Teledentistry: A Future Solution in the Diagnosis of Oral Lesions: Diagnostic Meta-Analysis and Systematic Review
Introduction
Teledentistry (TD), as part of telemedicine, is an information technology-based diagnosis, treatment, and education delivery system in the dental field.1 The recent rise of the use of TD due to COVID-19 has shown its potential for implementing different solutions to primary dental care.2
TD involves store-and-forward (SAF) solutions, which allow the patient–doctor communication to take part both at different time points (e.g., via a chat application) and in real-time (RT) solutions that require both the patient and the doctor to be simultaneously present in the online space (e.g., videoconference). Through these solutions, the transfer of electronic health records, teleconsultation, telediagnosis, online therapy planning, follow-up, remote patient monitoring, online research, and tele-education can be applied to dentistry.1
As visual information can include important details about a clinical case, a photograph-based telediagnosis depends on image resolution.3 Teledermatology, where evidence for telediagnosis is based on image capturing, is one of the most investigated and developing fields in telemedicine.4 Recent research in teledermatology shows contradictory findings about the diagnostic accuracy of the field,5,6 however, this research has not yet been performed in oral medicine (OM).
The correct diagnosis of premalignant lesions is essential. Due to the fact that oral cancer (OC) in an early stage is asymptomatic, it lacks patient-reported symptoms: ulceration, bleeding, and induration.7 In developing countries such as India and Malaysia where smokeless tobacco (e.g., betel chewing) is popular, the problem arises.8 Other risk factors can include infections, solar radiation, dental hygiene, and genetic factors.8 A low-resource setting and the lack of specialists in rural areas lengthen the diagnosing procedure, allowing the manifestation of oral potentially malignant disorders (OPMDs). TD could help in bridging this problem. Providing high-risk populations with a general oral mucosa screening could decrease the rising amount of OC.
Diagnosing oral lesions (OL) can be difficult due to the variability of presence and the lack of training in OM.9 Detection and follow-up of OPMDs at an early stage could stop their manifestation to OC. Also, the similarities between precancerous lesions and some benign lesions cause difficulties in differential diagnosis.7 Therefore, many unnecessary referrals are made by primary dentists to OM specialists, resulting in wait time and travel difficulties for patients.10
TD has the potential to connect patients with dentists via the online space, thus improving access to care in rural areas.11
This systematic review and meta-analysis aimed to collect available data on how OM could benefit from TD solutions. Due to the lack of data about different TD tools and concise data about the use of TD in OM, an extended summary is needed.
Our aim is to investigate the existing TD tools in the diagnosis of OLs using clinical oral examination (COE) as a reference standard. The main outcome is the presence and the differential diagnostic accuracy of OPMDs, which can lead to lowering the number of referrals to special care.
Methods
Reporting of this systematic review and meta-analysis is based on the recommendation of the PRISMA 2020 guideline,12 while we followed the Cochrane Handbook.12 The protocol of the study was registered on PROSPERO (registration number CRD 42021282645) with no deviations from it.
SEARCH STRATEGY AND ELIGIBILITY CRITERIA
The systematic search was conducted on October 25, 2021, and was revised on October 11, 2022, in the following electronic databases: MEDLINE (via PubMed), Embase, and the Cochrane Central Register of Controlled Trials (CENTRAL). The search key we used in each database is detailed in Supplementary Data S1.
The search applied no filters. Only English language articles were considered eligible. In addition, the reference lists of eligible studies were also searched. Nonoriginal articles (reviews, editorials, letters, and comments) and nonpeer-reviewed articles (gray literature) were excluded. In vitro and animal studies were also excluded.
Study eligibility was determined based on the PICO framework: adults with suspected OLs (P), who were examined using TD tools (that included imaging, I) were acceptable with a reference standard (C) of COE or biopsy. Specificity and sensitivity, positive predictive value (PPV), and negative predictive value (NPV) were considered the included outcomes (O). Regarding the study type, observational studies were considered eligible.
SELECTION PROCESS
The selection was performed by two independent review authors (E.U. and Z.D.). After the duplicates were removed using reference management software (EndNote X9, Clarivate Analytics), the two authors independently screened titles, abstracts, and then full texts. Disagreements were resolved by a third independent reviewer (O.N.).
