Artificial Intelligence In Degenerative Lumbar Spine Management: Machine Learning Models For Data-Driven Clinical Decision-Making

Abstract

Background:Artificial Intelligence (AI) is transforming medicine by enhancing diagnostic accuracy, treatment planning, and patient monitoring through advanced data analysis[1][2] [3]. We conducted a study at Narayana Health, Bangalore, to evaluate the utility of Machine Learning (ML) models—specifically Random Forest (RF), Support Vector Machine (SVM), and K-Nearest Neighbor (KNN)—in predicting and managing degenerative lumbar spine disease. Objectives:The aim of this study was to evaluate the potential of ML models in predicting and determining the appropriate treatment for degenerative lumbar spine disease. Additionally, we sought to explore how these models could assist neurosurgery and spinal surgery residents in the decision-making process during outpatient evaluations. Methods:Using retrospective data from 92 patients with degenerative lumbar spine disease, we identified 83 clinical and radiological parameters critical for preoperative evaluation. The models were trained on this dataset and validated using 20 new prospective cases. Assessment of treatment modalities—conservative management, surgical decompression, or decompression with fusion—were conducted by nine medical professionals, including four neuro-spinal surgeons and five neurosurgery residents. Results:RF and SVM models demonstrated superior accuracy, achieving 87.7%, significantly outperforming the KNN model. These models accurately predicted treatment modalities and outcomes, showing promise as decision-support tools. Conclusions:The RF and SVM models can complement clinical expertise and support training for spinal surgery residents, enhancing their learning in decision-making for degenerative lumbar spine disease. Future work will focus on expanding datasets, refining models, and exploring advanced ML techniques to optimize their application in education and patient care.

Ähnliche Arbeiten