THU0616 Elevated thyroid stimulating hormone as a potential biomarker for rheumatic immune-related adverse events following pd-1 inhibitor therapy

Background Programmed cell death protein (PD-1) inhibitor immunotherapy is being increasingly used in oncology, but may cause immune-related adverse events (irAEs) resembling classical rheumatic and non-rheumatic autoimmune diseases. While oncological response to therapy has been associated with the development of rheumatic irAEs, no biomarker to predict the development of rheumatic irAEs has yet been identified. Thyroid stimulating hormone (TSH) is readily available in clinical practice and is frequently tested prior to the administration of PD-1 inhibitors to screen for existing thyroid disease. Objectives To investigate whether thyroid stimulating hormone, or changes to it, are associated with the development of immune-related adverse events (irAEs) following PD-1 inhibitor therapy for cancer. Methods This was a retrospective chart review of all patients at a single centre who had a TSH level performed in the institutional laboratory prior to the patient being dispensed nivolumab or pembrolizumab before January 1, 2017, with follow-up until July 1, 2017. TSH levels before and during PD-1 inhibitor therapy were recorded. Patients with any diagnosis of a rheumatic irAE or any non-cutaneous irAE were identified. Linear regression was performed to determine the relationship between TSH and the development of irAEs. Youden’s index was used to derive the optimal cut-off point. Recursive partitioning methods with imbalanced prior probabilities were used to create a decision tree model. Results There were 213 episodes of therapy which met criteria, of which a non-cutaneous irAE occurred in 62 episodes (29.1%), thyroid irAEs occurred in 22 episodes (10.3%) and rheumatic irAEs were diagnosed in 16 episodes (7.5%). Even when corrected for duration of exposure to PD-1 inhibitor therapy, elevated TSH levels (u003e4.2 mU/L) were statistically significantly associated with the development of rheumatic irAEs (adjusted OR 6.08, 1.53–24.22), and this was not weakened by excluding patients who went on to develop thyroid irAEs. There was no significant association between elevated TSH levels and the development of any non-cutaneous irAE. Change in TSH levels with PD-1 inhibitor therapy was also not associated with the development of non-cutaneous irAEs or rheumatic irAEs specifically. Using a TSH level u003e2.4 to predict rheumatic irAEs led to a positive predictive value of 25% and a negative predictive value of 93% in our cohort. A decision tree model to predict rheumatic irAEs combining a pre-PD-1 inhibitor TSH level u003e2.4 and an oncological response to therapy led to a positive predictive value of 50% and a negative predictive value of 94% in our cohort. Conclusions Elevated TSH levels may be a potential biomarker for the development of rheumatic irAEs. In particular, a pre-PD-1 inhibitor TSH level of u003e2.4 mU/L, in combination with oncological response to therapy, may identify patients at risk of rheumatic irAEs. Associations observed in this cohort should be examined in larger cohorts to determine the clinical utility of TSH in predicting rheumatic irAEs. Disclosure of Interest None declared