Predictors, Phenotype and Imaging Correlates of Immune‑Mediated Cholangitis in a Large Single‑Center Cohort

Iijima and colleagues report a single-center retrospective cohort describing immune-mediated cholangitis among patients treated with immune checkpoint inhibitors (ICIs), focusing on baseline laboratory associations and biliary imaging correlates. Using a data-warehouse review of 1332 ICI-treated patients (2014–2023), the authors identified 81 cases of immune-mediated hepatotoxicity (IMH), including 10 classified as immune-mediated cholangitis (IMC). The report describes baseline markers associated with later IMC, the clinical phenotype and imaging patterns used for classification, and outcomes spanning treatment response and overall survival.

The investigators retrospectively evaluated patients receiving anti–PD-1, anti–PD-L1, and/or anti–CTLA-4 agents and assessed adverse events using CTCAE v5.0. They flagged potential IMH when AST, ALT, ALP, or GGT reached grade 2 or higher, then diagnosed IMH based on liver enzyme elevation with exclusion of other liver diseases per Japanese diagnostic criteria for ICI-induced liver injury. Within IMH, IMC was defined by characteristic biliary imaging abnormalities on contrast-enhanced CT and/or MRCP, including extrahepatic bile duct changes (localized dilation without obstruction and/or diffuse wall thickening) and/or intrahepatic duct changes such as PSC-like strictures. The authors compared IMC with non-cholangitis IMH within the IMH subset.

In multivariable models reported by the authors, baseline inflammatory and hematologic markers showed an association with subsequent IMC. Specifically, baseline eosinophil count >270/μL (OR 10.33; p=0.004) and baseline C-reactive protein (CRP) >0.8 mg/dL (OR 6.260; p=0.027) were identified as independent predictors of IMC onset, using study-derived cutoffs and logistic regression (with similar results in Firth penalized analysis). In univariate analyses, anti–PD-1 antibody monotherapy was associated with IMC occurrence; in the authors’ multivariable model (which included variables meeting their prespecified univariate threshold), baseline eosinophils and CRP emerged as independent predictors. In this dataset, higher baseline eosinophils and CRP marked patients who later met the authors’ IMC definition.

After onset, IMC was described as a later-onset phenotype during ICI treatment, with a predominantly cholestatic liver-injury pattern and a symptom profile notable for abdominal complaints, including abdominal pain. At IMH onset, the IMC group also had higher inflammatory measures (including white blood cell count, CRP, and neutrophil-to-lymphocyte ratio) alongside higher cholestatic enzymes (ALP and GGT) compared with non-cholangitis IMH. Imaging was central to the phenotype: contrast-enhanced CT and MRCP commonly showed bile duct dilation and wall thickening, and serial review indicated that biliary abnormalities were visible before biochemical onset in 5 of 10 cases (50%). The authors further report that imaging abnormalities persisted after treatment in 7 of 10 cases (70%) and were more common among non-responders, suggesting that radiologic persistence may be associated with treatment resistance in this case series.

In this retrospective cohort, reported IMC treatments included corticosteroids and (in some cases) ursodeoxycholic acid—used at physician discretion—as well as mycophenolate mofetil in selected patients; biochemical responses were variable, with improvement in some cases and relapse or non-improvement in others. For prognosis among patients with IMH, the authors report that IMC diagnosis itself was not independently associated with overall survival in multivariable analysis. In the same prognostic modeling, cholestatic-type liver injury (HR 2.318; p=0.023) and neutrophil-to-lymphocyte ratio ≥4.0 at IMH onset (HR 3.622; p=0.001) were identified as independent predictors of poorer overall survival. Taken together, the report links a cholestatic/inflammatory profile to the IMC phenotype and to the survival associations described within this single-center cohort.

Key Takeaways:

• In a large single-center ICI-treated cohort, IMC represented a small subset of immune-mediated hepatotoxicity cases, as classified by characteristic biliary imaging findings.
• The authors’ regression analyses identified higher baseline eosinophils and baseline CRP above study-specific cutoffs as independent predictors of later IMC.
• Serial CT/MRCP abnormalities (dilation and wall thickening) persisted after treatment in 7 of 10 IMC cases (70%) and may be associated with treatment resistance, while cholestatic-type injury and elevated NLR at IMH onset were independently associated with poorer overall survival in the cohort’s prognostic models.