ML Prediction for SDS Blood Transfusion.

Abstract

Background: Spinal deformity surgery (SDS) is usually accompanied by significant intraoperative blood loss and transfusion, which is not without risk, as transfusions can lead to transfusion reactions, transmission of infections, and immunosuppression. Therefore, limiting unnecessary intraoperative blood transfusion (IBT) by accurately predicting transfusion requirements is an important goal.

Purpose: Constructing a predictive model for IBT in SDS based on multiple machine learning (ML).

Method: Include patients with spinal deformities who received SDS at 11 large medical centers in China from 2012 to 2022. A total of 162 cases were randomized into a training cohort (70%) and a testing cohort (30%) with the outcome of IBT. A total of 39 candidate factors were collected, including basic personal data, medical comorbidities, surgery-related indicators, and preoperative blood draw indicators, among others. Lasso regression was used to screen potential modeling features. 10 ML algorithms incorporated include Logistic regression (LR), Decision tree, Elastic network, k-Nearest Neighbor (KNN), Neural Networks (NN), Light Gradient Boosting Machine (LightGBM), Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Support Vector Machine (SVM) and Stacking ensemble model. The performance of these models was evaluated using operating characteristic curve (ROC), Precision-Recall, Calibration, and Decision curve analysis (DCA). In addition, SHapley Additive exPlanations (SHAP) was applied to interpret the predictive models. Finally, a web calculator and logistic analysis were created to quantify the hazard level of the features.

Result: By comparing the training group, validation group and multiple parameter comparisons, the RF model had the strongest performance generalization ability (AUC of ROC: 0.8716; AUC of Precision-Recall: 0.8246; BS of Calibration curve: 0.142). Seven key variables were determined including age, BMI, Preoperative hematocrit, Fibrinogen, Prefunction, Bone graft and Number of levels fusion. Finally, Logistics determined that level 4 vertebral fusion surgery may have the greatest IBT risk (OR=20.78, 95% CI 3.9-110.83; P<0.001). A web calculator has also been established for clinical personnel to assess the risk of IBT.

Conclusion: In this study, multiple ML algorithms were successfully established to predict the risk of IBT in SDS, thereby making reasonable use of blood resources and optimizing blood transfusion strategies.