Abstract Orthotropic liver transplantation (OLT) is a major surgical procedure associated with significant blood loss. Unanticipated blood requirements during OLT can delay the procedure and compromise lifesaving intraoperative resuscitation, adding strains to the blood bank and blood suppliers' resources, especially when units are requested on short-notice or for patients with alloantibodies. Age, preoperative laboratory values, severity of underlying liver disease, and coagulopathy are factors that influence the number of units required. Currently, each transfusion medicine service issues a specific number of crossmatched red blood cell (RBC) units for OLT patients. At our institution, 10 crossmatched units are issued for each patient, which may be an overestimation or underestimation of the number of units required for certain patients. Aim: To characterize OLT patients who require more than 10 RBC units intraoperatively, and compare them to those who require less, identify pre-operative clinical and laboratory factors that predict which patients will require more than 10 units. Finally develop an artificial intelligence platform (AI-OLTRBC-1) aiming at reducing inappropriate RBC transfusions. Methods: We collected data from 485 patients that underwent OLT from January 1st, 2021 to December 31st 2022. Data collected included demographics, coagulopathies, anticoagulants, ABO/RH group, and laboratory testing collected within 24 hours of admission (complete blood count, prothrombin time (PT), partial thromboplastin time (PTT), INR, ALT, AST, total bilirubin, conjugated bilirubin, alkaline phosphatase (ALP), creatinine, Blood urea nitrogen (BUN), calcium and thromboelastogram (TEG). Data analysis was performed in R-studio (V.1.4.1). Missing values were imputated using multivariate imputation by chained equations algorithm. AI-OLTRBC-1 was built using machine intelligence learning optimizer platform. Results: Based on univariate regression analysis (p-value<0.05, 95% confidence interval) the following laboratory factors were found predictive of OLT patients requiring more than 10 RBC units: low albumin (OR 1.47 [CI 1.05-2.04]), increased total bilirubin (OR 1.06 [CI 1.041.08]), increased conjugated bilirubin (OR 1.08 [CI 1.04-1.11], increased creatinine (OR 1.55 [CI 1.33-1.83]), BUN (OR 1.03 [CI 1.02-1.04]), low hemoglobin (OR 1.61 [CI 1.81-1.42]), MCV (OR 1.04 [CI 1.02-1.07]), high INR (OR 3.48 [CI 2.35-5.26]), PTT (OR 1.04 [CI 1.03-1.06]), TEG-coagulation index (OR 1.12 [CI 1.05-1.19]). Among the anticoagulants, patients on enoxaparin required more RBC units (OR 1.47 [CI 1.66-11.1]). Other factors including demographic and ABO/Rh were not significant. The median number of transfused RBC units in patients requiring more than 10 units was 20, while the median number of transfused RBC units in patients requiring less was 5. Conclusions: Our study aimed to characterize OLT patients who required more than 10 units and identified preoperative factors to estimate RBC unit requirements, which would aid transfusion medicine service in providing RBC units properly and on-time. AI-OLTRBC-1 achieved a sensitivity of 88.0%, specificity of 70.4%, and an accuracy of 71.8% in classifying patients.