Autoantibody-positivity before and seroconversion during treatment with anti-PD-1 is associated with immune-related adverse events in patients with melanoma.

Abstract

INTRODUCTION

Treatment with the immune checkpoint inhibitor anti-programmed cell death protein-1 (PD-1) often causes immune-related adverse events (irAEs). Since irAEs resemble autoimmune diseases, autoantibodies might play a role and could potentially be used to identify patients at risk. Therefore, we investigated the association between autoantibody-positivity and toxicity as well as clinical response in patients with melanoma treated with anti-PD-1.

MATERIALS AND METHODS

This two-center, retrospective study included 143 patients with melanoma treated with anti-PD-1. Toxicities grade ≥2 and recurrences/responses were captured until 6 months after treatment initiation. Autoantibody measurements were performed at baseline and 3 months after treatment initiation, including IgM-rheumatoid factor (RF), antinuclear antibodies (ANA), extractable nuclear antigen, anti-cyclic citrullinated peptide antibodies (anti-CCP2) and anti-thyroid antibodies.

CONCLUSION

The presence of autoantibodies prior to anti-PD-1 therapy is associated with irAEs in patients with melanoma. Both baseline positivity and seroconversion of anti-thyroid antibodies were strongly associated with thyroid dysfunction. This association was stronger in women, with all women who were baseline positive developing thyroid dysfunction.

RESULTS

169 irAEs were experienced by 86/143 patients (137 grades 1-2, 32 grades 3-4), the most common being thyroiditis (n=25), dermatitis (n=24), and sicca problems (n=19). Patients with autoantibodies at baseline experienced more irAEs (p=0.001), predominantly associated with anti-thyroid antibodies and thyroid dysfunction. No association was observed between any irAE and anti-CCP2, RF or ANA. In women, baseline and on-treatment anti-thyroid antibody-positivity as well as seroconversion during treatment was associated with thyroid dysfunction. In men, this association was only observed on-treatment. The presence of autoantibodies was not associated with melanoma recurrence (p=0.776) or response (p=0.597).