Abstract 5388: Association of baseline systemic corticosteroid use with time to next treatment in patients with advanced melanoma, non-small cell lung cancer or urothelial cancer receiving cancer immunotherapy in US clinical practice

Background: Corticosteroids (CS) are often prescribed with cancer treatment to alleviate symptoms, treat comorbidities and manage treatment-related adverse events. The immunosuppressive properties of CS may decrease the effectiveness of cancer immunotherapy (CIT), and the effect may vary by tumor site. This study explored the association of baseline CS (bCS) use with time to next treatment (TTNT) in CIT-treated patients with advanced melanoma (aMel), advanced non-small cell lung cancer (aNSCLC) or advanced urothelial cancer (aUC). Methods: The Flatiron Health electronic health record-derived de-identified database was used to select patients diagnosed with aMel, aNSCLC or aUC between January 1, 2011, and June 30, 2017, who received CIT alone in any treatment line and were followed through March 30, 2018. bCS included intramuscular or intravenous administrations or oral orders ≤ 14 days before and ≤ 30 days after the CIT start date. TTNT is a measure of intermediate outcomes, such as cancer progression, in real-world data sources. A TTNT event was defined as initiation of a new line of non-maintenance treatment or death within 60 days of the last CIT treatment during the index line of therapy. Other patients were censored at the latter of either their last CIT administration or last recorded visit. The association of bCS with TTNT was estimated using multivariable Cox proportional hazards models adjusting for key baseline characteristics, including prior CS use. Results: Most patients were white men aged 66 to 72 years; median follow-up across tumor types was 3.9 to 5.5 months. One-fifth to one-third (19%-30%) of patients received bCS. More than half of patients (58%-61%) received a new treatment line during the study period, the milestone for measuring TTNT. Patients receiving bCS were more likely to have stage IV disease at diagnosis, brain or liver metastases (aNSCLC, aUC) and poorer ECOG performance status (aUC) at baseline. After adjusting for these and other important potential confounders, including prior CS use, bCS use was associated with shorter TTNT compared with no bCS use in multivariable models (aMel HR, 1.28 [95% CI: 1.03, 1.58]; aNSCLC HR, 1.40 [1.16, 1.70]; aUC HR, 1.36 [1.06, 1.82]). Conclusions: After adjustments for measured confounders, patients receiving bCS had a shorter TTNT than those who did not receive bCS, suggesting a shorter treatment duration that may limit the potential long-term benefits of CIT use. TTNT is commonly used as