Safety and survival outcomes of immune checkpoint inhibitors in metastatic lung cancer patients with pre-existing autoimmune disease: A propensity-matched real-world analysis.

e20615 Background: Patients with cancer and pre-existing autoimmune diseases (AID) are frequently excluded from immune checkpoint inhibitor (ICI) trials due to concerns about severe immune-related adverse events (irAEs) and AID flares. Real-world data on safety and outcomes of immunotherapy in this population remain limited. Methods: This retrospective cohort study used the TriNetX US Collaborative Network to identify adults with lung cancer (ICD-10 C34) and metastatic disease codes (C77–C79) treated with PD-1/PD-L1 inhibitors (pembrolizumab, nivolumab, atezolizumab, durvalumab, or cemiplimab) between 2018 and 2025. Patients with prior transplant, HIV, chronic viral hepatitis, or CTLA-4 inhibitor exposure were excluded. Those with pre-existing AID diagnosed prior to ICI initiation were compared with those without AID using 1:1 propensity score matching to balance demographics and major comorbidities. The primary outcome was overall survival; secondary outcomes included hospitalization, emergency department visits, systemic steroid use and organ-specific irAE proxies. Results: After propensity score matching, 3,159 patients were included in each cohort. Overall mortality did not differ between groups (54.5% vs 52.9%; HR, 1.03; 95% CI, 0.96–1.10; p = 0.43). However, patients with pre-existing AID experienced significantly higher rates of hospitalization (63.4% vs 59.2%; RR, 1.072; p = 0.001). Organ-specific irAEs were significantly elevated in the pre-existing AID cohort, including pulmonary irAEs (28.5% vs 24.0%; RR, 1.186; p < 0.001), gastrointestinal irAEs (19.8% vs 12.1%; RR, 1.638; p < 0.001), endocrine irAEs (14.8% vs 5.1%; RR, 2.895; p < 0.001), and myositis (1.4% vs 0.7%; RR, 2.143; p = 0.003). Systemic steroid use was common and similar between groups. Emergency department visits and dermatologic irAE proxies did not differ significantly. Conclusions: In this real-world study, metastatic lung cancer patients with pre-existing AID receiving PD-1/PD-L1 inhibitors demonstrated comparable overall survival to those without AID, supporting the use of ICIs in this population. However, significantly elevated rates of organ-specific irAEs—particularly gastrointestinal, endocrine, and pulmonary toxicities—and increased hospitalizations underscore the need for enhanced monitoring and multidisciplinary management. These findings support risk-adapted use of PD-1/PD-L1 inhibitors rather than routine exclusion of patients with AID.