Panoramic Radiographic Images Research Articles

Artificial Intelligence in Fractured Dental Implant Detection and Classification: Evaluation Using Dataset from Two Dental Hospitals.

Fracture of a dental implant (DI) is a rare mechanical complication that is a critical cause of DI failure and explantation. The purpose of this study was to evaluate the reliability and validity of a three different deep convolutional neural network (DCNN) architectures (VGGNet-19, GoogLeNet Inception-v3, and automated DCNN) for the detection and classification of fractured DI using panoramic and periapical radiographic images. A total of 21,398 DIs were reviewed at two dental hospitals, and 251 intact and 194 fractured DI radiographic images were identified and included as the dataset in this study. All three DCNN architectures achieved a fractured DI detection and classification accuracy of over 0.80 AUC. In particular, automated DCNN architecture using periapical images showed the highest and most reliable detection (AUC = 0.984, 95% CI = 0.900–1.000) and classification (AUC = 0.869, 95% CI = 0.778–0.929) accuracy performance compared to fine-tuned and pre-trained VGGNet-19 and GoogLeNet Inception-v3 architectures. The three DCNN architectures showed acceptable accuracy in the detection and classification of fractured DIs, with the best accuracy performance achieved by the automated DCNN architecture using only periapical images.

The Age- and Gender-wise Assessment of the Position and Shape of Mental Foramen in Patients’ Panoramic Radiography

Background and Aim: Recognizing the position of the Mental Foramen (MF) is essential in numerous cases, such as anesthesia injection and periapical surgeries in the anterior region of the mandible. Furthermore, the diversity in the location and position of MF can develop problems during surgery in this region. Methods & Materials: The present study examined anatomical landmarks based on panoramic radiographic images obtained in Bandar Abbas City, Iran. In total, 450 panoramic radiographic images of men and women were assessed. All explored images were converted to JPEG format and entered in real size in Auto CAD software (2014). The collected data were analyzed in SPSS by one-way Analysis of Variance (ANOVA), Student’s t-test, and Chi-squared test. Ethical Considerations: This study was approved by the Ethics Committee of Hormozgan University of Medical Sciences (Code: IR.HUMS.REC.1394.189). Results: The obtained data revealed that the mean distance from the MF to the lower edge of the mandible on the right and left was 10.53 mm and 10.51 mm, respectively. The mean distance from the MF to the posterior side of the mandible equaled 49.36 mm on the right and 48.72 mm on the left. Moreover, the mean distance of MF to the midline of the lower jaw on the right and left was calculated as 27.16 and 26.27 mm, respectively. Furthermore, in most cases, the anterior-posterior position of MF was symmetrical concerning anatomical landmarks. The anterior-posterior distance of MF to mandibular midline was significantly different between the explored males and females (P<0.001). There was no significant gender-wise difference in the shape of the MF (P=0.89). Conclusion: The present research results signified that the most frequent anterior-posterior position of the MF is between the apex of the first and second premolars. Furthermore, the most prevalent shape of MF is oval per panoramic images.

Deep learning based prediction of extraction difficulty for mandibular third molars

This paper proposes a convolutional neural network (CNN)-based deep learning model for predicting the difficulty of extracting a mandibular third molar using a panoramic radiographic image. The applied dataset includes a total of 1053 mandibular third molars from 600 preoperative panoramic radiographic images. The extraction difficulty was evaluated based on the consensus of three human observers using the Pederson difficulty score (PDS). The classification model used a ResNet-34 pretrained on the ImageNet dataset. The correlation between the PDS values determined by the proposed model and those measured by the experts was calculated. The prediction accuracies for C1 (depth), C2 (ramal relationship), and C3 (angulation) were 78.91%, 82.03%, and 90.23%, respectively. The results confirm that the proposed CNN-based deep learning model could be used to predict the difficulty of extracting a mandibular third molar using a panoramic radiographic image.

