MACHINE LEARNING-BASED PREDICTION OF CLINICAL OUTCOMES IN MICROSURGICAL CLIPPING TREATMENTS OF CEREBRAL ANEURYSMS

Abstract

Background

This study investigates the application of Machine Learning (ML) techniques to predict clinical outcomes in microsurgical clipping treatments of cerebral aneurysms. The goal is to enhance healthcare processes through informed clinical decision-making by utilizing a dataset of 344 patients' preoperative characteristics.

Methods

Various ML classifiers were trained to predict outcomes measured by the Glasgow Outcome Scale (GOS). The dataset included 344 patients who underwent microsurgical clipping for intracranial aneurysms. Key features in the dataset included age, aneurysm diameter and neck, arterial hypertension, atherosclerosis, obesity, intubation, vasospasm, and hemorrhage. The study evaluated the models using ROC-AUC scores for outcome prediction and identified key predictors using SHAP analysis.

Results

-    Model Performance: The models achieved ROC-AUC scores of 0.72 ± 0.03 for specific GOS outcome prediction and 0.78 ± 0.02 for binary classification of outcomes.

-    Key Predictors: SHAP analysis identified intubation as the most impactful factor influencing treatment outcome predictions. Other significant predictors included vasospasm and the Hunt and Hess Scale.

-    Patient Demographics: The study included 344 patients with a mean age of 55.05 ± 11.45 years. Gender discrepancy was noted, with 219 females (63.7%).

-    Aneurysm Characteristics: The mean aneurysm neck was 3.93 ± 1.64 mm. Most aneurysms were of the saccular or berry type, with 308 cases (89.5%) being ruptured and 311 cases (90.4%) presenting with hemorrhage.

Conclusions

The study demonstrates the potential of ML in predicting surgical outcomes of cerebral aneurysm treatments. The findings emphasize the need for high-quality datasets and external validation to enhance model accuracy and generalizability. The integration of ML into clinical practice can potentially improve patient management and treatment outcomes in neurosurgery. The study contributes to the ongoing efforts in the medical AI community, supporting the use of ML techniques to enhance clinical decision-making and patient care.