DATA COLLECTION PROCESS
Data were collected by two authors independently (E.U. and Z.D.) based on a predefined data collection form. The disagreements were resolved by a third independent author (O.N.).
The following data were extracted: first author, year of publication, type of study, country, study population, mean age, surveyed risk factors, number of participants, target condition definition, person conducting telediagnosis, type of photography tool, description of TD tool, and reference standard.
DIAGNOSTIC ACCURACY MEASURES
The comparison of a new diagnostic test (TD) with the gold standard of the field (COE) is conventionally summarized in a two-by-two table called the diagnostic contingency table. It contains the numbers of true-positive (TP), false-positive (FP), true-negative (TN), and false-negative (FN) cases. The most common measures calculated based on the table include sensitivity, specificity, accuracy, and PPV and NPV. All these data were collected in a predefined data table.
ORAL PREMALIGNANT OR MALIGNANT DISEASE
We established a detailed diagnosis classification table based on Haron’s predefined list13 to define OPMD lesions in those articles that have only the final diagnosis listed. We considered the WHO classification of OLs14 as the standard decision base in cases of questionable diagnosis. The final table was discussed with an OM expert (O.N.).
RISK-OF-BIAS ASSESSMENT
Based on the Cochrane Handbook for Systematic Reviews of Diagnostic Test Accuracy,12 we used the QUADAS-2 tool15 for risk-of-bias assessment. This was performed by two review authors (E.U. and Z.D.). Disagreements were resolved by a third independent author (O.N.).
CERTAINTY OF EVIDENCE
Certainty of evidence was evaluated based on the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) workgroup’s recommendations.16 The endpoints of the outcomes were assessed by two independent reviewers (E.U. and Z.D.). In case of disagreement, a third independent reviewer resolved it (O.N.).
We made the GRADE evidence profiles with GRADEpro GDT Software17 for the investigated outcomes.
SYNTHESIS METHODS
Two by two contingency tables were extracted from the studies containing TP, FP, FN, and TN values. Where studies published data corresponding to more than one remote expert, we randomly selected one of them. We repeated the analysis using the contingency tables of previously nonselected experts as a sensitivity analysis.
We used the bivariate model of Chu and Cole and Reitsma et al.18,19 Ellipsoids of this analysis reflect the weights of the studies according to the method by Burke et al.20 Univariate analyses of sensitivity and specificity were performed using the generalized mixed-effect approach.21 Statistical analyses were carried out using the online tool described in Freeman et al.22 and the meta package of R statistical software (version 4.1.2.). The statistical analyses followed the advice of Harrer et al.23 Further details of the synthesis methods are described in Supplementary Data S2.
Results
SEARCH AND SELECTION
In all, 7,608 studies were identified by our systematic search, 13 of which were included in the qualitative synthesis.3,10,13,24–33 By the end of the process, 10 articles were found eligible for the quantitative synthesis,3,10,13,25,26,29–32,34 and 4 articles were excluded due to lack of data.24,27,28,33 The selection process and Cohen’s kappa values are presented using a PRISMA flowchart (Fig. 1).
Fig. 1. PRISMA flowchart.
BASIC CHARACTERISTICS OF INCLUDED STUDIES
The baseline characteristics of the studies included in the qualitative analysis are detailed in Table 1.
