e18585 Background: Racial/ethnic minority (REM) and limited English proficiency (LEP) patients with cancer receive more intensive treatment at the end-of-life (EOL) compared to non-Hispanic whites. Various factors at the health system, patient, and provider level contribute to this disparity. While some data exists on the role of provider communication, little is known about medical oncologists' knowledge and attitudes regarding EOL care communication with REM and LEP patients with cancer. Methods: We conducted a cross-sectional, anonymous, pilot survey of medical oncology attendings and fellows at one academic center. Over 3 months, 57 individuals were invited to participate. The survey addressed 6 domains about EOL care communication with REM and LEP patients (role of language, barriers, facilitators, provider beliefs, provider communication skills, and cultural competency training). Questions used Likert scales and were modified from previously validated items when possible. Results: Of 35 individual respondents (61% response rate) 21 (60%) were male, 14 (40%) were female, and 29 (83%) were attendings. Average participant age was 42 years (range 31-60 years) and average years in practice was 9 (range 2-25 years). Twenty-three (66%) self-identified as non-Hispanic white, 11 (31%) as Asian, and 1 (3%) as other. Thirty-three (94%) reported their primary language as English and 23 (66%) reported speaking a second language. A majority (77%) strongly agreed discussing EOL care is part of an oncologist’s job and 69% strongly agreed patients with cancer, regardless of race/ethnicity or English language ability, should be engaged in EOL care discussions. Among several facilitators identified, two most highly rated as “Very Important” were eliciting patient belief systems about their disease (71%) and about death/dying (69%). However, only 14% strongly agreed they ask patients their beliefs about death/dying, and a minority strongly agreed they feel comfortable having EOL care discussions with REM (34%) and LEP (29%) patients. Importantly, 77% rated themselves at 6 or 7 (on a scale from 0 to 10, with 10 being the best) in how well they currently communicate about EOL care with REM and LEP patients, 74% reported wanting to be at 8 or 9 in 6 months, and 89% completely or somewhat agreed with wanting help to improve. Conclusions: Our results suggest there is conceptual understanding among academic medical oncologists of the need to engage all patients in EOL care discussions and elicit patient belief systems on the topic. Though there is lack of comfort in having EOL care discussions with REM and LEP patients, and infrequent inquiry about patient belief systems, there is also motivation to improve communication skills in this area. Interventions that foster medical oncologists' comfort and skills in cross-cultural, language non-congruent communication about EOL care are needed.