Purpose: Conflicting data have been reported concerning the prognostic value of autoimmune antibodies in patients with melanoma treated with adjuvant interferon alfa-2b (IFN). We evaluated the prognostic significance of autoantibodies in the European Organisation for Research and Treatment of Cancer 18991 trial, comparing long-term administration of pegylated IFN (PEG-IFN) with observation.
Patients and methods: Anticardiolipin, antithyroglobulin, and antinuclear antibodies were determined by enzyme-linked immunosorbent assays in 296 patients before random assignment and every 6 months after random assignment for up to 5 years. Prognostic impact of autoantibodies on recurrence-free survival (RFS) was assessed using the following three Cox models: a model that considered autoantibody appearance as a time-independent variable (model 1); a model that considered a patient to be autoantibody positive from the first positive test (model 2); and a model in which the most recent autoantibody test was used to define the status of the patient (model 3).
Results: Patients who were autoantibody negative at baseline were analyzed (n = 220). Occurrence of autoantibodies during follow-up was higher in the PEG-IFN-treated patients (18% in the observation arm v 52% in the PEG-IFN arm). Autoantibody appearance was of prognostic importance by using model 1 (hazard ratio [HR] = 0.56; 95% CI, 0.36 to 0.87; P = .01). However, when guarantee-time bias was taken into account using model 2 (HR = 1.19; 95% CI, 0.75 to 1.88; P = .46) or method 3 (HR = 1.14; 95% CI, 0.71 to 1.83; P = .59), significance was lost. Results were similar when treatment groups were analyzed separately.
Conclusion: Appearance of autoimmune antibodies is neither a prognostic nor a predictive factor for improved outcome in patients with melanoma treated with PEG-IFN.