Guruprasad P. Aithal教授：重新审视甲氨蝶呤肝毒性与免疫检查点抑制剂诱导的肝损伤

Q1: Your presentation at this conference is titled “Methotrexate and Liver: Half-life of truth.” How should we interpret this theme?

Prof. Guruprasad P. Aithal: The central theme of my presentation challenges long-standing concerns regarding methotrexate-induced liver fibrosis. For decades, clinicians have relied on biomarkers such as alanine aminotransferase (ALT) to monitor hepatotoxicity. However, ALT is a poor marker for chronic liver injury, particularly fibrosis. Subsequently, markers like procollagen type III amino-terminal peptide (P3NP) were introduced, but these too primarily reflect inflammation, not specific to the liver, rather than true fibrotic changes.

Our multicenter cohort study, encompassing 999 patients across numerous sites and utilizing a nested case-control design, demonstrated that methotrexate, particularly at low doses, is not a causal factor in the development or progression of chronic liver disease. Instead, hepatic fibrosis observed in these patients is predominantly driven by underlying metabolic risk factors, such as elevated body mass index (BMI) and type 2 diabetes mellitus. These findings underscore the necessity of shifting our monitoring focus toward patients with increased susceptibility to steatotic liver disease (SLD), for which transient elastography currently represents one of the most accurate and non-invasive assessment tools.

Q2: Current research suggests that although the cumulative dose of methotrexate is not directly associated with hepatic fibrosis, patients with metabolic syndrome or metabolic-associated fatty liver disease (MAFLD) exhibit a significantly increased risk of hepatotoxicity. In clinical practice, how should we define the therapeutic threshold for this high-risk population? Should the monitoring frequency be adjusted, or should more sensitive tools for liver fibrosis assessment be employed?

Prof. Guruprasad P. Aithal: It is important to distinguish between two distinct categories of methotrexate-associated liver toxicity: chronic fibrosis and acute liver injury.

In the setting of low-dose methotrexate, as studied in our cohort, we observed no excess risk of fibrosis attributable to methotrexate itself. Any liver injury in these patients typically reflects the burden of pre-existing metabolic liver disease rather than methotrexate exposure. Thus, clinical decisions regarding methotrexate continuation should be guided primarily by therapeutic efficacy rather than hypothetical concerns over fibrosis.

Acute liver injury remains a valid concern, particularly within the first year of therapy initiation. Our data indicate that during this period, elevations of liver enzymes exceeding five times the upper limit of normal occur at approximately 40 per 1,000 patient-years (roughly 4%). Applying more stringent criteria, the incidence falls to approximately 20 per 1,000 patient-years (2%). Beyond the first year, the incidence of significant liver injury drops further to 2 per 1,000 patient-years (0.2%), supporting a substantial reduction in monitoring intensity after the initial treatment year.

To optimize monitoring, we have developed a computerized predictive model incorporating early markers such as transaminase elevations within the first six months, early cytopenias, rise in creatinine and the presence of diabetes. This model enables personalized risk stratification: approximately 60% of patients fall into a very low-risk category (＜5% five-year risk), 10% are classified as moderate risk (5%-10%), and 25% are identified as requiring more intensive monitoring. For the majority of patients, annual liver function testing is sufficient once the first year of therapy has been successfully navigated.

Q3: In patients with a high tumor burden who require rapid re-initiation of immune checkpoint inhibitor (ICI) therapy, how can clinicians balance corticosteroid dosing with the potential risk to oncologic outcomes? Should steroid regimens be tailored according to the subtype of liver injury?

Prof. Guruprasad P. Aithal: This is a crucial and increasingly encountered clinical dilemma in oncology-hepatology interface. The management approach must be tailored to the histologic and biochemical pattern of liver injury.

For patients exhibiting cholestatic or mixed-pattern liver injury, immunosuppression should generally be avoided. In such cases, ursodeoxycholic acid may be considered, although high-quality evidence remains limited. Importantly, corticosteroids have demonstrated no clinical benefit in cholangiopathies, including those induced by ICIs.

Conversely, in hepatocellular injury patterns—provided bilirubin levels remain within normal limits—close biochemical monitoring over one to two weeks often reveals spontaneous improvement, negating the need for corticosteroids. The JAMA study I presented demonstrated that patients who received corticosteroids within the first two months of ICI therapy experienced inferior oncological outcomes compared to those who did not receive steroids, further supporting this conservative approach.

When corticosteroid therapy is considered unavoidable, histological confirmation via liver biopsy is ideal to ensure the presence of significant inflammatory infiltrates. In scenarios where biopsy is not feasible but hyperbilirubinemia is present, cautious initiation of low-dose corticosteroids may be warranted. High-dose corticosteroids, however, are not recommended, as current evidence does not support their efficacy in expediting recovery from drug-induced liver injury.

Q4: Looking ahead, is it possible to develop predictive models through multi-omics integration to enable pre-treatment risk stratification of hepatotoxicity associated with ICI therapy?

Prof. Guruprasad P. Aithal: Indeed, this represents one of the most promising frontiers in hepatotoxicity research. While I did not present these data during this conference, we have recently published findings in an oncology journal describing a novel circulating biomarker associated with ICI-induced liver injury. Simultaneously, we are advancing research into genetic biomarkers, and we have successfully developed a cell-based assay that accurately diagnoses ICI hepatotoxicity.

Although these tools remain under evaluation, our goal is to integrate these multi-omics platforms into comprehensive predictive algorithms within the next one to two years. Such personalized risk stratification models hold great promise for optimizing patient selection, maximizing therapeutic benefit, and minimizing the risk of hepatotoxicity in patients receiving immune checkpoint inhibitors.