Somatosensory evoked potentials in patients with high-grade aneurysmal subarachnoid hemorrhage

Explainable machine learning in outcome prediction of high-grade aneurysmal subarachnoid hemorrhage

OBJECTIVE

Accurate prognostic prediction in patients with high-grade aneruysmal subarachnoid hemorrhage (aSAH) is essential for personalized treatment. In this study, we developed an interpretable prognostic machine learning model for high-grade aSAH patients using SHapley Additive exPlanations (SHAP).

METHODS

A prospective registry cohort of high-grade aSAH patients was collected in one single-center hospital. The endpoint in our study is a 12-month follow-up outcome. The dataset was divided into training and validation sets in a 7:3 ratio. Machine learning algorithms, including Logistic regression model (LR), support vector machine (SVM), random forest (RF), and extreme gradient boosting (XGBoost), were employed to develop a prognostic prediction model for high-grade aSAH. The optimal model was selected for SHAP analysis.

RESULTS

Among the 421 patients, 204 (48.5%) exhibited poor prognosis. The RF model demonstrated superior performance compared to LR (AUC = 0.850, 95% CI: 0.783-0.918), SVM (AUC = 0.862, 95% CI: 0.799-0.926), and XGBoost (AUC = 0.850, 95% CI: 0.783-0.917) with an AUC of 0.867 (95% CI: 0.806-0 .929). Primary prognostic features identified through SHAP analysis included higher World Federation of Neurosurgical Societies (WFNS) grade, higher modified Fisher score (mFS) and advanced age, were found to be associated with 12-month unfavorable outcome, while the treatment of coiling embolization for aSAH drove the prediction towards favorable prognosis. Additionally, the SHAP force plot visualized individual prognosis predictions.

CONCLUSIONS

This study demonstrated the potential of machine learning techniques in prognostic prediction for high-grade aSAH patients. The features identified through SHAP analysis enhance model interpretability and provide guidance for clinical decision-making.

The Evoked Potential Score for SSEP and BAEP-A Prognostic Marker for the Long-Term Neurological Outcome in Patients after Poor-Grade Subarachnoid Hemorrhage

Objective: Evoked potentials are widely used in comatose patients to evaluate neurological function; however, prognostic relevance in patients after SAH is barely investigated. Therefore, we aimed to investigate the prognostic value of the proposed Evoked Potential Score (EPS) for somatosensory (SSEP) and brainstem auditory evoked potentials (BAEP) on the neurological outcome in patients after poor-grade SAH. Methods: We retrospectively analyzed patients after poor grade SAH (Hunt and Hess (HH) grade IV and V) that were admitted to the ICU at the Department of Neurosurgery, MUV, between 2014 and 2017. Measurements of SSEP and BAEP were evaluated separately as well as in a combined model, using the EPS at admission and before ventilator weaning and correlated with the grade of the modified ranking scale at the last available follow up. Results: In total, 48 patients after SAH HH IV/V were included in this study. The EPS for SSEP at admission (p = 0.007) and both the EPS for SSEP (p = <0.0001) and BAEP (p = 0.036) before ventilator weaning were significant prognostic markers for neurological improvement at a mean follow-up period of 14.1 months. In addition, the combined model of the EPS for SSEP/BAEP performed as a prognostic marker for neurological improvement (“at admission” p = 0.007; “before ventilator weaning” p < 0.001). Conclusions: In the first series to date we found a high prognostic significance for the EPS as a combined model, as well as a separate analysis for SSEP and BAEP in patients after SAH IV and V. In the future, these findings potentially support physicians in ethically challenging decision-making processes and in advice for patients’ families under consideration of an individual evaluation of each patient.