Sexual dimorphism of tooth agenesis patterns in non-syndromic oligodontia in Japanese patients

ObjectiveWide individual variation observed in tooth agenesis patterns in patients with non-syndromic oligodontia could be affected by sexual dimorphism. Therefore, the present study investigated sex-associated differences in tooth agenesis patterns in patients with non-syndromic oligodontia. DesignSubjects were 108 male (ages 7–46 years) and 184 female (ages 7–43 years) orthodontic patients missing ≥6 permanent teeth, excluding third molars. The number of agenetic teeth per patient, excluding third molars, and the prevalence of agenesis of each tooth type were evaluated using panoramic radiographic images. Rankings of the most common tooth agenesis patterns in the maxillary and mandibular arches were calculated. The number of agenetic teeth per patient was compared between sexes using the Mann–Whitney test. The prevalence of tooth agenesis for each tooth type was compared between sexes using the χ2 and Fisher's exact tests with Bonferroni correction. ResultsNo significant sex difference in the distribution of the number of agenetic teeth was observed. The prevalence of agenesis of the maxillary second premolars was significantly higher in females (P < 0.007) and the prevalence of agenesis of the mandibular central incisors was significantly higher in males (P < 0.007). Sexual dimorphism was observed in the highly ranked tooth agenesis patterns. ConclusionsNo significant difference in the severity of tooth agenesis was observed between sexes; however, sexual dimorphism was observed in the tooth agenesis patterns of patients with non-syndromic oligodontia. Although various tooth agenesis patterns were identified, sex-specific tendencies suggest that categories can be utilized in future epidemiologic research and planning.

Condylar and ramal vertical asymmetry in patients with different vertical skeletal patterns

Abstract Objective: The objective of the present study was to evaluate condylar, ramal and condylar+ramal mandibular vertical asymmetry in a group of patients with different vertical skeletal patterns. Methods: Mandibular condylar and ramal measurements of 104 patients were performed on panoramic radiographic images and asymmetry indices were obtained according to the Habets’ formula. The study groups consisted of 35 hypodivergent, 34 normodivergent and 35 hyperdivergent patients. Data were statistically analysed by means of two-way ANOVA, Kruskal–Wallis and Pearson chi square tests. Results: Ramal height was significantly higher in the hypodivergent group (p &lt; 0.05). No significant differences were found between the vertical skeletal pattern groups in relation to the asymmetry indices (p &gt; 0.05). Conclusion: Condylar asymmetry values were found to be higher than a 3% threshold value in all groups, but no significant differences were observed between the groups. The effect of the vertical skeletal pattern on vertical mandibular asymmetry was found to be insignificant.

Axiograph and the Panoramic Radiographic Imaging Technique, an Alternative to Conventional Technique to Record Condylar Guidance

Introduction: The documentation of the maxillomandibular relationship and the recording values of condylar guidance is the most critical, yet essential step in fabrication of complete denture prosthesis and it also helps in diagnosing, treatment planning, and fabrication of the balanced complete denture prosthesis that aids in enhancing the aesthetics, and maintenance of the stomatognathic system. Various techniques are available to record the condylar guidance; One of them being axiographic technique and panoramic radiograph followed by programming the semi adjustable articulators. Aim: The aim of the present study was to evaluate and compare the condylar guidance using the conventional technique the gothic arch tracing with that of axiographic technique and the panoramic radiographs. Materials and Methods: The observational study with the clinical set up was carried out over a period between October 2018- March 2019. Twenty completely edentulous patients were selected for the study. The comparative evaluation of axiographic tracing technique, panoramic radiographic technique and the extraoral gothic arch tracing technique was conducted. The horizontal condylar guidance values obtained with these techniques were assessed. The data obtained was statistically analysed using Karl Pearson correlation test to determine the correlation between the values obtained from all the three techniques on the right side and the left side of the subjects. Paired t-test and Bonferroni Post-Hoc test were used to compare the extraoral gothic arch tracing values and Orthopantomogram values and axiographic tracing values on right side and left side of the subjects. Results: A statistically significant correlation was observed between the extraoral Gothic arch tracing technique and the panoramic radiographic technique on right (r=0.047, p-value 0.04) and left sides (r=0.55, p-value 0.01) of the subjects. The axiographic tracing showed lower values when compared with gothic arch tracing technique and the panoramic radiographic technique. (95%CI of difference was -0.49 lower). Significant differences were obtained on comparison between all three techniques on right (p-value&lt;0.001) and left sides.(p-value 0.008). On pairwise comparison of right sides between Extraoral gothic arch Tracing and Axiographic tracing (p-value&lt;0.001) and between axiographic tracing and Orthopantomographic tracing (p-value&lt;0.001), were significant, while pairwise comparison on left side between Orthopantomogram and Axiographic tracings (p-value 0.01) was significant. Conclusion: The panoramic radiographic images can be used as an alternative method to that of conventional gothic arch tracing technique to obtain the condylar guidance values. The values obtained using axiographic tracings method was comparatively lower. Thus, the values obtained from axiographic technique should be used with caution or the values needs to be compared to gold standard techniques to further analyse the accuracy of the values that can be used as a valuable aid in setting the condylar guidance in semi-adjustable articulators.

