Background: Immune checkpoint inhibitors (ICI) cause immune-related adverse events (irAEs). The spectrum of irAEs requiring referral to non-oncology specialists has not been well described. We established an immune-related toxicity (IR-Tox) team of oncology (n = 8) and medical subspecialists (n = 20), to support multidisciplinary irAE diagnosis and management. Methods: Patients (pts) treated with ICIs were electronically referred to the IR-Tox team between 01/2017-03/2018. IR-Tox team met monthly to discuss complex irAEs, and identify areas of clinical need. Pt demographics, treatment, and irAE details were collected. Pt features and irAE associations were analyzed using Chi-square tests. Results: The IR-Tox Team received 92 referrals for 80 pts (outpt: 61%, inpt: 39%). Median age was 65 years (range: 21-91), 55% were male, and 14% had prior autoimmune disease. Pts most commonly had non-small cell lung cancer (35%), melanoma (19%), or gynecologic malignancies (10%). Pts received ICI monotherapy (56%) or combination immunotherapy (44%), as standard-of-care (46%) or on clinical trials (54%). Referrals related to diagnosis (32%), management (11%), or both (51%) were received from faculty (68%), fellows (12%), and nurses (20%). Referrals were for suspected irAE (90%), pre-ICI assessment in known autoimmune disease (9%), or ICI re-challenge (1%). Sixty-three irAEs were confirmed (CTCAE grade 1 = 19%; 2 = 43%; 3+ = 38%), and 26 patients developed > 1 irAE (2 irAEs = 19, 3 irAEs = 7). IrAEs included pneumonitis (22%), arthritis (17%), dermatitis (13%), colitis/diarrhea (13%), idiopathic thrombocytopenia purpura (5%) hepatitis (5%) and thyroid dysfunction (5%). A new irAE of bony inflammation was identified (n = 2). Grade 3+ irAEs (p = 0.04) and colitis/diarrhea (p = 0.01) were more likely with combination immunotherapy. Medical specialty input was obtained in 85% of pts, and 35% required an invasive procedure for irAE diagnosis or management. Conclusions: Creation of an IR-Tox Team facilitated identification and management of complex irAEs. This new model provides a valuable forum to identify educational/service needs, manage multi-system irAEs, identify new irAEs, stratify irAE risk, and establish research collaborations.