ABSTRACT We respond to Notice of Special Interest: Research in the Emergency Setting, and will build upon the Compre- hensive Oncologic Emergency Research Network (CONCERN). For many cancer patients, Immune Check- point Inhibitors (ICPIs) can be life-saving. However, the immune-related adverse events (irAEs) from ICPIs can be debilitating, and can quickly become severe, or even be fatal. Often irAEs will precipitate visits to the emer- gency department (ED). Therefore, early recognition and the decision to admit, observe or discharge these pa- tients from the ED can be key to a cancer patient’s morbidity and mortality. ED clinicians typically make their decision for disposition (admit, observe or discharge) within 2-6 hours from their patient’s ED presentation. However, irAEs are particularly challenging in the ED because of atypical presentations, the absence of classic symptoms, delayed availability of diagnostic tests during the ED encounter, and the fast pace in the ED setting. At present, there is no single sufficiently large ED data source with clinical, biological, laboratory, and imaging data that will allow for the development of a tool that will guide early recognition and appropriate ED dispo- sition of patients with potential irAEs. Therefore, we propose to capitalize on a multi-site collaboration among 4 CONCERN EDs (MD Anderson Cancer Center, Ohio State University, Northwestern University and University of California San Diego) in different States, to achieve the following aims: 1) To develop a probability model [the Immune-related Emergency Disposition Index (IrEDi)] to risk stratify ED patients on ICPIs for ED disposition. We will leverage our existing data (n=~2000) of unique ED patients who received ICPIs within 3 months of ED presentation at the 4 research sites. We hypothesize that host immune response underlie the development of irAEs and that inflammation/immune biomarkers available during the ED encounter will im- prove the prediction of Hospital admission; Observation; or Discharge, along with traditional factors, i.e. epide- miological factors (age, race/ethnicity), biological factors (sex, BMI), cancer (type, stage/metastases) and treat- ment-related variables (class of ICPI, monotherapy versus combination, dose/duration), and clinical status (comorbidities, preexisting autoimmune diseases, vital signs, laboratory results, imaging study results); and 2) To validate IrEDi using prospective data and determine the predictive validity of irEDi. We will conduct a pro- spective cohort study of ED cancer patients who had received ICPIs, recruiting 1500 total from 4 sites over a 3- year period. A common limitation of ED studies is that patients may receive care from multiple EDs. Thus, we will conduct follow-up calls in 30 days to assess ED revisits or hospitalization. We hypothesize that IrEDi devel- oped in aim 1 will have high sensitivity (≥90%) and high specificity (≥90%) for predicting appropriate ED dispo- sition (hospital...