| AUTHORS (YEAR) | STUDY SITE | STUDY DESIGN | NO OF PARTICIPANTS | MEAN AGE (YEARS) | INCLUSION CRITERIA | EXCLUSION CRITERIA | RISK FACTOR SURVEYED | TARGET CONDITION DEFINITION | INDEX TEST | REFERENCE STANDARD | ||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| PERSON CONDUCTING REMOTE DIAGNOSIS | TYPE OF PHOTOGRAPHY TOOL | DESCRIPTION OF TELEDENTISTRY TOOL | ||||||||||
| Birura et al. (2015)25 | India | Cohort | 3,440 | 18–85 | Targeted group from two rural villages, opportunistic group from oral surgeon’s screening | N/A | Smoking tobacco, chewing betel leaf/gutka, regular use of alcohol | N/A | Not applicable | Mobile phone | Sana application was designed that is integrated with OpenMRS (medical record system) | Not applicable |
| Birur et al. (2019)24 | India | Cross-sectional | 3,445 | 18–57 | Workers of a pipeline factory | N/A | High risk of tobacco use | WHO | Remote specialist using m-health | Mobile phone | Mobile phone-based questionnaire and photographs of oral cavity | Screening by onsite specialist |
| Flores et al. (2022)34 | Brazil | Cross-sectional | 100 | 51.3 | Patients referred to one of the three included oral medicine centers | N/A | N/A | Exact lesion diagnosis | Remote dentist | Mobile phone (at least 10 MP resolution) or camera | Mobile phone application that includes patient data, and characteristics of the lesion and photographs of the lesion | Expert professor |
| Gomes and colleagues (2017)26 | Brazil | Prospective | 55 | 60–80+ | >40 years old, smoking | N/A | Smoking, alcohol consumption, oral/pharyngeal cancer in family | Grouped lesions (9) | 2 trained examiners (>3 years of experience) | Mobile phone | Newly developed app: videos of oral cavity and in-app data | Examiners with experience in oral diagnosis |
| Fonsecaa et al. (2021)27 | Brazil | Cross-sectional | 113 | 52 | >5 years of age with oral lesion | Difficulties with mouth opening | N/A | Final diagnosis, listed | Oral medicine and oral pathology professors (15 years of experience) | iPhone 5, 8 MP camera | Photographs were e-mailed with clinical implication to evaluators | COE or biopsy |
| Haron et al. (2017)3 | Malaysia | Cross-sectional | 8 | N/A | 8 targeted lesions, 8 normal or variant of normal | Not applicable | Not applicable | OPMD/non-OPMD/normal | Oral medicine specialist | 5–13 MP mobile phone camera | Photographs taken by mobile phone, retrieved later for review | COE by oral medicine specialist |
| Haron et al. (2020)13 | Malaysia | Prospective | 355 | 53.9 | Adults >18 years old referred by primary dentist | N/A | Alcohol consumption, smoking, betel-chewing | Presence of lesion/normal variant, predetermined table | Off-site specialist | 13 MP mobile phone | Using MeMoSa application | COE by oral medicine or oral surgery specialist |
| Namakiana et al. (2012)28 | United States | Observational | 29 | 47 | Patients of Special Care Clinic: intellectual disability, cerebral palsy, Down syndrome, autism, seizures, HIV, liver disease, neurologic disorders, stroke, schizophrenia | If patient was not able to cooperate for record collection (at least intraoral and extraoral photographs) | N/A | Community-based treatment or referral to dentist | Study dentists | Intraoral camera, point-and-shoot camera | COE followed by an evaluation questionnaire, after 3-week-washout period a virtual examination was conducted | Study dentists |
| Perdoncini et al. (2021)29 | Brazil | Cross-sectional | 33 | 53 | >18-year-old patients referred to the clinic with oral lesion | Symptoms but no oral sign | N/A | Diagnosis per lesion | Oral medicine specialist | iPhone SE | The dentist sent the photographs to specialist via WhatsApp, and a video call was initialized | Oral medicine specialist |
| Petruzzi and De Benedittis (2016)10 | Italy | Cross-sectional | 96 | N/A | Referred patients by general dentists/oral hygienist/physician OR patient | N/A | N/A | Traumatic, infective, preneoplastic/neoplastic, autoimmune, not diagnosable | Oral medicine specialist | Mobile phone | Photographs sent via WhatsApp, and after clinical examination was made. Biopsy was made where needed. | Oral medicine specialist |
| Tesfalul et al. (2016)30 | United States | Observational | 23 | 45.5 | Referred adult patients with complicated oral lesion, receiving care via the mobile oral telemedicine system | <18 years, incomplete telemedicine consultation, failure to obtain consent | N/A | Cancer, infection, fracture, benign mass, dermatologic conditions, other | Specialist | HTC myTouch mobile phone | Mobile telemedicine application including image-sharing and clinical data | COE conducted by dentist |
