Expert System for Mandibular Condyle Detection and Osteoarthritis Classification in Panoramic Imaging Using R-CNN and CNN

Abstract

Temporomandibular joint osteoarthritis (TMJ OA) is a degenerative condition of the TMJ led by a pathological tissue response of the joint under mechanical loading. It is characterized by the progressive destruction of the internal surfaces of the joint, which can result in debilitating pain and joint noise. Panoramic imaging can be used as a basic screening tool with thorough clinical examination in diagnosing TMJ OA. This paper proposes an algorithm that can extract the condylar region and determine its abnormality by using convolutional neural networks (CNNs) and Faster region-based CNNs (R-CNNs). Panoramic images are collected retrospectively and 1000 images are classified into three categories—normal, abnormal, and unreadable—by a dentist or orofacial pain specialist. Labels indicating whether the condyle is detected and its location enabled more clearly recognizable panoramic images. The uneven proportion of normal to abnormal data is adjusted by duplicating and rotating the images. An R-CNN model and a Visual Geometry Group-16 (VGG16) model are used for learning and condyle discrimination, respectively. To prevent overfitting, the images are rotated ±10° and shifted by 10%. The average precision of condyle detection using an R-CNN at intersection over union (IoU) >0.5 is 99.4% (right side) and 100% (left side). The sensitivity, specificity, and accuracy of the TMJ OA classification algorithm using a CNN are 0.54, 0.94, and 0.84, respectively. The findings demonstrate that classifying panoramic images through CNNs is possible. It is expected that artificial intelligence will be more actively applied to analyze panoramic X-ray images in the future.

Artificial intelligence in detecting temporomandibular joint osteoarthritis on orthopantomogram

“…Based on this result, AI performed best for TMJOA diagnosis on OPG when indeterminate OA was omitted during training and verification. The pilot study showed similar accuracy (0.77-0.84) with various AI models when indeterminate OA was omitted 13 .…”

“…Various studies have applied AI algorithms to read OPG for clinical conditions such as tooth segmentation 6 , age estimation 7 , third molar and inferior alveolar nerve detection 8 , cysts and tumors of the jaw 9 , osteoporosis 10 , and maxillary sinusitis 11 . However, there are few AI studies on TMJOA diagnosis; one study that used CBCT and one study based on OPG have been reported 12 , 13 .…”

Artificial Intelligence in Detecting Temporomandibular Joint Osteoarthritis on Orthopantomogram

“…Based on this result, AI performs best for TMJOA diagnosis on OPG when indeterminate OA was omitted during it trained and veri ed. The pilot study showed similar accuracy (0.77-0.84) with various AI models when indeterminate OA was omitted 13 .…”

“…In this study, AI was designed to extract ROI from OPG by an algorithm based on object detection technique. The average precision of mandibular condyle detection was reported as 99.4% on right side and 100% on left side in previous study 13 . The ROI extracted automatically for the study covers mandibular condyle, articular fossa and eminence, which is very important because TMJOA is not de ned only by bony change of the condyle.…”

“…Various studies have been reported to apply AI algorithms to read OPG for clinical conditions such as tooth segmentation 6 , age estimation 7 , third molar and inferior alveolar nerve detection 8 , cyst and tumors of jaw 9 , osteoporosis 10 , and maxillary sinusitis 11 . However, there is few AI studies on TMJOA diagnosis, and a study using CBCT and a study based on OPG have been reported 12,13 . This study aimed to investigate the clinical usability of an AI diagnostic tool developed for TMJOA diagnosis on OPG in terms of deep learning by comparing its read with that of an expert.…”