eISSN: 2299-551X
ISSN: 0011-4553
Journal of Stomatology
Current issue Archive Manuscripts accepted About the journal Editorial board Reviewers Abstracting and indexing Subscription Contact Instructions for authors Ethical standards and procedures
Editorial System
Submit your Manuscript
SCImago Journal & Country Rank
vol. 77
Review paper

The role of radiomics in dentistry and oral radiology

Tannishtha .
Shruthi Hegde
G. Subhas Babu
Vidya Ajila
B.S. Shama

  1. Department of Oral Medicine and Radiology, Nitte (Deemed to be University), AB Shetty Memorial Institute of Dental Sciences (ABSMIDS), Mangalore, India
J Stoma 2024; 77, 2: 147-152
Online publish date: 2024/05/29
Article file
- JOS-00867.pdf  [0.16 MB]
Get citation
PlumX metrics:


Artificial intelligence (AI) research has advanced, and is used outside of a computer science department. As a result, there was a steady rise in the use of AI applications in healthcare. AI technology is intended to help clinical practitioners in clinical decision-making, and to reduce repetitive tasks performed in their daily work [1]. Radiomics is a quantitative approach to medi­cal imaging that aims to enhance the available radiographic data through sophisticated mathematical analysis. The main idea of radiomics is based on the concept that biomedi­cal imaging comprises data reflecting disease-specific processes and is accessible by quantitative image analyses [2]. Radiomics technology is produced as a result of combining genomic data, imaging, and pathology results [3]. The application of artificial intelligence-aided programs in maxillofacial radiology is increasing [4, 5].
Vast availability and an increased use of digital imagining techniques, such as computed tomography (CT) and cone beam computed tomography (CBCT), enabled the introduction of AI and radiomics in oral and maxil­lofacial imaging. The advent of radiomics in dentistry made it possible to improve diagnostic, prognostic, and predictive accuracy, by combining radiographic, biolo­gical, and clinical data. Qualitative data extracted from the radiographs using data-characterization algorithms in radiomics allow specialists to obtain additional diagnostic information for personalized care [6, 7]. Research in radiomics and AI have advanced in the field of dentistry, demonstrating these technologies’ potential to significantly enhance clinical treatment [6-9]. This article highlighted the potential role of radiomics in dentistry and oral radiology.

Concept of artificial intelligence

Intelligence is referred to as the capacity to learn and use information and skills. Artificial intelligence (AI) can be defined as the non-biological potential to perform complicated tasks [10]. Currently, AI become one of the most sought-after domains in terms of research, and has a vital role in many people’s daily activi­ties [11].
Artificial intelligence in dentistry mainly depends on the output of a number of schemes based on pre-determined programming models and computer analy­sis, using complicated mathematical models and formulas [12]. Machine learning (ML) is a sub-set of AI. Computer-aided detection and diagnosis are two important areas of medical applications in machine learning [13]. Algorithms for image processing are often used in radiology, especially for in-depth evaluation of multi-planar medical images. These techniques start from extracting the features of a given image before conducting target detection or classifying the image into pre-defined groups, to achieve image detection or classification. Out of all the existing ML models, convolutional neural networks (CNNs) are the most utilized technology in medical imaging due to their exceptional effectiveness in enhancing image diagnostic characteristics [14]. Despite their role as promising adjunctive tools in radiology, the applications of ML algorithms in dental and maxillofacial radiology are still in early phases of development [15].

