Predictive modeling for step II therapy response in periodontitis - model development and validation.

Abstract

Steps I and II periodontal therapy is the first-line treatment for periodontal disease, but has varying success. This study aimed to develop machine learning models to predict changes in periodontal probing depth (PPD) after step II therapy using patient-, tooth-, and site-specific clinical covariates. Models accurately predicted that healthy sites stay healthy, but performed suboptimally for diseased sites. Tuning improved performance, with PPD, tooth-site, and tooth-type identified as key predictors. Pocket closure was predicted with fair accuracy, with baseline PPD as the most relevant covariate. Models predicted improving pockets well but underperformed for non-responding sites, with antibiotic treatment and tooth type being the most influential features. While predictive performance for step II periodontal therapy based on routine clinical data remains limited, models can stratify periodontal sites into meaningful categories and estimate the probability of pocket improvement. They provide a foundation for site-specific outcome prediction and may support patient communication and expectations.