Article Text
PDF
Abstract
Introduction/Background Institutional transfusion protocols are not universal, and a variety of transfusion policies may exist across participating institutions. As a result, there is no well-defined threshold for intraoperative blood transfusion (BT) in advanced epithelial ovarian cancer (EOC) surgery. According to a recent ESGO consensus guidance, many patients need chemotherapy, thus more liberal transfusion thresholds may be used. We developed a Machine Learning (ML)-prediction that could trigger a BT communication alert based on the extent of surgical cytoreduction.
Methodology We analyzed prospectively collected data from 560 patients with advanced epithelial ovarian cancer (EOC) who underwent cytoreductive surgery at a UK tertiary center between 2014 and 2019. We excluded those with pre-operative anaemia and non-intact anticoagulation system, totaling 464 patients. We employed the eXtreme Gradient Boosting (XGBoost) algorithm to model pre-operative, intra-operative, and human factors. We calculated the estimated blood volume (EBV) using the formula EBV= weight x 80 and set off 10%EBV as threshold for individual intervention. Based on the known estimated blood loss (EBL) we identified two groups. Receiver operating characteristic (ROC) curves were used for performance comparison. We used The SHapley Additive exPlanations (SHAP) framework to explain the predictive model.
Results The model performance for the above threshold prediction was satisfactory (AUC 0.723, 95% CI 0.69–0.75). The top feature commonly shared between both interrogators was operative time (OT) (figure 1). Intra-operative blood loss of at least 20%EBV was associated with OT>130 minutes, primary surgery, peritoneal carcinomatosis index >8, surgical complexity score >3, age of consultant surgeon <48 years, and ascites.
Conclusion Based on the EBV and EBL, we identified a threshold for potential individual intervention, regardless of BT policies. Precise prediction of blood requirements is not possible unless a rough estimate of OT is known in advance.
Disclosures There are no conflict of interests.