Adolescent Idiopathic Scoliosis Surgery Decision Making with Fuzzy Model

Objective: The surgical decision in scoliosis patients exhibits variability based on angle parameters and the characteristics of patients in the adult or adolescent age group. Existing studies demonstrate that the Cobb's angle, particularly in the range of 25-45, shapes the surgical decision depending on the measures and characteristics of the patients. This study evaluated the performance of a fuzzy logic-based decision support system in making surgical decisions. Methods: A total of 888 patient scenarios were generated in a computer environment, with age, Cobb's, and Risser values. Surgical probability predictions were recorded according to the values in the patient scenarios using rules established by field experts through fuzzy modeling. Results: Although surgical necessity was found in 28.8% of the patients in the reference model, the model detected it at a rate of 11.6%. The sensitivity of the model was 33.9% [95% confidence interval (CI) 27.8-39.7%], specificity 97.3% (95% CI 95.7-98.4%), positive predictive value 83.5% (95% CI 74.9-90.1%), negative predictive value 78.34% (95% CI 75.3-81.2%), accuracy 78.9% (76.1-81.6%), Youden index 0.308, and area under the curve value 0.654. Conclusion: Fuzzy logic is a viable method, particularly in situations where boundaries cannot be clearly determined. Considering variables such as Cobb's and Risser in scoliosis surgery, it could be a method to use in the choice of surgery or conservative follow-up.