(PO64) Predicting the Association Between Hypertension on Periapical Disease using Machine Learning

Abstract: Apical periodontitis and hypertension are chronic conditions triggered by inflammatory processes that share similar molecular players. Previous studies have shown a relationship between cardiovascular disease and apical periodontitis; however, few have used machine learning (ML) to predict their association. The purpose of this study was to predict the association between hypertension and periapical disease from the Temple University School of Dentistry Electronic Dental Record using an ML algorithm. Only patients who required primary endodontic therapy from January 2018 to December 2022 were included. Periapical diagnosis (Y/N) was defined as having either symptomatic apical periodontitis, asymptomatic apical periodontitis, chronic apical abscess, or acute apical abscess. Extreme gradient boosting (XGBoost) was used to predict periapical disease under 5-fold cross-validation. Model performance was evaluated using the Brier score, accuracy, and area under the curve (AUC) with the MachineShop Package in R. A total of 3,888 patients over the age of 18 were eligible for this study, of which 1,511 were diagnosed with periapical disease. About 16% of the patients had hypertension. The prediction performance of XGBoost showed a Brier score of 0.124, an accuracy of 83.7%, and an AUC of 88.6%. Out of 13 variables assessed, six were found to be important predictors of periapical disease diagnosis. They are (in descending order): age, excessive alcohol use, hypertension, asthma, diabetes, and tobacco use. Our machine learning model is effective in predicting periapical disease using clinical exposure variables. Hypertension was one of the most significant health history factors found to predict periapical disease within the Temple dental school population.