Machine Learning and Neurosurgical Outcome Prediction: A Systematic Review
Joeky T Senders1, Patrick C Staples2, Aditya V Karhade3
1Department of Neurosurgery, University Medical Center Utrecht, Utrecht, The Netherlands; Computational Neurosciences Outcomes Center, Department of Neurosurgery, Brigham and Women's Hospital, Harvard Medical School, Boston, Massachusetts, USA.
World Neurosurgery
|October 8, 2017
Summary
Machine learning (ML) accurately predicts neurosurgical outcomes, outperforming traditional methods. Further research is needed to integrate ML tools into clinical practice for improved patient care.
Area of Science:
- Neurosurgery
- Medical Informatics
- Artificial Intelligence
Background:
- Accurate surgical outcome prediction is crucial for optimizing patient selection and surgical decision-making.
- Machine learning (ML) offers advanced capabilities for data analysis and prediction.
- Identifying patients who will benefit from surgery pre-intervention is a key challenge.
Purpose of the Study:
- To systematically review the potential of machine learning (ML) for predicting neurosurgical outcomes.
- To evaluate the performance of ML algorithms in diverse neurosurgical applications.
Main Methods:
- A systematic literature search was conducted in PubMed and Embase up to January 1, 2017.
- Studies evaluating ML algorithms for predicting survival, recurrence, symptom improvement, and adverse events were included.
Main Results:
- Thirty studies assessed ML for predicting outcomes in various neurosurgical conditions (epilepsy, brain tumors, etc.).
- ML models achieved a median accuracy of 94.5% and an area under the receiver operating curve (AUC) of 0.83.
- ML models significantly outperformed logistic regression, showing median improvements of 15% in accuracy and 0.06 in AUC.
Conclusions:
- Machine learning demonstrates excellent performance in predicting neurosurgical outcomes in research settings.
- ML models showed superior predictive capabilities compared to logistic regression, prognostic indices, and clinical experts.
- Further investigation is required to implement ML as a practical tool in neurosurgical care.