Dental and maxillomandibular incidental findings in panoramic radiography among individuals with mucopolysaccharidosis: a cross-sectional study.

Mucopolysaccharidosis (MPS) is a group of rare and inherited metabolic disorders caused by the accumulation of macromolecule glycosaminoglycans inside lysosomes. Affected individuals may have dental and craniofacial tissue alterations, facilitating the development of several oral diseases. Objectives To assess, with panoramic radiographic images, the frequency of dental and maxillomandibular incidental findings among MPS individuals and compare them with non-MPS individuals.Methodology A cross-sectional study evaluating a sample of 14 MPS individuals and 28 non-MPS individuals aged from 5 to 26 years was carried out. They were matched for sex and age on a 2:1 proportion. Panoramic radiographs were assessed for the presence/absence of the following dental and maxillomandibular alterations: dental anomalies of number (hypodontia/dental agenesis, supernumerary teeth); anomalies of form (microdontia, macrodontia, conoid teeth, taurodontism, and root dilaceration); anomalies of position (impacted tooth, inverted tooth, tooth migration, partially bony teeth, complete bony teeth); periapical alterations (furcation lesion, circumscribed bone rarefaction); other alterations (radiolucent bone lesions, radiopaque bone lesions, radiopacity in the maxillary sinus, condylar hypoplasia). Differences between groups were tested by the Fisher’s exact test and chi-square test (p<0.05).Results For intrarater agreement, Kappa values were 0.76 to 0.85. The presence of supernumerary teeth (p=0.003); conoid teeth (p=0.009); taurodontism (p<0.001); impacted teeth (p<0.001); partial bony teeth (p=0.040); complete bony teeth (p=0.013); and root dilaceration (p=0.047) were statistically more frequent in MPS individuals compared to non-MPS individuals. Bone rarefaction/furcation lesions (p=0.032), condylar hypoplasia (p<0.001), radiolucent bone lesions (p=0.001), and dentigerous cysts (p=0.002) were also more frequent in MPS individuals.Conclusion The presence of specific oral manifestations is more common in MPS individuals than non-MPS individuals.

Biometric identification using panoramic dental radiographic images with few-shot learning

Biometric identification using panoramic dental radiographic images with few-shot learning

Agreement between Panoramic Radiography and CBCT in Interpretation and Diagnosis of Jaw Lesions

Objective: this study aims to assess the rate of agreement between panoramic radiography (PR) and CBCT in evaluating the radiographic features of intraosseous lesions and detecting whether these differences had a significant impact on the accuracy of differential diagnoses. Methods: Twenty four (24) intraosseous biopsy-proven lesions were reviewed using PR and CBCT images by two oral and maxillofacial radiologists. Both observers answered eighteen questions for each case describing the lesions` radiographic criteria and listed two possible differential diagnoses for each case. Results: McNemar’s test was used for comparing the two modalities and for calculating the inter-observer agreement. The intra-observer agreement showed that questions 18 (cortical destruction) and 10 (expansion in the anatomical landmarks) had the lowest percentage of agreement between both observers. Furthermore, there was no statistically significant difference between the two modalities in terms of the agreement between differential diagnosis and final diagnosis (P > 0.05). Conclusions: Both PR and CBCT provide similar agreement in describing the lesion location, epicenter, internal structure. Moreover, similar agreement was also observed regarding the effect of the lesion on the teeth including resorption, impaction and displacement, root resorption, lamina dura, and periodontal membrane space of the teeth. The main disagreement between the two modalities was found in the expansion of the normal surrounding anatomic boundaries, and cortical destruction.

Comparative evaluation of condylar guidance obtained by three different interocclusal recording materials in a semi-adjustable articulator and digital panoramic radiographic images in dentate patients: An in vivo study.