AI-enabled maxillofacial imaging

Artificial intelligence is gradually influencing every part of our lives with various conveniences, such as speakers with built-in AI, content recommendation systems, etc. The development of deep learning also opens exciting opportunities for the automation of picture analysis in the fields of dentistry and medicine. Signifi­cant developments have been made in every subject of artificial intelligence, including robotics, data mining, medical image analysis, and processing [16]. Studies on artificial intelligence for diagnosing dental caries, perio­dontal disease, odontogenic cysts and tumors, diseases of the maxillary sinus, temporomandibular joints, and osteosclerosis, have shown promising results in the field of oral radiology [4]. AI-enabled programs are used in maxillofacial radiology for cephalometric tracing, detection of caries, alveolar bone loss, periapical pathosis, detection and tracing of the inferior alveolar nerve canal, and other similar functions [17].
A deep learning hybrid approach was recently developed by Chang et al. [18] for the automated staging of periodontitis on dental panoramic radiographs. The basis of their AI system was a combination of deep learning architecture and traditional computer-assisted diagnosis approach. Their system demonstrated greater accuracy and reliability in staging periodontitis based on the amount of alveolar bone loss.
Image analysis of a maxillofacial radiograph using artificial intelligence has been employed for a variety of tasks, including segmentation or localization of teeth, evaluation of bone quantity and quality, estimation of age using hand-wrist radiographs based on analysis of the pattern of ossifications, and localization of cephalometric landmarks [17, 19].
Panoramic radiographs are considered an important modality for identifying osteopenia and osteoporosis. In post-menopausal females, osteoporosis can be diagnosed using the reduction in mandibular cortical width and severity of erosion of the mandibular lower cortex. There are artificial intelligence models in panoramic radiography available to diagnose osteopenia and osteoporosis [17]. Lee et al. [20] analyzed the efficacy of a deep convolutional neural network (DCNN)-based CAD system in assessing diagnostic performance for the detection of osteoporosis on panoramic radiographs, and reported satisfactory results. Promising findings from AI-enabled technologies have led to a sharp rise in the innovation of new deep learning techniques for oral and cranio-facial radiology, including head and neck oncology [11, 15].


The “-omics” ideology has emerged in the recent decade due to the improvements in high-throughput computing and machine learning methods. It refers to the collective characterization and quantification of biological information pools, including genomics, proteomics, and metabolomics. Radiomics consists of the automatic extraction of numerical descriptors with mathematical definitions, referred to as “radiomics features”, from 2-dimensional and 3-dimensional radiographic image series [21].
Physical characteristics of the imaged region, such as tissue cellularity, heterogeneity, and necrosis, are represented by radiomics features. According to the literature evidence on radiomics research, radiomics traits usually correlate with diagnostic and outcome variables [22, 23]. Moreover, radiomics comprises algorithms that divide input images into basic properties, such as edges, gradients, form, signal intensity, wavelength, and textures, which can be further used to categorize or analyze the radiographic image. Therefore, a straightforward definition of radiomic analysis is the extraction of quantitative data in a radiograph to measure and extract specific parameters. As a result, radiomics’ software can define or find numerous abstract mathematical properties on images based on variations in the spatial patterns of grey levels and signal intensity, which are typically not visible to human sight in images and cannot be assessed visually [7].
The workflow of radiomics involves image acquisition, image segmentation, image processing, feature extraction, feature selection, dimension reduction, and analysis. Acquisition of high-quality and standardized imaging is necessary [1, 24]. The source of data for radiomics are retrospective medical images. Different techniques for imaging may lead to changes in image signals and image textures due to various parameters used for acquisition and reconstruction in medical imaging. Standardization is essential for radiomics image analysis, and segmentation also must be reproducible and reliable. Automated segmentation is preferred over manual and semi-automatic. The automated professional software usually performs the extraction of features in radiomics, and shape and texture features are mined [1, 24]. The workflow of radiomics is shown in Figure 1 [1, 24].
Valid and quantitative features from medical images extracted in radiomics can be combined with other characteristics, such as clinical staging, pathological features, and tissue molecular markers [1, 24].
Various clinical decision-making systems have been developed in recent years based on radiomics, which may aid in the correlation of radiographic data with biological and clinical endpoints, and could increase the accuracy of diagnosis and prediction of possible prognosis. Investigation of correlations between quantitative bio-imaging features and disease properties resulted in a sub-specialty known as radio-genomics [25, 26].

