11007 Background: Hematology/oncology fellows must achieve bone marrow biopsy proficiency. However, a cost-effective, high-fidelity system to practice these procedures has not been described. Other specialties utilize 3D printing to practice procedures. Using design thinking, we developed, implemented, and evaluated a bone marrow biopsy training session with 3D pelvis models. Methods: We printed two models using an NIH template and optimized them through iterative prototyping. We covered a hole at the intended biopsy site with a replaceable cap simulating cortical bone, used cork as medullary bone, and used sand and water as marrow. Caps of two densities simulated softer and harder bone. Fellows could lift silicone skin pads to view anatomy. A flat base minimized movement. In July 2019, we conducted a one-hour practice session (“3D session”) with eight fellows during orientation. After an anatomy review, fellows practiced biopsies using the models with faculty feedback. Fellows also attended a one-hour session with a hematologist demonstrating a biopsy on a patient (“patient session”), the only session offered in previous years. We used a t-test to compare course ratings and pre/post-orientation self-assessed comfort with biopsies (5-point scales). Six months later, we surveyed attendings about fellow biopsy skill and success rate compared to prior years. S.B. conducted a content analysis of a focus group with four fellows and email feedback from one fellow. Results: Fellows rated the 3D and patient sessions highly (4.50 vs 4.75, p = 0.51). Procedural comfort improved significantly after orientation (2.13 to 3.63, p = 0.03). Attendings noted no difference between the 2019 fellows and prior years. Fellows called the 3D session “helpful” and “high-yield.” They praised the opportunity to practice repeatedly with high-fidelity anatomy, rehearse mechanics, receive feedback, and internalize anatomy and muscle memory for later recall. Fellows noted that the model did not allow for patient positioning practice and that the denser cap was too hard. Fellows suggested incorporating a female pelvis and more soft tissue. Conclusions: We developed, implemented, and evaluated a design-based bone marrow biopsy training session. Though we did not find outcome differences compared to traditional training, 3D printing represents a feasible, cost-effective, and high-fidelity educational tool. 3D sessions, in conjunction with patient sessions, may augment understanding of anatomy and provide opportunities for practice and feedback. Future iterations should incorporate user feedback to optimize model fidelity and utility.