Blood transfusion (BT) remains an integral part of oncosurgical management but is often associated with adverse outcomes and resource overuse. Predicting transfusion requirements before surgery could enhance patient blood management and optimize inventory utilization. This study aimed to identify perioperative predictors of BT and to develop a transfusion risk prediction tool for oncosurgical patients in a tertiary cancer care setting.

A retrospective analysis was conducted on 578 patients who underwent major oncologic surgeries between December 2020 and April 2021. Demographic, surgical, and hematological parameters were compared between transfused and non-transfused groups. Factors associated with BT were analyzed using univariate and multivariate logistic regression. A transfusion risk prediction score model was constructed from independent predictors, and its discriminatory performance was assessed using the receiver operating characteristic (ROC) curve.

Among 578 patients, 73 (13%) required perioperative BT. Multivariate analysis identified preoperative hemoglobin (OR = 0.6; 95% CI: 0.52–0.71; p < 0.001) and duration of surgery (4–8 h: OR = 2.36; p = 0.03; > 8 h: OR = 19.16; p = 0.003) as independent predictors of transfusion. Type of surgery (deep or laparoscopic) showed significance on univariate but not on multivariate analysis. The model exhibited good discriminatory ability (AUC = 0.789; 95% CI: 0.762–0.860; p < 0.001). Patients with a cumulative score ≤ 8 had a higher probability of requiring BT.

Preoperative hemoglobin concentration and surgical duration are independent determinants of perioperative transfusion in oncosurgical patients. The proposed Transfusion Risk Prediction Score provides a simple, validated, and evidence-based tool to promote rational blood utilization and improve perioperative planning in cancer surgery.