Artificial Intelligence-Based Prediction of Oroantral Communication after Tooth Extraction Utilizing Preoperative Panoramic Radiography

Abstract

Oroantral communication (OAC) is a common complication after tooth extraction of upper molars. Profound preoperative panoramic radiography analysis might potentially help predict OAC following tooth extraction. In this exploratory study, we evaluated n = 300 consecutive cases (100 OAC and 200 controls) and trained five machine learning algorithms (VGG16, InceptionV3, MobileNetV2, EfficientNet, and ResNet50) to predict OAC versus non-OAC (binary classification task) from the input images. Further, four oral and maxillofacial experts evaluated the respective panoramic radiography and determined performance metrics (accuracy, area under the curve (AUC), precision, recall, F1-score, and receiver operating characteristics curve) of all diagnostic approaches. Cohen’s kappa was used to evaluate the agreement between expert evaluations. The deep learning algorithms reached high specificity (highest specificity 100% for InceptionV3) but low sensitivity (highest sensitivity 42.86% for MobileNetV2). The AUCs from VGG16, InceptionV3, MobileNetV2, EfficientNet, and ResNet50 were 0.53, 0.60, 0.67, 0.51, and 0.56, respectively. Expert 1–4 reached an AUC of 0.550, 0.629, 0.500, and 0.579, respectively. The specificity of the expert evaluations ranged from 51.74% to 95.02%, whereas sensitivity ranged from 14.14% to 59.60%. Cohen’s kappa revealed a poor agreement for the oral and maxillofacial expert evaluations (Cohen’s kappa: 0.1285). Overall, present data indicate that OAC cannot be sufficiently predicted from preoperative panoramic radiography. The false-negative rate, i.e., the rate of positive cases (OAC) missed by the deep learning algorithms, ranged from 57.14% to 95.24%. Surgeons should not solely rely on panoramic radiography when evaluating the probability of OAC occurrence. Clinical testing of OAC is warranted after each upper-molar tooth extraction.

Performance Analysis of Supervised Machine Learning Algorithms for Automatized Radiographical Classification of Maxillary Third Molar Impaction

“…Tobel et al also demonstrated that practitioners with 5 to years of experience were capable of achieving results similar to the AI algorithms [28]. Our previous study on the prediction of OAC in OPGs was able to confirm this observation [16]. Kim et al developed similar protocols to predict the existence of inflamed maxillary sinuses.…”

“…In addition to buccal advancement flap, rotation flaps with buccal cheek fat, local flap plastics with gingival tissue or substitute materials, and various different local rotation flaps, new techniques such as plated rich fibrin, which induces increased wound healing, are increasingly being discussed in the literature [9,25,26]. As we have found in a prior study, we were unable to obtain satisfactory results when using panoramic radiographs to predict OAC [16]. The object of this study was to identify classification systems that were significantly associated with an increased probability of occurrence of OAC.…”

“…Alveolar bone loss, the proximity of the third lower molar to the alveolar nerve, or the prediction of orthodontic extractions could be investigated for similar questions as in the present work [15]. The predictive assessment of OAC based on preoperative panoramic radiography has been investigated in our previous study [16]. Here, we found only limited comparability of AI-based results compared to experienced surgeons.…”

“…Analyzing these classifications in panoramic radiographs with the help of an AI-based algorithm could help automate the prediction of OAC in panoramic radiography imaging. In a previous study of our workgroup, the prediction based solely on panoramic radiographs was found unreliable [16]. To the best of our knowledge, no study has utilized an AI-based algorithm to evaluate the predictive value of these classification systems for automated panoramic radiography analysis.…”

“…Relying on the prediction of an AI-based algorithm alone can thus not be conclusively recommended based on available evidence. Further studies with larger case numbers, other algorithms, and multicenter center studies might be necessary [16]. Several risk factors can be employed to predict the development of OAC.…”