Background: The cost of cancer care is an enormous healthcare burden. Most inpatient chemotherapy (IC) is not reimbursed because of the diagnosis-related group code structure for reimbursement. We implemented the use of a novel objective scoring rubric to guide and automate IC stewardship at an academic cancer center to decrease the inappropriate use of costly IC in patients especially at the end of life. Methods: We created a scoring rubric based on treatment and patient specific criteria. Intravenous (IV) IC that is on formulary and standard of care is automatically approved. IV IC that is non-formulary requires evaluation using developed criteria including: type and phase of trial, FDA and NCCN approvals, performance status, line, and goal of therapy. Clinicians enter these criteria in a Redcap form which automatically calculates a score which is reviewed by 2 disease specific physicians and a clinical pharmacist for accuracy. If the threshold score is not met, IC is not approved for administration. We compared differences from the post-implementation time period (3/2022-10/2022) to the same time the year prior on our primary outcomes: inpatient mortality and drug costs. Analyses were conducted using Chi-square tests. Results: A total of 98 IC requests were submitted; 16% (16) were disapproved. Solid Tumor (ST) accounted for 52% (51) of requests with 48% (41) for Heme Malignancy (HM). Of the approved cases, ST comprised 61% (22) of deaths compared to 39% (14) of patients with HM. Of the disapproved cases that died, 55% (5) were ST and 45% (4) were HM. Requests for checkpoint inhibitors comprised 16% (15) of cases and 81% of those patients died (p= 0.028) with the majority (60%) dying in the hospital (p= 0.05). Among those who died, 26% (n=4; p=0.08) and 46% (n=7; p=0.01) died within 30 and 60 days of last chemotherapy administration, respectively. Pre-implementation the spend on IV NF chemotherapy was $744,105 and post was $549,363; a 26% ($194,741) reduction in cost. Conclusions: Physician choice for administration of IC aligned with objective criteria 84% of the time. Alternatively, this tool prevented inappropriate administration of IC 16% of the time. Checkpoint inhibitor administration was significantly associated with inpatient mortality. Our pilot indicates that there is a role for an objective tool for automated IC stewardship with an emphasis to limit inpatient immunotherapy.