A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes

Abstract Background Orthopantomograms (OPGs) are essential diagnostic tools in dental and maxillofacial care, providing a panoramic view of the jaws, teeth, and surrounding bone structures. Detecting bone loss, which indicates periodontal disease and systemic conditions like osteoporosis, is crucial...

Full description

Saved in:
Bibliographic Details
Main Authors: Pradeep Kumar Yadalam, Amit Rajabhau Pawar, Prabhu Manickam Natarajan, Carlos M. Ardila
Format: Article
Language:English
Published: BMC 2025-07-01
Series:Head & Face Medicine
Subjects:
Online Access:https://doi.org/10.1186/s13005-025-00528-3
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849331641602801664
author Pradeep Kumar Yadalam
Amit Rajabhau Pawar
Prabhu Manickam Natarajan
Carlos M. Ardila
author_facet Pradeep Kumar Yadalam
Amit Rajabhau Pawar
Prabhu Manickam Natarajan
Carlos M. Ardila
author_sort Pradeep Kumar Yadalam
collection DOAJ
description Abstract Background Orthopantomograms (OPGs) are essential diagnostic tools in dental and maxillofacial care, providing a panoramic view of the jaws, teeth, and surrounding bone structures. Detecting bone loss, which indicates periodontal disease and systemic conditions like osteoporosis, is crucial for early diagnosis and treatment planning. Periodontists use OPGs to identify subtle radiographic features that signify different stages of bone loss. Automated systems integrating radiographic imaging with textual notes can enhance diagnostic accuracy and minimize interobserver variability. Radiographic notes, which summarize clinical observations and preliminary interpretations, can be utilized for classification through natural language processing techniques, including Transformer-based models. This study will classify bone loss severity (normal, mild, or severe) from OPG notes using a novel dual-embedding few-shot learning framework. Methods This study used a dataset of radiographic notes from OPGs gathered at Saveetha Dental College and Hospital in Chennai. Bone loss was classified according to Glickman’s Classification system. The proposed DualFit model architecture consists of two main branches: a Text Processing Branch for converting textual data into dense vectors and a Feature Processing Branch for analyzing numerical and categorical data. Key techniques such as batch normalization and dropout layers were implemented to improve learning stability and reduce overfitting. A Fusion Layer was utilized to merge outputs from both branches, optimizing classification performance. Results The DualFit model outperformed leading models like BioBERT, ClinicalBERT, and PubMedBERT. It attained an accuracy of 98.98%, precision of 98.71%, recall of 99.14%, and an F1-score of 98.92%, marking a 5.53% accuracy increase over PubMedBERT. Additionally, the model excelled in multi-class classification tasks, ensuring class balance and achieving near-perfect values for precision, recall, and area under both the ROC and precision-recall curves. Conclusions The DualFit model significantly advances the automated classification of OPG radiographic notes related to periodontal bone loss. Outperforming existing Transformer-based models streamlines the diagnostic workflow, reduces the workload of radiologists, and enables timely interventions for improved patient outcomes. Future work should explore external validation and integration with multimodal diagnostic systems.
format Article
id doaj-art-667bb21dce56452da67ef9391c0fdd0a
institution Kabale University
issn 1746-160X
language English
publishDate 2025-07-01
publisher BMC
record_format Article
series Head & Face Medicine
spelling doaj-art-667bb21dce56452da67ef9391c0fdd0a2025-08-20T03:46:28ZengBMCHead & Face Medicine1746-160X2025-07-0121111310.1186/s13005-025-00528-3A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notesPradeep Kumar Yadalam0Amit Rajabhau Pawar1Prabhu Manickam Natarajan2Carlos M. Ardila3Department of Periodontics, Saveetha Dental College and Hospitals, Saveetha Institute of Medical and Technical Sciences (SIMATS)Department of Periodontics, Saveetha Dental College and Hospitals, Saveetha Institute of Medical and Technical Sciences (SIMATS)Department of Clinical Sciences, Center of Medical and Bio-allied Health Sciences and Research, College of Dentistry, Ajman UniversityDepartment of Periodontics, Saveetha Dental College and Hospitals, Saveetha Institute of Medical and Technical Sciences (SIMATS)Abstract Background Orthopantomograms (OPGs) are essential diagnostic tools in dental and maxillofacial care, providing a panoramic view of the jaws, teeth, and surrounding bone structures. Detecting bone loss, which indicates periodontal disease and systemic conditions like osteoporosis, is crucial for early diagnosis and treatment planning. Periodontists use OPGs to identify subtle radiographic features that signify different stages of bone loss. Automated systems integrating radiographic imaging with textual notes can enhance diagnostic accuracy and minimize interobserver variability. Radiographic notes, which summarize clinical observations and preliminary interpretations, can be utilized for classification through natural language processing techniques, including Transformer-based models. This study will classify bone loss severity (normal, mild, or severe) from OPG notes using a novel dual-embedding few-shot learning framework. Methods This study used a dataset of radiographic notes from OPGs gathered at Saveetha Dental College and Hospital in Chennai. Bone loss was classified according to Glickman’s Classification system. The proposed DualFit model architecture consists of two main branches: a Text Processing Branch for converting textual data into dense vectors and a Feature Processing Branch for analyzing numerical and categorical data. Key techniques such as batch normalization and dropout layers were implemented to improve learning stability and reduce overfitting. A Fusion Layer was utilized to merge outputs from both branches, optimizing classification performance. Results The DualFit model outperformed leading models like BioBERT, ClinicalBERT, and PubMedBERT. It attained an accuracy of 98.98%, precision of 98.71%, recall of 99.14%, and an F1-score of 98.92%, marking a 5.53% accuracy increase over PubMedBERT. Additionally, the model excelled in multi-class classification tasks, ensuring class balance and achieving near-perfect values for precision, recall, and area under both the ROC and precision-recall curves. Conclusions The DualFit model significantly advances the automated classification of OPG radiographic notes related to periodontal bone loss. Outperforming existing Transformer-based models streamlines the diagnostic workflow, reduces the workload of radiologists, and enables timely interventions for improved patient outcomes. Future work should explore external validation and integration with multimodal diagnostic systems.https://doi.org/10.1186/s13005-025-00528-3OrthopantomogramsBone lossPeriodontitisClinicalBERTBERTText classification
spellingShingle Pradeep Kumar Yadalam
Amit Rajabhau Pawar
Prabhu Manickam Natarajan
Carlos M. Ardila
A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes
Head & Face Medicine
Orthopantomograms
Bone loss
Periodontitis
ClinicalBERT
BERT
Text classification
title A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes
title_full A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes
title_fullStr A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes
title_full_unstemmed A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes
title_short A novel dual embedding few-shot learning approach for classifying bone loss using orthopantomogram radiographic notes
title_sort novel dual embedding few shot learning approach for classifying bone loss using orthopantomogram radiographic notes
topic Orthopantomograms
Bone loss
Periodontitis
ClinicalBERT
BERT
Text classification
url https://doi.org/10.1186/s13005-025-00528-3
work_keys_str_mv AT pradeepkumaryadalam anoveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT amitrajabhaupawar anoveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT prabhumanickamnatarajan anoveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT carlosmardila anoveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT pradeepkumaryadalam noveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT amitrajabhaupawar noveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT prabhumanickamnatarajan noveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes
AT carlosmardila noveldualembeddingfewshotlearningapproachforclassifyingbonelossusingorthopantomogramradiographicnotes