Time to use the right classification to predict the severity of checkpoint inhibitor-induced liver injury, as assessed for causality using the updated RUCAM

Abstract

Background and aims: While immune checkpoint inhibitors (ICIs) are revolutionising cancer therapy, checkpoint inhibitor-induced liver injury is a significant immune-related side effect of this immunotherapy. This study focuses on the severity classifications and characteristics of patients with checkpoint inhibitor-induced hepatitis.

Methods: A retrospective analysis of patients with severe Checkpoint Inhibitor-induced hepatitis grade 3 and 4 according to the recommended Common Terminology Criteria for Adverse Events (CTCAE) classification was conducted. Data on clinicobiological characteristics, treatment and outcomes were collected from 3 university hospitals, and causality was assessed by using the updated Roussel Uclaf Causality Assessment Method. The severity of hepatitis was assessed using the Model for End-stage Liver Disease score, the Drug-Induced Liver Injury Network, and the Drug-Induced Liver Injury International Expert Working Group classifications.

Results: We retrospectively included 100 patients presenting various hepatitis patterns with a median time to onset of 20 days after checkpoint inhibitors. Severity grading varied significantly among the classifications used. A lower incidence of severe cases was observed when using the Drug-Induced Liver Injury classifications instead of the recommended CCTCAE classification, and this was correlated with outcomes.

Conclusions: This retrospective study challenges the efficacy of the CTCAE classification in defining the severity of Checkpoint Inhibitor-induced hepatitis and suggests that the traditional hepatology-focused scores may be more relevant. The CTCAE classification is inconsistent and gives equal weight to jaundice and elevated transaminases, which leads to steroid overtreatment and limits the rechallenge of ICIs.

FIGURE 1

Flow chart.

FIGURE 2

Acute liver injury estimated according to severity classifications CTCAE, DILI‐N and DILI‐IEWG. (A) Acute liver injury stratified by severity according to CTCAE: G3 vs. G4; p = 0.002. (B) Acute liver injury stratified by severity according to DILI‐N; p < 0.0001. (C) Acute liver injury stratified by severity according to DILI‐IEWG; p < 0.0001. Acute liver injury rate was estimated using the Kaplan–Meier method and Cox test regression in percentage. Non severe is grades 1 and 2; Severe is grades 3 and 4.

FIGURE 3

Correlation of severity classifications to predict acute liver injury. Graph shows receiver operating characteristic curve for predicting Acute Liver Injury. Solid blue line indicates Common Terminology Criteria for Adverse Events (CTCAE), solid red line indicates MELD (Model for End‐stage Liver Disease), solid green line indicates DILI‐N (Drug Induced Liver Injury‐Network), and solid orange line indicates DILI‐IEWG (Drug Induced Liver Injury‐International Expert Working Group).

FIGURE 4

Proposed strategy for the management of patients with severe CHILI based on our results. Non severe is grades 1 and 2; Severe is grades 3 and 4. ALI, acute liver injury; CTCAE, Common Terminology Criteria for Adverse Events; DILI‐IEWG, Drug Induced Liver Injury‐International Expert Working Group; DILI‐N, Drug Induced Liver Injury‐Network.