The Novel Diagnostic Index Based on HLA-DRB1 Genotype and PD-L1 Expression can Predict Severe irAEs in Patients with Metastatic Melanoma Taking Immune Checkpoint Inhibitors. The Results of the Pilot Study

Abstract

Immune-related adverse events (irAEs) occur in up to 50% of patients treated with an anti-CTLA-4 antibody and 30% of patients treated with PD-1/PD-L1 antibodies. Severe forms of toxicity are observed in 3% of patients and require systemic steroid therapy and constant monitoring. One of the considered predictor biomarkers of irAEs development is HLA-genotypes. This research aims to evaluate the diagnostic significance of HLA-DRB1 genotypes and other clinical and laboratory parameters to predict the development of irAEs. The study involved 28 patients with metastatic melanoma taking checkpoint inhibitors therapy [nivo 53.6%, Ipi+nivo 32.1%, other (pembro, prolgo) 14.3%]. The PD-L1 expression and HLA-DRB1 genotype were evaluated. After 2-3 months the development of irAES was assessed. The complications of 3-4 grade or multi-organ damage were termed as severe irAEs. Various IrAEs developed in 57.1% (16/28) of patients, while severe irAEs occurred in 35.7% (10/28). Among all patients, HLA-DRB1 genotypes associated with the risk of autoimmune diseases were found in 78.5% (22/28). The PD-L1 expression was detected in 60.7% (17/28) of individuals. Combination treatment increases the risk of toxicity, p = 0.0028, with a diagnostic sensitivity of 56% and a diagnostic specificity of 100% (RR = 2.71, OR = 31.67). An index based on the parameters studied (HLA-DRB1, absence of PD-L1 expression, and type of treatment) was created. It allows assuming the risk of developing severe irAES (p = 0.0126). When comparing this indicator between irAEs 1-2 and irAEs 3-4, the presence of an index value of more than 2 gives a sensitivity for predicting severe toxicity of 40.00% and a specificity of 83.33%.

Non-Invasive Predictive Biomarkers for Immune-Related Adverse Events Due to Immune Checkpoint Inhibitors

“…In other words, further research is necessary to create a practical tool for irAE prediction. For example, Zhukova et al created a three-point scoring system consisting of the presence of HLA-DRB1, PD-L1 negativity, and dual ICI therapy and claimed its usefulness as a predictor for severe irAEs [ 132 ]. Third, most irAE studies have been performed in a retrospective fashion, requiring the vigilance of patients and physicians to document irAEs in patient charts.…”

“…6-item clinical likelihood score [131] Score > 5: associated with higher risk of irAE 3-item clinicopathological score [132] Score > 2: associated with severe irAE Psoriasis-associated polygenic risk scores for dermatologic irAEs [133] Higher score: associated with higher risk of dermatologic irAE 140-gene germline polygenic risk score [134] Higher score: associated with thyroid irAE 859-gene germline variant, HLA alleles, and blood counts, deep neural network model [34] Predictive of each organ-specific irAE 16-gene RNA expression signature [135] Higher score at 30 days after treatment: increased risk of colitis Combined 11 cytokine (CYTOX) score [136] Higher score: associated with higher risk of any irAE Abbreviation: irAE, immune-related adverse event.…”