| Torres-Pereira et al. (2008)31 | Brazil | Cross-sectional | 25 | N/A | People with oral lesion | People without oral lesion | N/A | Predefined list of terms | Oral medicine specialist (10 years of experience) | FUJI S7000 digital camera | Images sent via e-mail, with an electronic form containing patient’s data | COE by oral medicine specialist or biopsy when needed |
| Torres-Pereira et al. (2013)32 | Brazil | Cross-sectional | 60 | N/A | People with oral lesion | N/A | N/A | Predefined list of terms | Oral medicine specialist | Canon EOS 300 Rebel digital camera | Images sent via e-mail, with an electronic form containing patient’s data | Biopsy |
| Vinayagamoorthya et al. (2019)33 | India | Cross-sectional | 131 | 37.34 | Patients included from oral screening programs | Having problem with comprehension, limited mouth opening | Medical and habit history | Normal/abnormal, exact lesion diagnosis | Trained and calibrated examiner | Samsung mobile phone | A set of five photographs were made, then sent to specialist via WhatsApp | Trained and calibrated examiner |
PRIMARY OUTCOMES
Oral lesion detection with TD tools
Our pooled analysis of 3 articles3,13,24 including 3,783 patients showed that TD tools can be a reliable option for replacing face-to-face dental visits in the detection of OLs (sensitivity: 0.92 confidence interval [CI] = 0.59–0.99; specificity: 0.93 CI = 0.17–1.00) (Fig. 2). The data used for this analysis are shown in Supplementary Table S1. The examined outcome showed a substantially heterogenous population (sensitivity: I2 = 84%, p < 0.01; specificity: I2 = 98%, p < 0.01). Fitting the bivariate model was not possible in this analysis due to the small number of available studies.
Fig. 2. Sensitivity and specificity of diagnosing the presence of oral lesions with teledentistry tools. CI, confidence interval; TP, true positive; TN, true negative.
Diagnosing oral premalignant lesions or oral cancer
As the exact number of TP, TN, FP, and FN lesions was not mentioned in every article, the abovementioned classification table was able to help us resolve the problem, as can be seen in Supplementary Table S2.
We assessed the available data (Supplementary Table S3) for the diagnosis of OPMD or OC lesions. Estimated high sensitivity and specificity of nine pooled articles3,10,13,26,29–32,34 showed that TD can be a reliable tool for differential diagnosis of OLs (sensitivity: 0.923 CI = 0.835–0.966; specificity: 0.987 CI = 0.947–0.997), as shown in Figure 3 and Supplementary Figure S1. Heterogeneity is low (the size of the prediction region is moderate; sensitivity: I2 = 0%, p = 0.66; specificity: I2 = 0%, p = 1.00), as shown in Supplementary Figure S1.
Fig. 3. Sensitivity and specificity of diagnosing the oral premalignant/malignant lesions with teledentistry tools-2D analysis.
Conducting sensitivity analysis, the resulting sensitivity and specificity are 0.953 (CI = 0.856–0.986) and 0.983 (CI = 0.930–0.996), respectively. See Supplementary Data for further details of the extractions (Supplementary Table S4). This result confirmed the findings of the main analysis; thus, the outcome shows that the results do not depend on the examiner’s knowledge (Supplementary Figs. S2 and S3).
SECONDARY OUTCOMES
Time-effectiveness
No statistical analysis could be performed on time-effectiveness due to the heterogeneity of reported data. Namakian et al.28 measured the difference in time during in-person examination (mean: 4.2 min, SD: 1.6) and virtual examination (2.83 min, SD: 1.0). Perdoncini et al.29 listed the transportation time (mean: 58.3 min, SD: 52.9), which shows that mostly people from rural areas attended the clinic. The COE took 10.44 min on average (SD: 2.88). In the case of the WhatsApp messaging platform, Petruzzi and De Benedittis10 measured an average distance of 70 km (SD: 78 km).
Screening person
Birur et al.25 showed that a trained frontline health care worker could detect OPMD with a PPV of 45%. They also studied the efficacy of community health care workers (CHWs) in the diagnosis of the presence of OL, where they found a kappa score of 0.92 between CHWs and onsite specialists.24 Therefore, not only dentists but also trained CHWs could take part in the detection of OLs.
Referral decision
The referral decision was measured in four articles,3,13,27,28 however, its policy showed variability. Fonseca et al.27 recorded that an average of 35.4% of the examined patients could be treated in primary dental care. Haron et al.13 recorded a kappa value of 0.892 (CI: 0.843–0.940) for referrals using a predefined list.
