Background

Immune Checkpoint Inhibitors (ICIs), which have shifted the cancer treatment paradigm, induce the human immune system to attack tumor cells by blocking pathways such as CTLA-4 or PD-1. Behind their excellent anticancer effects lies the fatal risk of Immune-Related Adverse Events (irAEs), similar to autoimmune diseases. Toxicities that manifest unexpectedly in systemic organs, such as colitis, hepatitis, pneumonitis, and myocarditis, determine patient survival rates depending on early diagnosis and the decision to discontinue treatment.

The problem is that most patients enrolled in the existing Phase 3 randomized controlled trials (RCTs) used as the basis for drug approval were White. In many clinical cohorts of immune-oncology drugs that received FDA approval, the proportion of Black, Asian, and Hispanic patients remained in the single digits. The unique genetic predispositions, immune system reactivity, and environmental characteristics of non-White populations were not fully reflected in the safety assessment stage prior to new drug approval.

Key Findings

To overcome sample bias in clinical trials, researchers conducted precise tracking by constructing a large-scale Real-World Data (RWD) cohort based on Electronic Health Records (EHR). They stratified the medical records of tens of thousands of cancer patients prescribed ICI monotherapy or combination therapy across major medical institution networks in the United States by race and ethnicity.

Analysis revealed distinct differences in the frequency and patterns of specific organ toxicities between White and non-White groups. The Black patient group showed significantly higher rates of immune-mediated pneumonitis and severe skin adverse events compared to the White patient group. Higher rates of Grade 3 or higher severe adverse events, requiring high-dose corticosteroids or additional immunosuppressants, were also observed. In the Asian patient population, the risk of endocrine toxicities, including hepatotoxicity and thyroid dysfunction, was higher than that in the White control group.

The incidence of multi-organ toxicity, which had been obscured in clinical trials due to insufficient sample size, was also newly revealed. Even after controlling for confounding variables such as patient age, cancer type, stage, comorbidities, and concomitant cytotoxic chemotherapy using propensity score matching, these racial disparities in odds ratios remained robust. This effectively proves with vast data that existing approved clinical trials failed to transparently reflect the clinical reality of actual multi-ethnic patient populations.

Implications and Outlook

These results clearly demonstrate that ensuring diversity in clinical trial participants is not merely a normative recommendation but a scientific necessity for patient safety in global new drug development and regulatory review. In line with the trend of the US FDA mandating Diversity Action Plan guidelines, developers of immune-oncology agents now face the task of strictly managing racial enrollment ratios from the initial clinical protocol design.

There is also expected to be increased momentum for advancing active pharmacovigilance systems using EHR and insurance claims data post-market. However, retrospective real-world data analysis leaves room for confounding by errors in the recording of racial categories within medical records or disparities in healthcare access due to socioeconomic factors. The possibility that differences in clinic visit intervals based on patient residence or income level influenced the early detection rate of toxicity cannot be ruled out. Follow-up research should proceed toward elucidating biological mechanisms at the molecular level, such as ethnic polymorphisms in Human Leukocyte Antigen (HLA) genotypes and T-cell receptor diversity.