Aim:The Aim of the study was to evaluate the reproducibility of condylar inclination values obtained using three different interocclusal recording materials in Hanau Wide-Vue semi-adjustable articulator and to compare those values of condylar inclination with panoramic radiographic images.Settings and Design:In Vivo Comparative study.Materials and Methods:The clinical trial was carried out in the Department of Prosthodontics, M. R Ambedkar Dental College and Hospital, Bengaluru. A total of 30 dentulous subjects aged between 20 and 40 years were selected, without any signs and symptoms of temporomandibular joint disorders. With the help of a custom-made protrusive guide, protrusive records for three materials were made. These records were used for programming the articulator and the respective horizontal condylar inclination (HCI) value, and panoramic radiographs were made. The values obtained were statistically analyzed.Statistical Analysis Used:Independent t-test, multiple comparisons using Bonferroni method followed by Analysis of variances (ANOVA) were used.Results:Among each pair of groups there exists a significant difference, i.e., the mean condylar inclination (degree) was found to be statistically significant between wax and polyvinyl siloxane (P < 0.01) as well as between wax and polyether (P < 0.01). No significant difference was observed between polyvinyl siloxane and polyether (P > 0.05).Conclusion:HCI values differed depending on the type of material employed. The study concluded that the protrusive condylar guidance angles obtained by panoramic radiograph can be used in programming the semi-adjustable articulator and among the materials polyether gave the highest Condylar Guidance values with less variability and wax gave lowest values with greater variation.

Tooth Numbering and Condition Recognition on Dental Panoramic Radiograph Images Using CNNs

Tooth Numbering and Condition Recognition on Dental Panoramic Radiograph Images Using CNNs

SUCCESS OF ARTIFICIAL INTELLIGENCE SYSTEM IN DETERMINING ALVEOLAR BONE LOSS FROM DENTAL PANORAMIC RADIOGRAPHY IMAGES

Objectives: The aim of this study was to detect alveolar bone loss from dental panoramic radiographic images using artificial intelligence systems. Material and Methods: A total of 2276 panoramic radiographic images were used in this study. While 1137 of them belong to cases with bone destruction, 1139 were periodontally healthy. The dataset is divided into three parts as training (n=1856) , validation (n=210) and testing set (n= 210). All images in the data set were resized to 1472x718 pixels before training. A random sequence was created using the open-source python programming language and OpenCV, NumPy, Pandas, and Matplotlib libraries effectively. A pre-trained Google Net Inception v3 CNN network was used for preprocessing and data sets were trained using transfer learning. Diagnostic performance was evaluated with the confusion matrix using sensivitiy, specificity, precision, accuracy and F1 score. Results: Of the 105 cases with bone loss, 99 were detected by the AI system. Sensitivity was 0.94, specificity 0.88, precision 0.89, accuracy 0.91 and F1 score 0.91. Conclusion: The convolutional neural network model is successful in determining periodontal bone losses. It can be used as a system to facilitate the work of physicians in diagnosis and treatment planning in the future.

Evaluation of Clinical Application Potential based on a Deep Learning Technique with Real-Size Dental Panoramic Radiography: A Preliminary Study

With the recent advancement of artificial intelligence (AI), data-based research is being actively conducted in the dental medical field. However, there is a limited amoun of research yet based on algorithms using panoramic radiography. This study was conducted to find the standard AI reading that distinguishes the young from the elderly using panoramic radiographic images, and to confirm the applicability of the method as a means of increasing the reliability of a diagnosis. A total of 117 panoramas in A dental clinic were used. The selected radiographic images were classified into two groups: the old group and the young group. To load the classified images into the suggested and designed multi-layer neural network model (modified DarkNet), they were split into 70 % training data and 30 % testing data using the ‘SplitEachLable()’ Matlab function. To identify the old group, the focal class activation mapping or CAM (the height of the alveolar bone and the major places where other treatment actions took place) area was estimated. To identify the young group, a wide CAM area over the entire area was estimated as a feature. These data could be important quantitative indicators of the health of the alveolar bone and of the overall dental condition. Significant results and features were derived to show the potential of quantitative indicators for dental care. The results of this study confirmed the possibility of estimating the alveolar bone age based on AI.