Technical setting
The technical setting of TD tools varied in the available articles. Haron et al.3 found a correlation between using better resolution mobile phones as photograph-taking devices in diagnosing lesion category: the higher the resolution, the higher the diagnostic agreement (accuracy: 0.515–0.881). Using a better mobile phone helped not only in better detection (sensitivity: 0.813), but also in correct diagnosis (sensitivity: 0.93).
Perdoncini et al.29 investigated the accuracy of RT video consultation with a specialist; 91% of the photographs were of good quality, and the internet connection was stable in 58% of the cases.
A professional camera was used in two articles31,32 to simulate primary dental care settings. The images then were sent to the specialist via e-mail, thus avoiding internet connection problems.
Receiving photographs and testing the role of free chat applications in dentistry can simulate population-based screening.10,29,33 Not only dentists, but also dental hygienists and patients sent homemade photographs of OLs. WhatsApp was tested by Vinayagamoorthy et al.,33 where a sensitivity of 0.981 (examiner 1) and 0.987 (examiner 2) was found for exact lesion diagnosis.
RISK-OF-BIAS ASSESSMENT
Using the QUADAS-2 tool,15 our results for risk-of-bias assessment are presented in Supplementary Figures S4, S5, and S6. Most of the domains showed low risk of bias, however, having inappropriate exclusion criteria in patient selection in one article meant that only people with risk factors were involved in the actual study,27 and in another study, the patients were selected by having/not having OL/OPMD/OC.3 The index test result domain has unclear risk in those articles where COE (index test) was made by the same person who made the telediagnosis with a fixed, predefined washout period.27,28,30
Due to the small number of available studies, publication bias could not be assessed by performing the methodology of Deeks et al.35 Certainity of evidence Table can be seen in Supplementary Table S5. The PRISMA checklist can be seen in Supplementary Table S6.
Discussion
According to the high sensitivity and specificity in our study, TD could successfully help in primary dental care in the field of OM. Store-and-forward systems could be used in telediagnosis (e.g., photographs sent via e-mail), and RT solutions could be used in teleconsultation (e.g., video consultation).29
Visual examination is extremely important (and remains the primary diagnostic tool) in both dermatology and OM, causing a rapid technical rise in both teledermatology and TD.36 Image-based diagnosis by a specialist may help primary dental care facilities in the decision of referral to special care.
Also, TD could promote organizing population-based OL screening programs in the future.25 The differential diagnosis of OLs is important due to treatment planning reasons. TD could support general dentists in OL management through connection with OM specialists, enabling early diagnosis and correct treatment of oral (pre)malignant lesions.34,36,37 Knowing the best available treatment and prevention could lead to less manifestation of OC. Furthermore, referring only OPMDs and malignant OLs might result in a higher percentage of treated patients.
As the prevalence of OC is rising, the correct diagnosis of premalignant lesions is essential. In developing countries, low-resource settings and the lack of specialists in rural areas make the diagnosing procedure longer, allowing manifestations of OPMD. TD could help in solving this problem.
TD tools in the diagnosis of OLs show high accuracy, however, there are weaknesses in the available data that need to be mentioned. The protocol of diagnosis by the same person was followed by several articles, which despite the washout time (average 3 weeks) still brings into question the reliability and correctness of the results.3,13,28,33 Also, in one case, the general dentist could send his or her own diagnostic opinion, compromising the impartiality of the diagnosis.30
Education in the field of OM starts during the predoctoral phase, however, it varies afterward. A huge difference can be spotted regarding specialists: for example, in the United States, this field is a recognized dental specialty of the American Dental Association,38 however, in Europe, it varies from country to country, and a consensus has not been made yet about the competencies and limits of OM specialists.39 This fact questions the use of the term OM “specialist” in the articles. In addition, experience in the field does not guarantee a correct diagnosis using TD tools. Lack of information in the online form and the lack of in-person personalized questions may complicate the diagnosing process.
Referring every patient with an OL to special care might overload OM praxis. A high number of referrals may be avoidable using a pretriage system40 or population-based screening.24 This could increase the efficacy of patient flow in OM practices.
