Abstract
Context The aim of the study was to develop a prognostic model using artificial intelligence
for patients undergoing lumbar spine surgery for degenerative spine disease for change
in pain, functional status, and patient satisfaction based on preoperative variables
included in following categories—sociodemographic, clinical, and radiological.
Methods and Materials A prospective cohort of 180 patients with lumbar degenerative spine disease was included
and divided into three classes of management—conservative, decompressive surgery,
and decompression with fixation. Preoperative variables, change in outcome measures
(visual analog scale—VAS, Modified Oswestry Disability Index—MODI, and Neurogenic
Claudication Outcome Score—NCOS), and type of management were assessed using Machine
Learning models. These were used for creating a predictive tool for deciding the type
of management that a patient should undergo to achieve the best results. Multivariate
logistic regression was also used to identify prognostic factors of significance.
Results The area under the curve (AUC) was calculated from the receiver-operating characteristic
(ROC) analysis to evaluate the discrimination capability of various machine learning
models. Random Forest Classifier gave the best ROC-AUC score in all three classes
(0.863 for VAS, 0.831 for MODI, and 0.869 for NCOS), and the macroaverage AUC score
was found to be 0.842 suggesting moderate discriminatory power. A graphical user interface
(GUI) tool was built using the machine learning algorithm thus defined to take input
details of patients and predict change in outcome measures.
Conclusion This study demonstrates that machine learning can be used as a tool to help tailor
the decision-making process for a patient to achieve the best outcome. The GUI tool
helps to incorporate the study results into active decision-making.
Keywords
neurosurgery - arificial intelligence - degenerative lumbar spine disease - machine
learning - Random Forest Classifier - lumbar canal stenosis