Automatic diagnosis of dental diseases using convolutional neural network and panoramic radiographic images

ABSTRACT People may experience tooth decay many times in a lifetime; however, its early diagnosis can greatly help accelerate the treatment process. Intelligent methods are now widely used in medicine, especially for diagnosis through panoramic radiographs. However, dental diseases have received little research focus due to the lack of a standard dataset. This study aimed to collect a standard dataset of panoramic radiographs of jaws and teeth. A deep neural network algorithm was then employed to classify teeth as healthy, decayed, root-canaled, and restored categories. The dataset was divided into five groups to analyse the proposed method in performance. Different evaluation criteria, confusion matrices, and precision percentages were determined for each group. On average, the final precision was 92% in the five groups. Moreover, the trained network results were compared with those of AlexNet and VGGNet16. Six months were spent collecting two new datasets labelled by experts. A smart method was also proposed to determine the dental location of jaws and reach dental diagnosis. According to the final results, the proposed network proved to be highly stable and managed to establish a better dental diagnosis.

A Performance Comparison between Automated Deep Learning and Dental Professionals in Classification of Dental Implant Systems from Dental Imaging: A Multi-Center Study.

In this study, the efficacy of the automated deep convolutional neural network (DCNN) was evaluated for the classification of dental implant systems (DISs) and the accuracy of the performance was compared against that of dental professionals using dental radiographic images collected from three dental hospitals. A total of 11,980 panoramic and periapical radiographic images with six different types of DISs were divided into training (n = 9584) and testing (n = 2396) datasets. To compare the accuracy of the trained automated DCNN with dental professionals (including six board-certified periodontists, eight periodontology residents, and 11 residents not specialized in periodontology), 180 images were randomly selected from the test dataset. The accuracy of the automated DCNN based on the AUC, Youden index, sensitivity, and specificity, were 0.954, 0.808, 0.955, and 0.853, respectively. The automated DCNN outperformed most of the participating dental professionals, including board-certified periodontists, periodontal residents, and residents not specialized in periodontology. The automated DCNN was highly effective in classifying similar shapes of different types of DISs based on dental radiographic images. Further studies are necessary to determine the efficacy and feasibility of applying an automated DCNN in clinical practice.

Identification of dental implants using deep learning\u2014pilot study

BackgroundIn some cases, a dentist cannot solve the difficulties a patient has with an implant because the implant system is unknown. Therefore, there is a need for a system for identifying the implant system of a patient from limited data that does not depend on the dentist’s knowledge and experience. The purpose of this study was to identify dental implant systems using a deep learning method.MethodsA dataset of 1282 panoramic radiograph images with implants were used for deep learning. An object detection algorithm (Yolov3) was used to identify the six implant systems by three manufactures. To implement the algorithm, TensorFlow and Keras deep-learning libraries were used. After training was complete, the true positive (TP) ratio and average precision (AP) of each implant system as well as the mean AP (mAP), and mean intersection over union (mIoU) were calculated to evaluate the performance of the model.ResultsThe number of each implant system varied from 240 to 1919. The TP ratio and AP of each implant system varied from 0.50 to 0.82 and from 0.51 to 0.85, respectively. The mAP and mIoU of this model were 0.71 and 0.72, respectively.ConclusionsThe results of this study suggest that implants can be identified from panoramic radiographic images using deep learning-based object detection. This identification system could help dentists as well as patients suffering from implant problems. However, more images of other implant systems will be necessary to increase the learning performance to apply this system in clinical practice.

Evaluation of the frequency of third molar agenesis according to different age groups.

Third molars (M3s) are the most common congenitally missing teeth. The aim of this study was to compare the frequency of M3 agenesis in two different age groups. This study examined the panoramic radiographic images of 1036 patients. Two groups, each consisting of 518 patients, were arranged based on age ranges. The patients aged 12 to 19 years were in the first group and those 20 years and older were in the second group. The frequency of M3 agenesis was examined in both age groups. The distribution of M3 agenesis by sex and jaw was also recorded, along with the number of M3s affected by agenesis. Chi-square tests were used to examine the relationships between categorical variables. M3 agenesis was observed in 29.3% and 20.5% in the first and second groups, respectively; the difference was statistically significant (P < .01). It was more common in female (27.3%) than in male patients (21.4%), and in the maxilla (11.2%) than in the mandible (5.8%), and the differences were statistically significant (P < .05). It was mostly observed forthemaxilla right M3 (18) (30.9%) and in the single quadrant of the jaws (9.7%). The frequency of M3 agenesis was more common in patients aged 12 to 19 years than in those aged 20 years and older. Additionally, M3 agenesis was more common in females and in the maxilla.