OL DETECTION
OC is unique compared with other cancers because visual detection at an early stage is possible with COE,41 however, we cannot predict the level of dysplasia based on clinical images. Diagnosing the presence and type of OL could lead to targeted treatment and prevention of manifestation to OC.41 Our results showed both high specificity and sensitivity in the detection of OLs, meaning that TD tools could be used as a pretriage system in the diagnosis of OLs.
DIFFERENTIAL DIAGNOSIS OF OLs
Correct diagnosis of an OL is necessary before treatment, however, not all patients need to be referred to an OM specialist. Our study showed high sensitivity, meaning that TP OPMD could be successfully diagnosed via TD image-sharing tools (e-mail, free chat applications, cloud-based storage applications, etc.). We found high specificity, meaning that TN lesions that do not necessitate referral to a specialist are correctly diagnosed, affirming previous articles’ clinical data.3,13
STRENGTHS
According to our knowledge, this is the first meta-analysis on this specific topic. We used rigorous methodology and followed a previously registered protocol. We selected for eligibility only those articles in which the clinical diagnosis (reference standard) was made only by doctors. Due to the strict inclusion criteria, we gained higher data quality (e.g., we excluded articles that did not include a reference standard). We did not include publications with retrospective or case–control designs.
LIMITATIONS
Regarding the limitations of this article, within the available data regarding the detection of OLs, we found significant heterogeneity among the included articles. The reason for this, despite the strict inclusion criteria, could be the included heterogenous population (e.g., a preselected population or factory workers). The methodology of applied TD tools was heterogeneous, meaning that RT and SAF systems were pooled together. The referral decision was not defined in each of our selected articles, and so, we could not include it in a quantitative analysis.
IMPLICATION FOR PRACTICE
A future photograph-taking protocol for OLs could be used in primary dental care. We suggest establishing the technical setting first, as telecommunication tools require a firm technical background. Also, a standardized questionnaire could be used in the future to provide more details about the patient.
IMPLICATION FOR RESEARCH
We suggest that researchers use biopsy as a reference standard (replacing COE) to have an objective diagnosis. Furthermore, exact lesion diagnosis could be helpful in comparison with other publications’ results, and referral decisions could be counted based on these data.
Conclusions
Detecting OLs with TD tools could lead to early diagnosis, treatment, and stricter follow-up of OPMDs. TD offers a great substitute for face-to-face dental visits in the detection and differential diagnosis of OLs, and thus, fewer referrals could be made to special OM care. Being a time-effective solution, when conducted in a proper way, access to care could be widened with the use of these tools.
Authors’ Contributions
E.U. and O.N.: Conceptualization, E.U. and Z.D.: Data curation, T.K.: Formal analysis, P.H.: Funding acquisition, E.U.: Investigation, P.H. and L.M.C.: Methodology and project administration, O.N.: Resources and validation, E.U. and T.K.: Visualization, E.U.: Writing—original draft preparation, O.N., Z.D., L.M.C., T.K., J.B., B.G.N.C., and P.H.: Writing—review and editing. All the authors provided critical conceptual input and approved the final version of the article.
Ethical Approval
No ethical approval was required for this systematic review and meta-analysis, as all data were already published in peer-reviewed journals. No patients were involved in the design, conducting, or interpretation of our study. The data sets used in this study can be found in the full-text articles included in the systematic review and meta-analysis.
Disclosure Statement
No competing financial interests exist.
Funding Information
Funding was provided by the Ministry of Innovation and Technology of Hungary from the National Research, Development, and Innovation Fund (ITM-NRDIF) (TKP2021-EGA-23). The study’s funder had no role in study design, data collection, data analysis, data interpretation, or the report’s writing.
