Development of an Artificial Intelligence System to Estimate Postoperative Discomfort After Impacted Third Molar Surgery

Abstract

Background: Artificial Neural Network (ANN) is relatively crude electronic model based on the neural structure of human brain which was used in the field of medicine in different purposes. It can be used for many medical branches especially for estimating the course of a certain disorder or treatment procedure. The aim of this study is to use ANN in maxillofacial surgery to estimate the postoperative symptoms after third molar surgery.Methods:The pre and post-operative information of 175 consecutive patients who needed extraction of impacted third molar teeth were employed to train an ANN. After the training process, the information of 26 cases was used in order to verify the network's ability to predict the post-operative symptoms such as swelling, pain, decrease of mouth opening, bleeding, number of days to return to normal activities and duration of activity restriction. The results obtained from ANN were compared with the results of patients self-reported information. The correlation between the postoperative symptoms of the patients and outcomes obtained from the ANN were analyzed statistically.Results: Close association was found between the patients’ reports and ANN results on post-operative pain, swelling, bleeding, number of days to return to normal activities and duration of activity restriction.Conclusions: The proposed ANN approach is easy to implement and adapted to predict the response of the postoperative outcomes. The model can be further extended to include more variables and experimental data to increase reliability.Keywords:Activity restriction, artificial neural network, postoperative discomfort, third molar surgery.