Th17-related indices) (23); however, patient-level immune measurements were not available in our case, and such effects should be interpreted as literature-based hypotheses. The 2023 guideline for artificial liver blood purification systems emphasizes the potential value of timely initiation in appropriate candidates (24). DPMAS has also been used for bilirubin clearance and reduction of selected inflammatory mediators in cholestatic injury (25). In this case, a sequential PE–DPMAS strategy was adopted with the intent to support detoxification while reducing overall plasma utilization.

For glucocorticoid-refractory severe IMH, escalation to a multimodal strategy integrating immunosuppressants and artificial liver support may be considered in selected patients as a bridge to hepatic recovery. In rapidly progressive cases, such an approach may help stabilize liver function while definitive recovery evolves, although the effectiveness of individual components cannot be determined from a single case.

This case has several notable strengths. We performed a structured etiological work-up to exclude competing causes, confirmed IMH histologically in a complex post-chemotherapy context, and provided detailed documentation of a multimodal rescue approach for subacute liver failure. Several limitations should also be noted. As a single-case report, the observations have limited generalizability. Although the temporal relationship and clinicopathologic correlation favored IMH, concomitant oxaliplatin and S-1 exposure remains a potential confounder. Because multiple interventions were initiated in parallel, the relative contribution of each modality cannot be determined. Finally, the absence of longitudinal immune profiling limits mechanistic insight.

Despite these limitations, this case supports considering closer liver function monitoring in selected high-risk patients (e.g., advanced age or pre-existing hepatic burden). In addition to baseline testing, repeat liver biochemistry and coagulation assessment early after treatment initiation and before each cycle may be reasonable. Early use of hepatoprotective agents when transaminases begin to rise may also be considered, although the optimal indications and protocols require further study. Future work should address (1) criteria and protocols for hepatoprotection in high-risk groups; (2) early biomarkers and timing of corticosteroid intervention; and (3) initiation thresholds and therapeutic windows for artificial liver support in immune-related acute/subacute liver failure.

In summary, we report a patient with camrelizumab-associated IMH (CTCAE v5.0 grade 4) complicated by subacute liver failure. The patient improved after a multimodal approach including hepatoprotective therapy, corticosteroids with adjunct immunosuppression, and sequential artificial liver support using PE and DPMAS, with subsequent normalization of liver biochemistry on follow-up. Given the single-case nature of this report and the absence of longitudinal immune profiling, mechanistic inferences—particularly regarding immune modulation—should be interpreted cautiously. These observations are hypothesis-generating and suggest that artificial liver support may be considered as an adjunct in refractory severe IMH, warranting evaluation in larger prospective cohorts.