Supplementary Material
References
- 1. . Teledentistry in practice: Literature review. Telemed J E Health 2013;19(7):565–567; doi:
10.1089/tmj.2012.0200 Link, Google Scholar - 2. . Impact of COVID-19 on the knowledge and attitudes of dentists toward teledentistry. JDR Clin Trans Res 2021;6(3):268–278; doi:
10.1177/2380084421998632 Crossref, Medline, Google Scholar - 3. Mobile phone imaging in low resource settings for early detection of oral cancer and concordance with clinical oral examination. Telemed J E Health 2017;23(3):192–199; doi:
10.1089/tmj.2016.0128 Link, Google Scholar - 4. Synchronous and asynchronous teledermatology: A narrative review of strengths and limitations. J Telemed Telecare 2022;28(7):533–538; doi:
10.1177/1357633X221074504 Crossref, Medline, Google Scholar - 5. , 3rd. Teledermatology: A review and update. Am J Clin Dermatol 2018;19(2):253–260; doi:
10.1007/s40257-017-0317-6 Crossref, Medline, Google Scholar - 6. Live interactive teledermatology compared to in-person care—A systematic review. J Eur Acad Dermatol Venereol 2020;34(4):733–745; doi:
10.1111/jdv.16070 Crossref, Medline, Google Scholar - 7. Early diagnosis of oral cancer. Cancer 1988;62(8 Suppl):1796–1799; doi:
10.1002/1097-0142(19881015)62:1+<1796::aid-cncr2820621319>3.0.co;2-e Crossref, Medline, Google Scholar - 8. Oral cancer: Etiology and risk factors: A review. J Cancer Res Ther 2016;12(2):458–463; doi:
10.4103/0973-1482.186696 Crossref, Medline, Google Scholar - 9. Dentists’ knowledge and opinions about oral mucosal lesions. Int J Oral Maxillofac Surg 2009;38(12):1283–1288; doi:
10.1016/j.ijom.2009.07.004 Crossref, Medline, Google Scholar - 10. . WhatsApp: A telemedicine platform for facilitating remote oral medicine consultation and improving clinical examinations. Oral Surg Oral Med Oral Pathol Oral Radiol 2016;121(3):248–254; doi:
10.1016/j.oooo.2015.11.005 Crossref, Medline, Google Scholar - 11. A systematic review of the research evidence for the benefits of teledentistry. J Telemed Telecare 2018;24(3):147–156; doi:
10.1177/1357633X16689433 Crossref, Medline, Google Scholar - 12. Cochrane Handbook for Systematic Reviews of Interventions version 6.3 2022. Cochrane, Available from: www.training.cochrane.org/handbook [Last accessed: February, 2022]. Google Scholar
- 13. m-Health for early detection of oral cancer in low- and middle-income countries. Telemed J E Health 2020;26(3):278–285; doi:
10.1089/tmj.2018.0285 Link, Google Scholar - 14. A digital manual for the early diagnosis of oral neoplasia. Available from: https://screening.iarc.fr/atlasoral.php [Last accessed: August 6, 2022]. Google Scholar
- 15. QUADAS-2: A revised tool for the quality assessment of diagnostic accuracy studies. Ann Intern Med 2011;155(8):529–536; doi:
10.7326/0003-4819-155-8-201110180-00009 Crossref, Medline, Google Scholar - 16. GRADE handbook for grading quality of evidence and strength of recommendations. Updated October 2013. The GRADE Working Group, 2013. Available from: guidelinedevelopment.org/handbook. 2013. Google Scholar
- 17. : GRADEpro Guideline Development Tool [Software]. McMaster University and Evidence Prime Afgo. 2021. Available from: https://www.gradepro.org/. Google Scholar
- 18. . Bivariate meta-analysis of sensitivity and specificity with sparse data: A generalized linear mixed model approach. J Clin Epidemiol 2006;59(12):1331–1332; doi:
10.1016/j.jclinepi.2006.06.011 Crossref, Medline, Google Scholar - 19. Bivariate analysis of sensitivity and specificity produces informative summary measures in diagnostic reviews. J Clin Epidemiol 2005;58(10):982–990; doi:
10.1016/j.jclinepi.2005.02.022 Crossref, Medline, Google Scholar - 20. Guidance for deriving and presenting percentage study weights in meta-analysis of test accuracy studies. Res Synth Methods 2018;9(2):163–178; doi:
10.1002/jrsm.1283 Crossref, Medline, Google Scholar - 21. . Random effects meta-analysis of event outcome in the framework of the generalized linear mixed model with applications in sparse data. Stat Med 2010;29(29):3046–3067; doi:
