Related Experiment Video
Updated: May 28, 2026
10:32
Detection and Quantitation of Label-Retaining Cells in Mouse Incisors using a 3D Reconstruction Approach after Tissue Clearing
Published on: June 10, 2022
Non-destructive classification of type 2 diabetic teeth using LIBS-derived elemental biomarkers and machine learning
Muhammad Mustafa Dastageer1, Khurram Siraj1, Giorgio Saverio Senesi2
1Laser & Optronics Centre, Department of Physics, University of Engineering and Technology (UET), Main Campus, G.T. Road, Lahore, Pakistan.
Analytica Chimica Acta
|May 26, 2026
Summary
Dental tissues can reveal elemental changes linked to Type 2 diabetes mellitus (T2DM). Laser-induced breakdown spectroscopy (LIBS) with machine learning (ML) offers a new, non-destructive method for early diabetes screening in dental settings.
Area of Science:
- Biomedical Engineering
- Analytical Chemistry
- Oral Medicine
Background:
- Type 2 Diabetes Mellitus (T2DM) significantly impacts oral health, increasing risks of periodontal disease and dental treatment failure.
- Hyperglycemia associated with T2DM alters oral health through various mechanisms.
- Mineralized tooth tissues, which retain long-term metabolic information, are largely unexplored for diabetes diagnostics.
Purpose of the Study:
- To investigate the potential of mineralized dental tissues as a substrate for detecting diabetes-induced elemental alterations.
- To evaluate the efficacy of laser-induced breakdown spectroscopy (LIBS) combined with machine learning (ML) for non-destructive diabetes screening.
- To identify specific elemental biomarkers in teeth associated with T2DM.
Main Methods:
- Collected 3600 LIBS spectra from four dental tissues (enamel, coronal dentine, radicular dentine, cementum) of 30 individuals (15 healthy, 15 diabetic).
- Analyzed elemental composition using LIBS and employed ML models (LR, SVM, ANN) with Principal Component Analysis (PCA) and Correlation-based Feature Selection (CFS-BFS).
- Assessed model performance using accuracy, sensitivity, and specificity metrics.
Main Results:
- LIBS analysis revealed higher iron (Fe) and tin (Sn) and lower zinc (Zn), silicon (Si), and potassium (K) in diabetic teeth, identifying them as potential diabetes biomarkers.
- The PCA-ANN model achieved the highest performance (96% accuracy, 94% sensitivity, 96% specificity).
- PCA with 6 components captured 94% of variance, consistently outperforming CFS-BFS integrated ML algorithms, with PCA-SVM showing high stability.
Conclusions:
- Dental tissues retain elemental signatures of diabetes detectable by LIBS-ML, offering a non-destructive screening approach.
- This LIBS-ML technique can enable pre-treatment risk identification in dental practice, potentially reducing treatment failures and improving patient outcomes.
- Further large-scale multicenter validation studies are necessary to integrate this method into clinical practice.
