Propuesta de un modelo machine learning para predecir la severidad de la reabsorción radicular inducida por ortodoncia

Abstract

Root resorption (RR) can be considered a common iatrogenic consequence of orthodontic treatment observed by orthodontists during treatment and its diagnosis is mainly radiographic. The aim of this study is to develop a model that allows predicting the severity of RR that a patient could present considering diagnostic and treatment variables. This will allow the orthodontist to anticipate the patient's willingness to develop RR at the beginning of their treatment, in order to promote clinical decision-making that allows maintaining the health of dental tissues. Methods: 191 records are taken from a study conducted by Silva et al. (2018), the respective labeling is carried out for the classification of the severity of the resorption (OIEARRmax: mild 0-15%, moderate/severe > 15%). A base model and four supervised learning models were trained and evaluated.