Professional Documents
Culture Documents
Ai Assisted Ceph
Ai Assisted Ceph
Ai Assisted Ceph
Introduction: Artificial intelligence (AI) developed rapidly in orthodontics, and AI-based cephalometric
applications have been adopted. This study aimed to assess AI-assisted cephalometric technologies related
knowledge, experience, and attitude among orthodontists and orthodontic students; describe their subject view
of the applications and related technologies in orthodontics; and identify associated factors. Methods: An online
cross-sectional survey based on a professional tool (www.wjx.cn) was performed from October 11-17, 2022.
Participants were recruited with a purposive and snowball sampling approach. Data was collected and
analyzed with descriptive statistics, chi-square tests, and multivariable generalized estimating equations.
Results: Four hundred eighty valid questionnaires were collected and analyzed; 68.8% of the respondents agreed
that AI-based cephalometric applications would replace manual and semiautomatic approaches. Practitioners
using AI-assisted applications (87.5%) spent less time in cephalometric analysis than the other groups using
other approaches, and 349 (72.7%) respondents considered AI-based applications could assist in obtaining
more accurate analysis results. Lectures and training programs (56.0%) were the main sources of respondents’
knowledge about AI. Knowledge level was associated with experience in AI-related clinical or scientific projects
(P\0.001). Most respondents (88.8%) were interested in future AI applications in orthodontics. Conclusions: Re-
spondents are optimistic about the future of AI in orthodontics. AI-assisted cephalometric applications were
believed to make clinical diagnostic analysis more convenient and straightforward for practitioners and even
replace manual and semiautomatic approaches. The education and promotion of AI should be strengthened to
elevate orthodontists’ understanding. (Am J Orthod Dentofacial Orthop 2023;164:e97-e105)
A
rtificial intelligence (AI) is an emerging technol- natural language processing, computer vision, and so
ogy based on computer science to develop ma- on.3 The development of precision medicine has been
chines that mimic intelligent behavior with greatly promoted by the stunning advancements of AI
little assistance from humans.1,2 AI generally encom- in recent decades.4-6 Similarly, orthodontic research
passes various technologies, including machine learning, has become focused on the application of AI,
a
State Key Laboratory of Oral & Maxillofacial Reconstruction and Regeneration, This study was reviewed and approved by the Ethics Committee of School and
Key Laboratory of Oral Biomedicine Ministry of Education, Hubei Key Laboratory Hospital of Stomatology, Wuhan University (no. 2022-B58). Informed consent
of Stomatology, School & Hospital of Stomatology, Wuhan University, Wuhan, was obtained from all individual participants included in the study.
China. This work was supported by the Undergraduate Education Quality Construction
b
Department of Orthodontics, School and Hospital of Stomatology, Wuhan Uni- and Comprehensive Reform Project, Wuhan University (no. 2021ZG328), and the
versity, Wuhan, China. Chinese Stomatological Association Chinese Orthodontic Society Basic Research
c
Center for Dentofacial Development and Sleep Medicine, School and Hospital of Fund (no. COS-B2021-08).
Stomatology, Wuhan University, Wuhan, China. Address correspondence to: Fang Hua, Center for Orthodontics and Pediatric
d
Center for Orthodontics and Pediatric Dentistry at Optics Valley Branch, School Dentistry at Optics Valley Branch, School and Hospital of Stomatology, Wuhan
and Hospital of Stomatology, Wuhan University, Wuhan, China. University, Luoyu Rd 237, Wuhan 430079, China; e-mail, huafang@whu.edu.
e
Center for Evidence-Based Stomatology, School and Hospital of Stomatology, cn or Hong He, Department of Orthodontics, School and Hospital of Stomatol-
Wuhan University, Wuhan, China. ogy, Wuhan University, Luoyu Rd 237, Wuhan 430079, China.; e-mail,
f
Division of Dentistry, School of Medical Sciences, Faculty of Biology, Medicine drhehong@whu.edu.cn.
and Health, The University of Manchester, Manchester, United Kingdom. Submitted, November 2022; revised and accepted, July 2023.
Lizhuo Lin and Bojun Tang are joint first authors and contributed equally to this 0889-5406/$36.00
work. Ó 2023 by the American Association of Orthodontists. All rights reserved.
