Background: Immune checkpoint inhibitor (ICI) induced pneumonitis (ICIp) can be severe and even fatal, but risk factors for severe ICIp remain poorly characterized. In this abstract, we characterize ICIp and assess factors associated with severe disease. Methods: This is a retrospective study of consecutive patients who received ICI therapy and developed ICIp at the Ohio State University (2013-2020). Patients with ICIp were identified from a pharmacy database using ICD-10 codes followed by chart review to confirm diagnosis. ICIp was graded based on the CTCAE v5.0 criteria. Laboratory values were collected at the onset of ICIp. A thoracic radiologist reviewed chest imaging for pattern and severity. Survival probabilities were estimated with Kaplan-Meier curve and compared with log rank test. Results: Of 2963 patients who received ICI, we identified 119 patients (4%) with ICIp: 75 (63%) were males; 79 (66%) received PD-1 inhibitor monotherapy; and 71 (60%) had lung cancer. ICIp was severe (grade 1-2) in 69 patients (58%) and mild (grade 3-5) in 50 (42%). 74 patients (62%) were hospitalized. 63 patients (53%) needed oxygen supplementation. ICI therapy was held in all patients except one. ICI was rechallenged in 26 patients; 6 of them (23%) had recurrence of ICIp symptoms. All but 2 patients received steroids for the treatment of ICIp. 10 patients received non-steroidal medications (7 IVIG, 2 infliximab, 1 rituximab, 1 abatacept, 1 cyclophosphamide). Steroids were tapered in 6 patients of those who received non-steroidal medications. 3 of the patients who received IVIG had improvement of symptoms. Factors associated with severe ICIp. 21 patients (18%) died while hospitalized for ICIp are shown. Patients with severe ICIp had shorter overall survival than those with mild ICIp (P < 0.001). Conclusions: ICIp is a serious adverse event that limits ICI treatment and can lead to hospitalization and death. Factors associated with severe ICIp in our study include lower lymphocyte and eosinophil counts, lower albumin levels, finding of consolidation or bronchiectasis on imaging, and lower predicted FVC and TLC. Prospective studies are merited to validate our findings and to investigate the appropriate treatment of ICIp.