Comparison of radiograph panoramic quality between conventional and indirect digital methods on mixed teeth period

Introduction: Panoramic radiograph is an image used by the dentist to maintain diagnosis in dentistry, for example, to evaluate teeth condition in mixed dentition period. There are two panoramic radiography techniques, which are conventional and digital (direct and indirect) method. This study was aimed to analyse the differences of panoramic radiograph image quality between the conventional and digital indirect method on mixed dentition radiograph. Methods: Thirty samples of conventional and indirect digital panoramic radiograph were randomly taken from the database according to inclusion criteria. The image quality of conventional radiograph was manually assessed by using illuminator box, while indirect digital radiographs was assessed directly using DBS Win 5.7.0 software (Durr Dental, Germany). The radiographs were divided into six zones: teeth area, nasal and sinus, mandible, TMJ, ramus-spine, and hyoid bone. The image quality assessment was performed in each area based on three criteria: anatomical coverage, density and contrast, and anatomical structure. Results: The Cronbach's Alpha Intra-Class Correlation (ICC) inter and intraobserver agreement test of conventional and digital indirect methods on the aspect of anatomical, density and contrast coverage, and anatomic structure showed an excellent agreement. The independent sample t-test result showed significant difference (p &lt; 0.05) between the conventional and indirect digital method in density and contrast as well as anatomical structure aspects, but not significantly different (p &gt; 0.05) in the anatomical coverage aspect. Conclusion: Indirect digital panoramic radiograph shows better image quality as compared to conventional radiograph, especially in the anatomical structure aspect, density and contrast. There is no difference in the anatomical coverage aspect between conventional and indirect digital panoramic radiograph.

Efficacy of deep convolutional neural network algorithm for the identification and classification of dental implant systems, using panoramic and periapical radiographs: A pilot study.

Convolutional neural networks (CNNs), a particular type of deep learning architecture, are positioned to become one of the most transformative technologies for medical applications. The aim of the current study was to evaluate the efficacy of deep CNN algorithm for the identification and classification of dental implant systems.A total of 5390 panoramic and 5380 periapical radiographic images from 3 types of dental implant systems, with similar shape and internal conical connection, were randomly divided into training and validation dataset (80%) and a test dataset (20%). We performed image preprocessing and transfer learning techniques, based on fine-tuned and pre-trained deep CNN architecture (GoogLeNet Inception-v3). The test dataset was used to assess the accuracy, sensitivity, specificity, receiver operating characteristic curve, area under the receiver operating characteristic curve (AUC), and confusion matrix compared between deep CNN and periodontal specialist.We found that the deep CNN architecture (AUC = 0.971, 95% confidence interval 0.963–0.978) and board-certified periodontist (AUC = 0.925, 95% confidence interval 0.913–0.935) showed reliable classification accuracies.This study demonstrated that deep CNN architecture is useful for the identification and classification of dental implant systems using panoramic and periapical radiographic images.

Development of a Deep Learning Algorithm for Periapical Disease Detection in Dental Radiographs.

Periapical radiolucencies, which can be detected on panoramic radiographs, are one of the most common radiographic findings in dentistry and have a differential diagnosis including infections, granuloma, cysts and tumors. In this study, we seek to investigate the ability with which 24 oral and maxillofacial (OMF) surgeons assess the presence of periapical lucencies on panoramic radiographs, and we compare these findings to the performance of a predictive deep learning algorithm that we have developed using a curated data set of 2902 de-identified panoramic radiographs. The mean diagnostic positive predictive value (PPV) of OMF surgeons based on their assessment of panoramic radiographic images was 0.69 (±0.13), indicating that dentists on average falsely diagnose 31% of cases as radiolucencies. However, the mean diagnostic true positive rate (TPR) was 0.51 (±0.14), indicating that on average 49% of all radiolucencies were missed. We demonstrate that the deep learning algorithm achieves a better performance than 14 of 24 OMF surgeons within the cohort, exhibiting an average precision of 0.60 (±0.04), and an F1 score of 0.58 (±0.04) corresponding to a PPV of 0.67 (±0.05) and TPR of 0.51 (±0.05). The algorithm, trained on limited data and evaluated on clinically validated ground truth, has potential to assist OMF surgeons in detecting periapical lucencies on panoramic radiographs.