All authors have completed and submitted the ICMJE Form for Disclosure of Po- https://doi.org/10.1016/j.ajodo.2023.07.006
tential Conflicts of Interest, and none were reported.
e97
e98 Lin et al
particularly machine learning techniques in diagnosis, survey guideline.34 The Informed Consent form in the
treatment planning, and decision-making optimiza- first section of the questionnaire included the study ob-
tion,7-9 including airway analysis,10-13 extraction jectives, the target population, and the principle of
decision,14-16 orthognathic surgery intervention,17-19 voluntariness and included the contact information of
and so on, to improve the efficiency and avoid bias of the investigator. The survey was anonymous, with all en-
inexperienced practitioners.20 crypted data stored on one investigator’s computer. Par-
Cephalometric analysis is universally adopted in the ticipants were specifically informed that there were no
clinical process of orthodontic and orthognathic prac- rewards or penalties for participating in the survey.
tice.21 The manual detection of landmarks via conven- They could withdraw anytime and were considered
tional and semiautomated approaches on x-ray images voluntarily participating once submitted.
is time-consuming and affected by clinicians’ experience, According to previous studies,30-32 the questionnaire
resulting in compromised accuracy.22 Recent studies have was developed and divided into 4 sections with 37 items
shown remarkable progress in AI-based landmark locali- (Supplementary Material). The first section was about
zation. Previous research extensively employed deep the background information of respondents. Questions
learning network models within machine learning, such regarding pertinent knowledge were set in the second
as convolutional neural networks, to accomplish auto- section, which included 6 items, with 1 Likert-scale
mated cephalometric analysis successfully.23-26 item (no. 15) assessing subjects’ self-perceived familiar-
Undoubtedly, AI, an emerging but rapidly developing ity with AI technologies. Five questions on the funda-
technology, can generate an all-around promotion of or- mental knowledge of AI-assisted cephalometric
thodontic services.8 Many commercially available cephalo- analysis and other AI-based applications in orthodontics
metric analyses software using AI-related technologies are are provided to evaluate the respondents’ actual knowl-
already emerging in clinical practice. However, because of edge level (2 single-choice and 3 multiple-choice, nos.
the potential technology inaccuracies in diagnosis and 16-20). To investigate the experience of respondents,
clinical decision-making at this stage, AI-based applica- the items in the third section focused on their percep-
tions in the medical field are still taken cautiously. Data tions of AI-based cephalometric applications. Further-
curation and sharing should be further promoted.27-29 more, a few questions in the fourth section (nos. 32,
With the swift advancement of AI technologies in vi- 33, and 35) were about the behavior of these applica-
sual object recognition detection, several surveys have tions. Other items of the fourth section were about par-
investigated the attitudes of medical professionals and ticipants’ attitudes toward their preferences for using AI-
students toward AI technologies, especially in radi- based applications in clinical practice and their expecta-
ology.30-33 However, there is no relevant study in tions for future AI development.
orthodontics. As a powerful assistant tool, is AI-related The questions were designed to be logically linked to
software likely to replace health care workers in certain make the questionnaire more concise and relevant when
tasks? Because AI-assisted cephalometric applications filled out. Based on the particular options selected by re-
have been used in orthodontic clinical practice, this study spondents, it is determined whether subsequent specific
focuses on this aspect and aims to investigate related questions were displayed or skipped. For example, in
knowledge, experience, and attitude of orthodontists question no. 26, respondents were asked whether they
and orthodontic students and describe their view of AI- would like to use AI-assisted cephalometric applications
based applications and associated factors. In addition, in clinical practice; depending on their answer (ie, yes or
this study was suitably extended to explore their experi- no), respondents would be directed to the question
ence and expectations for the future development of AI asking for their reasons for this attitude. Before the final
in orthodontics to provide suggestions for enhancements release of the questionnaire, it was sent to a panel of 7
to current AI applications and future directions of multi- orthodontists, 7 orthodontic students, and 1 graduate
disciplinary research in AI and orthodontics. student in computer science for a pilot survey. The pilot
survey intends to find potential ambiguities and ensure
the purpose of every item is clear and that there are no
MATERIAL AND METHODS flaws in the logic and structure of the questionnaire. In
This study was an anonymous online survey with a addition, the approximate time required to respond to
convenience sample. Ethical approval was granted by all items was determined to be 5 minutes. After the pilot
the Ethics Committee of the School and Hospital of Sto- survey, a few items were added or subtracted on the basis
matology, Wuhan University (no. 2022-B58). The ques- of the feedback, and the objectives of the survey and ex-
tionnaire was written on the basis of the CHEcklist for pressions that may have been semantically unclear were
Reporting Results of Internet E-Surveys, an online revised.
