492 Background: The development and selection of treatments for locally advanced or metastatic urothelial carcinoma (la/mUC) have historically focused on clinical outcomes while stakeholder preferences are often considered less frequently. To facilitate the consideration of various perspectives, this study explored factors that may influence preferences of patients, caregivers, and physicians around aspects of la/mUC treatments. Methods: Interview guides elicited perspectives on disease impact (symptoms, health-related quality of life [HRQOL], survival), therapy goals and unmet needs and were conducted with patients with la/mUC, their caregivers, and medical oncologists in the US. Qualitative semi-structured data were collected on symptoms, treatment expectations, and hypothetical treatments that required trade-offs between overall survival (OS), progression-free survival (PFS), cancer pain, and the risk of severe adverse events (SAEs). Participant’s willingness to accept AEs was also explored. Results: Thirty participants, including: 10 patients (mean age 58 years; 60% female, ineligible for cisplatin: 50%), 10 caregivers (mean age 50 years; 70% female), and 10 physicians (mean 17 years treating la/mUC; 30% female) were interviewed. The most frequently reported symptom at diagnosis was pain (patients 90%, caregivers 90%) and blood in urine (physicians 100%). All three groups reported emotional impacts with depression/sadness the most common. Patients (n=7) and caregivers (n=7) relied on physicians for decision-making but felt that alternative treatments were not discussed (patients: n=5; caregivers: n=3). All groups were willing to accept some level of risk of experiencing an AE, but the accepted risk of SAEs varied (patients: 0-50% risk of SAEs; caregivers: 5-100%; physicians: 3-30%). Physicians focused treatment discussions on AEs (n=8), and overall response rate (ORR; n=6), and rarely discussed survival (n=3). All three groups described PFS and treatment response as very or most relevant to them. In the hypothetical choice tasks, all would make trade-offs between OS, PFS, pain reduction and risk of SAEs; consistent with an approach that weighs benefits and risks in treatment selection. Conclusions: Patients likely may benefit from shared, informed decision-making to identify the most appropriate treatment option for them based on clinical outcomes, AEs, HRQOL, and pain control. All groups were willing to make benefit-risk trade-offs but preferences were heterogeneous. While this study included a relatively small number of patients, planned research informed by these results will expand and further identify which treatment attributes are most important to patients.