Applications of radiomics in oral and maxillofacial radiology

The potential of radiomics in the diagnosis and treatment planning of oral and maxillofacial lesions is increasing. Hung et al. [27] have recently reported that the performance of radiomics models on CT/CBCT images showed promising features for maxillofacial diseases. The roles of radiomics in odontogenic and non-odontogenic cysts and tumors of the jaws, developmental deformities of the dentomaxillofacial region, oral cancer, cervical lymph node metastasis, diseases of salivary glands, temporomandibular (TMJ) disorders, paranasal sinus pathologies, oro-maxillofacial fractures, age estimation, and endodontic therapy, are described in Table 1 [8, 9, 27-29].

Radiomics and oral onco-diagnosis

Head and neck cancer has a variety of challenges, which are hard to diagnose and treat. The role of imaging is inevitable in cancer, especially in head and neck carcinoma (HNC), both pre and post-operatively, for treatment planning and determination of prognosis. Contrast-enhanced computed tomography (CECT), magnetic resonance imaging (MRI), and positron emission tomography (PET) imaging, are the most widely used modalities in head and neck cancer (HNC) cases for the evaluation of anatomic extent, nodal involvement, perineural invasion, skull base and intra-cranial involvement, cartilaginous involvement, calcifications, infiltrations, and associated inflammations related to HNC [30].
However, the diagnosis is challenging due to complex regional anatomy, presence of numerous small-sized vital structures, pathways of primary and recurrent tumors, and high intra-tumoral heterogeneity that changes depending on the site and biological character of the tumor. Radiomics is considered an emerging solution for the potential diagnostic challenges associated with HNC imaging [31].
By applying radiomics in head and neck oncology, the radiographic data obtained from various imaging modalities facilitate the exploration of relevant diagnostic data by radiomic analysis. Additionally, it is possible to non-invasively capture the diverse structure and biological behavior of head and neck malignancies, which might be crucial for clinical decision-making [31]. Radiomics’ data must be extracted in a way that can be repeated and the same each time for clinical applications to be possible. In onco-diagnosis, the radiomics workflow starts with a standardized protocol for obtaining high-quality images. Next step is the segmentation of tumor, followed by details of the tumor area. Then, the extracted features that show the best performance, stability, or other similar defining metrics, are selected for using in clinical applications [32].
Previous research demonstrated that textural features taken from CT and MR images can be used to learn more about the tumors’ heterogeneity that is linked to necrosis, angiogenesis, and hypoxia [3, 33]. However, a study by Ger et al. [34] reported radiomics to be inconsistent with regard to the estimation of survival in CT or PET images of head and neck patients. Although discrepancies exist in the validation of radiomics-aided onco-diagnosis, active research is being undertaken worldwide for AI-aided clinical decisions feasible in oncology [35].

Radiomics and dental implants

The long-term success of implant therapy is based on peri-implant marginal bone stability. Within one year of functional loading, physiological re-modelling of the peri-implant crestal bone can occur [36]. Inade­quate crestal width, traumatic surgery, supra-cranial tissue height establishment, microbial colonization of implant-abutment interface, implant crest module characteristics, number of abutment connections/disconnections, height of prosthetic abutment, mechanical stability of implant/abutment connection, and adaptability, are just a few of surgical and restorative factors, which can affect physiological bone re-modelling (PBR), a complex process with multi-factorial etiology. It is still debatable whether the quality of the bone at the time of implant placement influences the development of peri-implantitis and/or late implant failure [37]. In contrast, radiomics takes a quantitative approach to medical imaging with the goal of enhancing data that doctors have access to through sophisticated mathematical calculations [38]. Radiomics develops prediction models with sufficient accuracy to help clinical decisions by extracting and analyzing “features” or “quantitative characteristics” from medical pictures. In a recent study by Troiano et al. [39], it was determined that radiomic analysis could reveal subtle bone characteristics helpful in the early prediction of PBR. The study’s results revealed that after three months of unsubmerged healing, specific radiomic features were associated with higher PBR around tissue-level implants, and the researchers came to the conclusion that using radiomics in conjunction with machine learning techniques appears to be a promising strategy.

