Background: Immune-related adverse events (IRAE) pose a significant diagnostic and therapeutic challenge in patients treated with immune-oncology (IO) drugs. IRAEs have been suggested to correlate with better outcome, but studies are conflicting. Estimating the true incidence of IRAEs is particularly difficult in the early phase I/II trial setting. A key issue is the lack of IRAE diagnostic criteria, necessary to discriminate "pure" IRAEs from other treatment-related adverse events not sustained by an autoimmune process.
Methods: In patients treated with immune-oncology (IO) drugs in phases I-II trials at our institute, we identified high confidence (HC) or low confidence (LC) IRAEs by clinical consensus. We empirically developed an IRAE likelihood score (ILS) based on commonly available clinical data. Correlation with outcome was explored by multivariate Cox analysis. To mitigate immortal time-bias, analyses were conducted (1) at 2-month landmark and (2) modeling IRAEs as time-dependent covariate.
Results: Among 202 IO-treated patients, 29.2% developed >1 treatment-related adverse events (TRAE). Based on ILS >5, we classified patients in no IRAE (n = 143), HC IRAE (n = 24), or LC IRAE (n = 35). hazard ratios (HR) for HC were significantly lower than LC patients (HR for PFS ranging 0.24-0.44, for OS 0.18-0.23, all P < .01).
Conclusion: ILS provides a simple system to identify bona fide IRAEs, pruning for other treatment-related events likely due to different pathophysiology. Applying stringent criteria leads to lower and more reliable estimates of IRAE incidence and identifies events with significant impact on survival.
Keywords: biomarker; clinical trial; immune-related adverse events; immunotherapy; predictive factors.
© The Author(s) 2023. Published by Oxford University Press.