10.1002/sim.4040 Crossref, Medline, Google Scholar - 22. Development of an interactive web-based tool to conduct and interrogate meta-analysis of diagnostic test accuracy studies: MetaDTA. BMC Med Res Methodol 2019;19(1):81; doi:
10.1186/s12874-019-0724-x Crossref, Medline, Google Scholar - 23. Doing Meta-Analysis with R A Hands-on Guide. CRC Press, Taylor & Francis Group; 2022. Google Scholar
- 24. Role of community health worker in a mobile health program for early detection of oral cancer. Indian J Cancer 2019;56(2):107–113; doi:
10.4103/ijc.IJC_232_18 Crossref, Medline, Google Scholar - 25. Mobile health application for remote oral cancer surveillance. J Am Dent Assoc 2015;146(12):886–894; doi:
10.1016/j.adaj.2015.05.020 Crossref, Medline, Google Scholar - 26. Development of a mobile application for oral cancer screening. Technol Health Care 2017;25(2):187–195; doi:
10.3233/THC-161259 Crossref, Medline, Google Scholar - 27. Telediagnosis of oral lesions using smartphone photography. Oral Dis 2021; doi:
10.1111/odi.13972 Crossref, Google Scholar - 28. In-person versus “virtual” dental examination: Congruence between decision-making modalities. J Calif Dent Assoc 2012;40(7):587–595. Crossref, Medline, Google Scholar
- 29. Use of smartphone video calls in the diagnosis of oral lesions: Teleconsultations between a specialist and patients assisted by a general dentist. J Am Dental Assoc 2021;152(2):127–135; doi:
10.1016/j.adaj.2020.10.013 Crossref, Medline, Google Scholar - 30. Evaluating the potential impact of a mobile telemedicine system on coordination of specialty care for patients with complicated oral lesions in Botswana. J Am Med Inform Assoc 2016;23(e1):e142–e145; doi:
10.1093/jamia/ocv140 Crossref, Medline, Google Scholar - 31. Email for distance diagnosis of oral diseases: A preliminary study of teledentistry. J Telemed Telecare 2008;14(8):435–438; doi:
10.1258/jtt.2008.080510 Crossref, Medline, Google Scholar - 32. Teledentistry: Distant diagnosis of oral disease using e-mails. Telemed J E Health 2013;19(2):117–121; doi:
10.1089/tmj.2012.0087 Link, Google Scholar - 33. Efficacy of a remote screening model for oral potentially malignant disorders using a free messaging application: A diagnostic test for accuracy study. Aust J Rural Health 2019;27(2):170–176; doi:
10.1111/ajr.12496 Crossref, Medline, Google Scholar - 34. Diagnostic accuracy of a telediagnosis service of oral mucosal diseases: A multicentric survey. Oral Surg Oral Med Oral Pathol Oral Radiol 2022;134(1):65–72; doi:
10.1016/j.oooo.2022.02.005 Crossref, Medline, Google Scholar - 35. . The performance of tests of publication bias and other sample size effects in systematic reviews of diagnostic test accuracy was assessed. J Clin Epidemiol 2005;58(9):882–893; doi:
10.1016/j.jclinepi.2005.01.016 Crossref, Medline, Google Scholar - 36. Using teledentistry in clinical practice as an enabler to improve access to clinical care: A qualitative systematic review. J Telemed Telecare 2018;24(3):129–146; doi:
10.1177/1357633X16686776 Crossref, Medline, Google Scholar - 37. Telediagnosis of oral lesions in primary care: The EstomatoNet Program. Oral Dis 2018;24(6):1012–1019; doi:
10.1111/odi.12851 Crossref, Medline, Google Scholar - 38.
American Dental Association’s recognized specialities . Available from: https://ncrdscb.ada.org/en/dental-specialties/specialty-definitions [Last accessed: October 5, 2022]. Google Scholar - 39. . Oral medicine in Europe: Past, present and future. Br Dent J 2017;223(9):726–728; doi:
10.1038/sj.bdj.2017.891 Crossref, Medline, Google Scholar - 40. A proof-of-concept evaluation of a cloud-based store-and-forward telemedicine app for screening for oral diseases. J Telemed Telecare 2016;22(6):319–325; doi:
10.1177/1357633X15604554 Crossref, Medline, Google Scholar - 41. . Exciting new advances in oral cancer diagnosis: Avenues to early detection. Head Neck Oncol 2011;3:33; doi:
10.1186/1758-3284-3-33 Crossref, Medline, Google Scholar