Future perspectives

Incorporating deep learning models in radiology has improved the efficiency of diagnostic data by reducing the time required for data interpretation without compromising diagnostic accuracy. It enabled bone age determination with accuracy comparable to that of a trained radiologist [40-42]. However, AI-enabled radio-diagnosis and fully automated image interpretation are evolving concepts, which will be soon available, the European Society of Radiology emphasized the inevitable role of human intervention by trained radiologists in radio-diagnosis because radiologists play a vital in both developing and validating AI applications in medical imaging [43].
Limitations of radiomics include lack of standardization in image acquisition, multiple software usage, and different statistical approaches, which restricts the comparison between studies and data reproduction. It is of utmost importance to use standardized imaging protocols with open-source software to eliminate the confounding variability with respect to radiomics. Variations with scanner manufacturer, model, calibrations, and differences in algorithms and software may lead to different results [44].


Radiomics is an advanced application of artificial intelligence in diagnostic radiology combining clinical and radiographic data, to establish diagnostic and prognostic accuracy. Although radiomics in the field of oral and maxillofacial radiology is still in the growing phase, it has a considerable potential to be applied in day-to-day radiology practice.


  1. Institutional review board statement: Not applicable.
  2. Assistance with the article: None.
  3. Financial support and sponsorship: None.
  4. Conflicts of interest: The authors declare no potential conflicts of interest concerning the research, authorship, and/or publication of this article.
1. Van Timmeren JE, Cester D, Tanadini-Lang S, Alkadhi H, Baessler B. Radiomics in medical imaging-”how-to” guide and critical reflection. Insights Imaging 2020; 11: 91. DOI: 10.1186/s13244-020-00887-2.
2. Neisius U, El-Rewaidy H, Nakamori S, Rodriguez J, Manning WJ, Nezafat R. Radiomic analysis of myocardial native T1 imaging discriminates between hypertensive heart disease and hypertrophic cardiomyopathy. JACC Cardiovasc Imaging 2019; 12: 1946-1954.
3. Ganeshan B, Goh V, Mandeville HC, Ng QS, Hoskin PJ, Miles KA. Non-small cell lung cancer: histopathologic correlates for texture parameters at CT. Radiology 2013; 266: 326-336.
4. Heo MS, Kim JE, Hwang JJ, Han SS, Kim JS, Yi WJ, et al. Artificial intelligence in oral and maxillofacial radiology: what is currently possible? Dento Maxillo Fac Radiol 2021; 50: 20200375. DOI: 10.1259/dmfr.20200375
5. Chen YW, Stanley K, Att W. Artificial intelligence in dentistry: current applications and future perspectives. Quintessence Int 2020; 51: 248-257.
6. Avanzo M, Stancanello J, El Naqa I. Beyond imaging: the promise of radiomics. Phys Med 2017; 38: 122-139.
7. Leite AF, Vasconcelos KF, Willems H, Jacobs R. Radiomics and machine learning in oral healthcare. Proteomics Clin Appl 2020; 14: e1900040. DOI: 10.1002/prca.201900040.
8. Jeon KJ, Kim YH, Choi H, Ha EG, Jeong H, Han SS. Radiomics approach to the condylar head for legal age classification using cone-beam computed tomography: a pilot study. PLoS One 2023; 18: e0280523. DOI: 10.1371/journal.pone.0280523.
9. Hasan HA, Saad FH, Ahmed S, Mohammed N, Farook TH, Dudley J. Experimental validation of computer-vision methods for the successful detection of endodontic treatment obturation and progression from noisy radiographs. Oral Radiol 2023; 39: 683-698.
