Image collection and annotation platforms to establish a multi-source database of oral lesions

A comprehensive dataset of annotated oral cavity images for diagnosis of oral cancer and oral potentially malignant disorders

OBJECTIVES

This study aims to address the critical gap of unavailability of publicly accessible oral cavity image datasets for developing machine learning (ML) and artificial intelligence (AI) technologies for the diagnosis and prognosis of oral cancer (OCA) and oral potentially malignant disorders (OPMD), with a particular focus on the high prevalence and delayed diagnosis in Asia.

MATERIALS AND METHODS

Following ethical approval and informed written consent, images of the oral cavity were obtained from mobile phone cameras and clinical data was extracted from hospital records from patients attending to the Dental Teaching Hospital, Peradeniya, Sri Lanka. After data management and hosting, image categorization and annotations were done by clinicians using a custom-made software tool developed by the research team.

RESULTS

A dataset comprising 3000 high-quality, anonymized images obtained from 714 patients were classified into four distinct categories: healthy, benign, OPMD, and OCA. Images were annotated with polygonal shaped oral cavity and lesion boundaries. Each image is accompanied by patient metadata, including age, sex, diagnosis, and risk factor profiles such as smoking, alcohol, and betel chewing habits.

CONCLUSION

Researchers can utilize the annotated images in the COCO format, along with the patients' metadata, to enhance ML and AI algorithm development.

Oral mucosal lesions triage via YOLOv7 models

BACKGROUND/PURPOSE

The global incidence of lip and oral cavity cancer continues to rise, necessitating improved early detection methods. This study leverages the capabilities of computer vision and deep learning to enhance the early detection and classification of oral mucosal lesions.

METHODS

A dataset initially consisting of 6903 white-light macroscopic images collected from 2006 to 2013 was expanded to over 50,000 images to train the YOLOv7 deep learning model. Lesions were categorized into three referral grades: benign (green), potentially malignant (yellow), and malignant (red), facilitating efficient triage.

RESULTS

The YOLOv7 models, particularly the YOLOv7-E6, demonstrated high precision and recall across all lesion categories. The YOLOv7-D6 model excelled at identifying malignant lesions with notable precision, recall, and F1 scores. Enhancements, including the integration of coordinate attention in the YOLOv7-D6-CA model, significantly improved the accuracy of lesion classification.

CONCLUSION

The study underscores the robust comparison of various YOLOv7 model configurations in the classification to triage oral lesions. The overall results highlight the potential of deep learning models to contribute to the early detection of oral cancers, offering valuable tools for both clinical settings and remote screening applications.

Practical applications of teledentistry during the Covid-19 pandemic in ASEAN member states - a systematic review

OBJECTIVE

The objective of this review is to determine the utilisation and adoption of teledentistry based solutions and technologies during the Covid-19 Pandemic in the Asean region.

BACKGROUND

Teledentistry is a branch of telemedicine that has rapidly advanced in the last few years and has the potential to provide solutions to oral health problems of patients and locations that do not have prompt and immediate access to a dentist or dental services. The Covid-19 has increased the adaption of all digital health technologies and teledentistry is no exception.

METHODOLOGY

The study utilized online databases such as Pubmed (Medline), Scopus (Embase) and CINAHL for the purpose of document search. Newcastle Ottawa (NOS) scale was used to determine the quality of the studies included in our systematic review. PRISMA guidelines were used as the criteria for reporting items in the systematic review.

RESULTS

A total of 1297 documents were found after applying the search criteria and the keywords for the selected study. After applying the Prisma guidelines, removal of duplicates and irrelevant entries, 10 studies that were conducted during the Covid-19 pandemic were selected, fitting the inclusion criteria. All the studies included were evaluated for quality and risk of bias through the Newcastle Ottawa scale. Only high-quality studies were included for the final review.

CONCLUSION

Teledentistry is a cost-effective solution to screen, diagnose and treat dental patients from a distance. Teledentistry also has the potential to continue seamless continuation of dental education to dental students, during disruptive and non-disruptive periods. ASEAN countries should fully utilise the potential of teledentistry, however sound and effective legislation would be the key first step to achieving that potential.

Calibration improves the agreement in grading oral epithelial dysplasia-Findings from a National Workshop in Malaysia

INTRODUCTION

A major pitfall of many of the established oral epithelial dysplasia (OED) grading criteria is their lack of reproducibility and accuracy to predict malignant transformation. The main objective of this study was to determine whether calibration of practicing oral pathologists on OED grading could improve the reproducibility of the WHO 2017 and the binary OED grading systems.

METHODS

A nationwide online exercise was carried out to determine the influence of calibration on the reproducibility of the WHO 2017 and the binary OED grading systems.

RESULTS

A significant improvement was observed in the inter-observer agreement for the WHO 2017 OED grading system (K 0.196 vs. 0.448; Kw 0.357 vs. 0.562) after the calibration exercise. The significant difference (p = 0.027) in the level of agreement between those with five or more years and less than 5 years of experience was no more observed (p = 0.426) after the calibration exercise. The percent agreement for binary grading was significantly higher (91.8%) for buccal mucosal lesions as compared to lesions on the tongue after the calibration exercise.

CONCLUSION

This study validates the significance of calibration in improving the reproducibility of OED grading. The nationwide exercise resulted in a statistically significant improvement in the inter-observer agreement for the WHO 2017 OED grading system among a large number of oral pathologists. It is highly recommended that similar exercises should be organized periodically by professional bodies responsible for continuing education among oral pathologists to improve the reliability of OED grading for optimal treatment of oral potentially malignant disorders.