The TOPHAN Score for Predicting Postoperative Intensive Care Unit Admission in Gynecologic Oncology Patients: A Cohort Study

Background: Surgical procedures are the gold standard for managing gynecologic cancer, with intensive care playing a crucial postoperative role. However, intensive care unit (ICU) availability in low- and middle-income countries is severely limited. This study aimed to develop a reliable preoperative scoring system to predict postoperative ICU needs in gynecologic oncologycases. Materials and Methods: Model development was conducted using a retrospective cohort from January 2013 to December 2019. Data from gynecologic oncology patients included clinical characteristics, lab values, intraoperative bleeding, and surgical duration. Variables with P < 0.157 in univariate analysis were entered into multivariate logistic regression. Internal validation used the bootstrap method. Discrimination and calibration were assessed using the area under the curve (AUC) and the Hosmer–Lemeshow test. Temporal validation was performed using a prospective cohort (June 2024–July 2025) and assessed using AUC, sensitivity, specificity, and calibration. Results: Among 843 surgeries, 380 (45.07%) required ICU admission. Significant predictors included age, hemoglobin, American Society of Anesthesiologists status, surgical access, estimated blood loss, duration, and tumor site. These variables consequently construct the proposed TOPHAN (T = Time of surgery duration; O = Organ site; P = Procedure access; H = Hemoglobin; A = Age/ASA; and N = depletioN of blood volume.) score. A 26.5 cutoff showed 75.8% sensitivity and 64.6% specificity. Temporal validation showed that a 20.5 cutoff improved alignment (76.5% sensitivity and 73.8% specificity). Conclusion: TOPHAN score represents a preliminary preoperative tool for predicting postoperative ICU admission in gynecologic oncology surgery. Temporal validation suggests that scores above 20.5 are associated with a higher likelihood of ICU requirement. Furthermore, multicenter studies are warranted to enhance its generalizability and refine its clinical application.