10. Tegmark M. Life 3.0: Being human in the age of artificial intelligence. Vintage; 2018 Jul 31.
11. Ren R, Luo H, Su C, Yao Y, Liao W. Machine learning in dental, oral and craniofacial imaging: a review of recent progress. Peer J 2021; 9: e11451. DOI: 10.7717/peerj.11451.
12. Erickson BJ, Korfiatis P, Akkus Z, Kline TL. Machine learning for medical imaging. Radiographics 2017; 37: 505-515.
13. Gao J, Jiang Q, Zhou B, Chen D. Convolutional neural networks for computer-aided detection or diagnosis in medical image analysis: an overview. Math Biosci Eng 2019; 16: 6536-6561.
14. Lee JG, Jun S, Cho YW, et al. Deep learning in medical imaging: general overview. Korean J Radiol 2017; 18: 570-584. DOI: 10.3348/kjr.2017.18.4.570.
15. Schwendicke F, Golla T, Dreher M, Krois J. Convolutional neural networks for dental image diagnostics: a scoping review. J Dent 2019; 91: 103226. DOI: 10.1016/j.jdent.2019.103226.
16. Murray C. Deep blue. Artificial Intelligence 2002; 134: 57-83.
17. Hwang JJ, Jung YH, Cho BH, Heo MS. An overview of deep learning in the field of dentistry. Imaging Sci Dent 2019; 49: 1-7.
18. Chang HJ, Lee SJ, Yong TH, et al. Deep learning hybrid method to automatically diagnose periodontal bone loss and stage periodontitis. Sci Rep 2020; 10: 7531. DOI: 10.1038/s41598-020-64509-z.
19. Hung K, Montalvao C, Tanaka R, Kawai T, Bornstein MM. The use and performance of artificial intelligence applications in dental and maxillofacial radiology: a systematic review. Dentomaxillofac Radiol 2020; 49: 20190107. DOI: 10.1259/dmfr.20190107.
20. Lee JS, Adhikari S, Liu L, Jeong HG, Kim H, Yoon SJ. Osteoporosis detection in panoramic radiographs using a deep convolutional neural network-based computer-assisted diagnosis system: a preliminary study. Dentomaxillofac Radiol 2019; 48: 20170344. DOI: 10.1259/dmfr.20170344.
21. Haider SP, Burtness B, Yarbrough WG, Payabvash S. Applications of radiomics in precision diagnosis, prognostication and treatment planning of head and neck squamous cell carcinomas. Cancers Head Neck 2020; 5: 6. DOI: 10.1186/s41199-020-00053-7.
22. Yip SS, Aerts HJ. Applications and limitations of radiomics. Phys Med Biol 2016; 61: 150-166.
23. Gillies RJ, Kinahan PE, Hricak H. Radiomics: images are more than pictures, they are data. Radiology 2016; 278: 563-577.
24. Peng Z, Wang Y, Wang Y, et al. Application of radiomics and machine learning in head and neck cancers. Int J Biol Sci 2021; 17: 475-486.
25. Pinker K, Shitano F, Sala E, Do RK, Young RJ, Wibmer AG, et al. Background, current role, and potential applications of radiogenomics. J Magn Reson Imaging 2018; 47: 604-620.
26. Gerlinger M, Rowan AJ, Horswell S, Math M, Larkin J, Endesfelder D, et al. Intratumor heterogeneity and branched evolution revealed by multiregion sequencing. N Engl J Med 2012; 366: 883-892.
27. Hung KF, Ai QYH, Wong LM, Yeung AWK, Li DTS, Leung YY. Current applications of deep learning and radiomics on CT and CBCT for maxillofacial diseases. Diagnostics (Basel) 2022; 13: 110. DOI: 10.3390/diagnostics13010110.
28. Jiang ZY, Lan TJ, Cai WX, Tao Q. Primary clinical study of radiomics for diagnosing simple bone cyst of the jaw. Dentomaxillofac Radiol 2021; 50: 20200384. DOI: 10.1259/dmfr.20200384.
29. Kwon O, Yong T, Kang S, Kim JE, Huh KH, Heo MS, et al. Automatic diagnosis for cysts and tumors of both jaws on panoramic radiographs using a deep convolution neural network. Dentomaxillofac Radiol 2020; 49: 20200185. DOI: 10.1259/dmfr.20200185.
