Selection of Patients for Early Intervention Following Blunt Splenic Injury using Machine Learning Models
Non-operative management, including intensive monitoring and splenic artery embolisation (SAE), is the standard of care for patients with blunt splenic injury who are haemodynamically stable. However, there is currently no consensus on which of these patients would benefit from SAE.
We approached this problem using machine learning modelling. Using data within a single Major Trauma Center across 7 years (2015-2022), we developed and refined various models using a training cohort (n=126) and a validation cohort (n=48).
The model presented in the MxBSI application has a high sensitivity (1.0), specificity (0.816), PPV (0.588), NPV (1.0), balance accuracy (0.908) and AUC (0.932) when tested on the validation cohort.
As far as we are aware, this is the first study to pilot machine learning modelling to address this clinical problem.