In a timely study, Hountondji et al. [1] highlight an important limitation of the Common Terminology Criteria for Adverse Events (CTCAE) classification, which is widely used by oncologists in the management of checkpoint inhibitor-induced liver injury (ChILI). Authors found Drug-induced Liver Injury International Expert Working Group (DILI-IEWG) classification of severity had superior performance characteristics in predicting acute liver injury (compared with CTCAE and Model for End-stage Liver Disease) [2]. One of the reasons for inferior performance of CTCAE is because the latter overestimates severity of ChILI as life-threatening on the basis of alanine transaminase levels of 20-fold upper limit normal alone in isolation of other clinical criteria; this leads to avoidable hospital admissions [3]. In contrast, DILI-IEWG was able to predict 3 months of mortality while v5 of the CTCAE that incorporates total bilirubin was not able to [1].

Majority of patients with ChILI are started on high-dose corticosteroid therapy based entirely on CTCAE grading despite lack of evidence in its support [4]. A recent study highlights the value of liver biopsy and incorporation of severity of inflammation on liver histology into decision-making regarding corticosteroid therapy in patients with ChILI [5]; following this strategy, two-thirds of patients classified as having severe ChILI were spared of corticosteroid therapy. Early identification of poor outcome is crucial for planning management of ChILI since these patients are not candidates for liver transplantation in case of acute liver failure development [6].

Although Hountondji et al. [1] focus on the accuracy of methods of grading severity of ChILI, overestimation of severity of this adverse event may lead to unnecessary, prolonged and high dose of corticosteroid therapy in some with consequent complications [4].

在一项及时的研究中，Hountondji 等人 [1] 强调了不良事件通用术语标准 （CTCAE） 分类的一个重要局限性，该分类被肿瘤学家广泛用于检查点抑制剂诱导的肝损伤 （ChILI） 的管理。作者发现药物性肝损伤国际专家工作组 （DILI-IEWG） 的严重程度分类在预测急性肝损伤方面具有更好的性能特征（与 CTCAE 和终末期肝病模型相比）[2]。CTCAE 性能不佳的原因之一是，后者根据单独丙氨酸转氨酶水平为正常上限的 20 倍，孤立于其他临床标准，高估了 ChILI 的严重程度，使其危及生命;这导致可避免的住院 [3]。相比之下，DILI-IEWG 能够预测 3 个月的死亡率，而包含总胆红素的 CTCAE v5 则不能 [1]。

大多数 ChILI 患者完全基于 CTCAE 分级开始大剂量皮质类固醇治疗，尽管缺乏证据支持 [4]。最近的一项研究强调了肝活检的价值，以及将炎症的严重程度对肝脏组织学的影响纳入 ChILI 患者皮质类固醇治疗的决策中 [5];按照这种策略，三分之二被归类为严重 ChILI 的患者免于皮质类固醇治疗。早期识别不良结局对于规划 ChILI 的管理至关重要，因为这些患者在急性肝衰竭发展的情况下不适合肝移植 [6]。

尽管 Hountondji 等人 [1] 专注于 ChILI 严重程度分级方法的准确性，但高估这种不良事件的严重程度可能会导致一些患者不必要、长期和高剂量的皮质类固醇治疗，并导致并发症 [4]。