Parul Dixit¹ and Rajeev R²*

¹Resident, Department of Oral Pathology and Microbiology, Government Dental College, Medical College, Thiruvananthapuram, India
²Additional Professor, Department of Oral Pathology and Microbiology, Government Dental College, Medical College, Thiruvananthapuram, India

*Corresponding Author: Rajeev R, Additional Professor, Department of Oral Pathology and Microbiology, Government Dental College, Medical College, Thiruvananthapuram, India.

Received: Sepetember 08, 2025; Published: Sepetember 16, 2025

Abstract

Introduction: Early detection of malignant transformation in oral potentially malignant disorders (OPMDs) is vital for improving survival in oral squamous cell carcinoma (OSCC). Conventional diagnostics, including histopathology, are invasive and subject to variability. Radiomics, which extracts quantitative features from imaging data, combined with machine learning (ML) and deep learning (DL), offers reproducible, non-invasive biomarkers to complement clinical assessment.

Discussion: Radiomics workflows involve image acquisition, segmentation, feature extraction, and predictive modeling. Applied across modalities such as CT, MRI, PET, ultrasound, autofluorescence, and optical coherence tomography (OCT), radiomic features integrated with ML algorithms enable risk stratification, guided biopsies, and longitudinal monitoring of OPMDs. Emerging technologies, including hyperspectral imaging, Raman spectroscopy, digital pathology, and liquid biopsy integration, further enhance diagnostic potential. Recent advances in software and AI platforms—such as PyRadiomics, 3D Slicer, CaPTk, MONAI, AutoRadiomics, cloud AI, federated learning, and explainability frameworks—are accelerating clinical translation. However, challenges persist due to heterogeneous data, small sample sizes, and segmentation variability. Solutions include harmonization techniques, data augmentation, robust automatic segmentation, and adherence to reporting standards.

Conclusion: Radiomics and AI-driven methods show strong promise for early, non-invasive detection of OPMD progression to OSCC. While standardization, reproducibility, and clinical validation remain barriers, advances in AI ecosystems and collaborative, multi-institutional research are paving the way toward precision diagnostics and clinical implementation.

Keywords: Radiomics; Machine Learning; Oral Potentially Malignant Disorders; Oral Squamous Cell Carcinoma; Artificial Intelligence; Imaging Biomarkers; Early Detection

References

1. Warnakulasuriya S. “Oral potentially malignant disorders: A comprehensive review”. Oral Oncology (2020).
2. Sung H., et al. “Global cancer statistics 2020”. CA Cancer Journal of Clinics (2021).
3. Mortazavi H., et al. “Oral potentially malignant disorders: An overview”. Journal of Dental Research, Dental Clinics, Dental Prospects (2014).
4. Kujan O., et al. “Oral epithelial dysplasia: Biopsy challenges and diagnostic pitfalls”. Journal of Oral Pathology and Medicine (2021).
5. Lambin P., et al. “Radiomics: Extracting more information from medical images”. European Journal of Cancer (2012).
6. Gillies RJ., et al. “Radiomics and its potential in precision medicine”. Nature Reviews Clinical Oncology (2016).
7. Zwanenburg A., et al. “The Image Biomarker Standardisation Initiative”. Radiology (2020).
8. Fedorov A., et al. “3D Slicer as an image computing platform for the quantitative imaging community”. Magnetic Resonance Imaging (2012).
9. van Griethuysen JJM., et al. “PyRadiomics: a flexible open source platform for radiomic feature extraction”. Cancer Research (2017).
10. Parmar C., et al. “Robust radiomic feature selection and its generalizability”. Scientific Report (2015).
11. Hawkins DM. “The problem of overfitting and model validation”. Journal of Chemical Information and Modeling (2004).
12. LeCun Y., et al. “Deep learning”. Nature (2015).
13. Lundberg SM and Lee SI. “A unified approach to interpreting model predictions (SHAP)”. NeurIPS (2017).
14. Collins GS and Moons KGM. “Reporting of prediction models — guidance and standards (TRIPOD)”. JAMA (2019).
15. Poh CF., et al. “Optical imaging and radiomics in oral cancer screening”. Cancer Prevention Research (2019).
16. Roblyer D., et al. “Clinical utility of autofluorescence imaging and algorithms”. Translational Oncology (2011).
17. Rao A., et al. “Radiomic signatures for predicting OPMD transformation”. Oral Oncology (2020).
18. Lu G and Fei B. “Hyperspectral imaging for cancer detection and diagnosis: a review”. Journal of Biomedical Optics (2014).
19. Yu K., et al. “Deep learning for OCT in oral dysplasia detection”. Biomedical Optics Express (2020).
20. Widjaja E., et al. “Raman spectroscopy and multiphoton microscopy in oral cancer research”. International Journal of Cancer (2010).
21. Dibble EH., et al. “PET radiomics in oncology: potential and pitfalls”. Journal of Nuclear Medicine (2019).
22. Zhang L., et al. “Delta-radiomics: temporal analysis for early progression detection”. Oral Oncology (2020).
23. Cai X., et al. “Pathomics model predicts malignant transformation of oral epithelial dysplasia”. Laboratory Investigation (2023).
24. Shephard AJ., et al. “AI in histopathology for prediction of OPMD transformation”. NPJ Precis Oncology (2023).
25. Aerts HJWL., et al. “Decoding tumour phenotype by radiomics and linking to genomics”. Nature Communication (2014).
26. van Griethuysen JJM., et al. “Computational radiomics system: PyRadiomics”. Cancer Research (2017).
27. Bakas S., et al. “The Cancer Imaging Phenomics Toolkit (CaPTk)”. Neuroinformatics (2018).
28. Cardoso MJ., et al. “MONAI: medical open network for AI”. Journal of Imaging (2022).
29. Johnson KW., et al. “Cloud AI platforms for medical imaging research”. NPJ Digital Medicine (2021).
30. Meyer M., et al. “Auto Radiomics and automated multi-modal radiomics frameworks”. European Radiology Experimental (2023).

Citation

Citation: Parul Dixit and Rajeev R. “Radiomics and Machine Learning in the Early Detection of Oral Potentially Malignant Disorders (OPMD) Transition to Oral Squamous Cell Carcinoma (OSCC)".Acta Scientific Otolaryngology 7.10 (2025): 11-16.

Copyright

Copyright: © 2025 Parul Dixit and Rajeev R. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.