Prediction of Surgical Satisfaction in Patients with Lumbar Spinal Canal Stenosis Using Artificial Neural Networks

Abstract

Background: The purpose of this study was to develop an artificial neural networks (ANNs) model for predicting 2-year surgical satisfaction and compared with traditional predictive tool in LSCS patients.

Materials and Methods: The two prediction models included an ANN model and a logistic regression (LR) model. The age, gender, duration of symptoms, walking distance, visual analog scale (VAS) of leg pain/numbness, the Japanese Orthopaedic Association (JOA) Score, the Neurogenic Claudication Outcome Score (NCOS) and the Stenosis Ratio (SR) values have been determined as the input variables for the developed ANNs and LR model. Patient surgical satisfaction was recorded by using standardized measure. ANNs were fed patient data in order to predict 2-year surgical satisfaction based on several input variables. Sensitivity analysis to the developed ANN model was applied to identify the important variables. The area under a receiver operating characteristic (ROC) curve (AUC), Hosmer-Lemeshow (H-L) statistics and accuracy rate were calculated for evaluating the two models.

Results: A total of 168 (59 male, 109 female, mean age 59.8±11.6 years) patients were divided into training (n = 84), testing (n = 42), and validation (n = 42) data sets. Post-surgical satisfaction was 88.7% at 2-year follow-up. The SR was important variable selected by the ANN. The ANN model displayed better accuracy rate in 96.9% of patients, a better H-L statistic in 42.4% of patients, and a better AUC in 80.0% of patients, compared to the LR model.

Conclusion: The findings show that an ANNs can predict 2-year surgical satisfaction for use in clinical application and more accurate compared to LR model.