30. Schlumpf MF, Haerle S. The current role of imaging in head and neck cancer: a clinician’s perspective. Swiss Med Wkly 2014; 144: w14015. DOI: 10.4414/smw.2014.14015.
31. Wong AJ, Kanwar A, Mohamed AS, Fuller CD. Radiomics in head and neck cancer: from exploration to application. Transl Cancer Res 2016; 5: 371-382.
32. Lambin P, Rios-Velazquez E, Leijenaar R, Carvalho S, van Stiphout RGPM, Granton P, et al. Radiomics: extracting more information from medical images using advanced feature analysis. Eur J Cancer 2012; 48: 441-446.
33. Jethanandani A, Lin TA, Volpe S, Elhalawani H, Mohamed ASR, Yang P, et al. Exploring applications of radiomics in magnetic resonance imaging of head and neck cancer: a systematic review. Front Oncol 2018; 14: 131. DOI: 10.3389/fonc.2018.00131.
34. Ger RB, Zhou S, Elgohari B, Elhalawani H, Mackin DM, Meier JG, et al. Radiomics features of the primary tumor fail to improve prediction of overall survival in large cohorts of CT- and PET-imaged head and neck cancer patients. PLoS One 2019; 14: e0222509. DOI: 10.1371/journal.pone.0222509.
35. Ding H, Wu C, Liao N, Zhan Q, Sun W, Huang Y, et al. Radiomics in oncology: a 10-year bibliometric analysis. Front Oncol 2021; 11: 689802. DOI: 10.3389/fonc.2021.689802.
36. Oh TJ, Yoon J, Misch CE, Wang HL. The causes of early implant bone loss: myth or science? J Periodontol 2002; 73: 322-333. 
37. Chrcanovic BR, Albrektsson T, Wennerberg A. Bone quality and quantity and dental implant failure: a systematic review and meta-analysis. Int J Prosthodont 2017; 30: 219-237.
38. Mayerhoefer ME, Materka A, Langs G, Häggström I, Szczypiński P, Gibbs P, et al. Introduction to radiomics. J Nucl Med 2020; 61: 488-495.
39. Troiano G, Fanelli F, Rapani A, Zotti M, Lombardi T, Zhuraki­v­ska K, et al. Can radiomic features extracted from intra-oral radiographs predict physiological bone remodelling around dental implants? A hypothesis-generating study. J Clin Periodontol 2023; 50: 932-941.
40. Larson DB, Chen MC, Lungren MP, Halabi SS, Stence NV, Langlotz CP. Performance of a deep-learning neural network model in assessing skeletal maturity on pediatric hand radiographs. Radiology 2018; 287: 313-322.
41. Kim JR, Shim WH, Yoon HM, Hong SH, Lee JS, Cho YA, et al. Computerized bone age estimation using deep learning based program: evaluation of the accuracy and efficiency. AJR Am J Roentgenol 2017; 209: 1374-1380.
42. Booz C, Yel I, Wichmann JL, Boettger S, Al Kamali A, Albrecht MH, et al. Artificial intelligence in bone age assessment: accuracy and efficiency of a novel fully automated algorithm compared to the Greulich-Pyle method. Eur Radiol Exp 2020; 4: 6. DOI: 10.1186/s41747-019-0139-9.
43. European Society of Radiology (ESR). Impact of artificial intelligence on radiology: a EuroAIM survey among members of the European Society of Radiology. Insights Imaging 2019; 10: 105. DOI: 10.1186/s13244-019-0798-3.
44. Santos GNM, da Silva HEC, Ossege FEL, de Souza Figueiredo PT, de Santos Melo N, Miron Stefani C, et al. Radiomics in bone pathology of the jaws. Dentomaxillofac Radiol 2023; 52: 20220225. DOI: 10.1259/dmfr.20220225.
This is an Open Access journal, all articles are distributed under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0). License (http://creativecommons.org/licenses/by-nc-sa/4.0/), allowing third parties to copy and redistribute the material in any medium or format and to remix, transform, and build upon the material, provided the original work is properly cited and states its license.
Quick links
© 2024 Termedia Sp. z o.o.
Developed by Bentus.