October 2023 Vol 164 Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
Lin et al e99
The questionnaire was uploaded to an online survey 22-64 years, with a mean age of 32.5 6 8.9, of which
platform (www.wjx.cn), which could ensure the 317 were females (66.0%) and 163 were males (34.0%).
completeness of the submitted questionnaires. The pur- For the professional attributes of the respondents,
posive and snowball sampling techniques were used for orthodontists accounted for more than half (62.1%),
participant recruitment.35 The questionnaire link was and the proportion of orthodontic students was
disseminated on the WeChat public platform, China’s 34.0%. Most respondents (97.7%) reported that they
biggest social media platform. Then the link was sent routinely used cephalometric analysis in their clinical
to orthodontic chat groups, and orthodontic practi- practice, with 62.0% using semiautomated
tioners had to invite their students and friends engaging cephalometric tracing most frequently, followed by AI-
in orthodontics to fill out the questionnaire. automated tracing (22.2%). Only 22.1% of the
The distribution of this survey was from October 11-17, participants considered themselves familiar with AI tech-
2022. The questionnaire was accessed through WeChat nologies, whereas those who had never engaged in
and could only be submitted once from each account. AI-related clinical or scientific projects accounted for
Questionnaires with apparent errors were excluded, such 88.5% of the overall respondents.
as contradictory answers, abnormal age, and anomalous Of the 584 participants who routinely used cephalo-
response times (\2 minutes or .30 minutes).36 metric aids in orthodontic clinical practice, the propor-
tion of those who took \5 minutes to perform lateral
Statistical analysis cephalometric tracing was 55.9%. Table II shows that
The collected data were all downloaded from the on- the proportion of respondents who traced a lateral ceph-
line platform and coded by 2 authors (L.L. and B.T.) alogram for \5 minutes differed significantly between
independently, and as planned a priori, invalid data the groups using different tracing approaches. The
were excluded. All discrepancies were resolved by discus- largest percentage used AI-automated tracing (87.5%),
sion. For the knowledge section (nos. 16-20), each ques- followed by semiautomatic tracing (48.1%) and manual
tion was worth 1 point if answered correctly. Then the tracing (32.6%).
total knowledge score was calculated (score range, 0-5). The self-evaluation score for AI familiarity was 2.89
The general information of respondents and the an- 6 0.94 (range, 0-5), whereas the mean total knowledge
swers to each question were described by descriptive score was 1.54 6 1.19. As shown in Figure 1, lectures
statics. For all Likert-scale items, the score ranged and training programs (56.0%) and academic confer-
from 1 (completely disagree) to 5 (completely agree). In- ences (55.6%) are the main sources of knowledge for
formation related to respondent demographic character- the respondents, followed by journal literature
istics and AI-related scientific experience and behaviors (43.5%), Web sites (35.6%), people around (35.6%), so-
were described in frequencies and percentages. Chi- cial media (30.6%), undergraduate or graduate courses
square and Kruskal-Wallis tests were used to analyze (33.8%) and books (26.0%).
categorical and skewed data. Table III presents the results of GEE regression ana-
To explore the associated factors to the self- lyses for the self-evaluation level for AI technologies fa-
perceived level of familiarity with AI and AI-based appli- miliarity and total knowledge score. Results indicated
cations and total knowledge score, generalized esti- self-perceptions were associated with workplace settings
mating equations (GEE) regression analyses were and AI-related clinical or scientific project experience.
performed. Predictors with tolerance \0.1 or variance Respondents who had never been involved in AI-
inflation factor .10 were removed from the final model related projects (B 5 0.685, 95% confidence intervals,
to detect multicollinearity. P \0.05 was regarded as sta- 0.949 to 0.421, P \0.001) had significantly lower
tistically significant. self-perceptions of AI technology familiarity. In contrast,
those whose workplaces were private (P 5 0.009) had
significantly higher self-ratings. According to GEE
RESULTS regression analyses, the total knowledge score of AI
Within the set survey period, 498 questionnaires were and AI-assisted cephalometric analysis was only signifi-
collected. Eighteen invalid questionnaires were excluded cantly lower among the participants who had no experi-
according to the previously developed exclusion inclu- ence in AI-related clinical or scientific projects (B 5
sion criteria, and finally, 480 valid questionnaires were 0.676, 95% CI, 1.058 to 0.294, P \0.001).
included for analysis. The questionnaires came from 32 Orthodontic professionals in the orthodontic clinic
of China’s 34 provincial-level administrative regions. have adopted several AI-assisted cephalometric analysis
Table I shows the demographic information of the re- applications. Most respondents (96.5%) had heard of AI-
spondents. The 480 respondents ranged in age from automated cephalometric applications and had
American Journal of Orthodontics and Dentofacial Orthopedics October 2023 Vol 164 Issue 4
e100 Lin et al
October 2023 Vol 164 Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
Lin et al e101
Table II. Time spent by respondents with different approaches to cephalometric analysis
Personal experience, Total Manual tracing Semiautomated tracing AI-automated tracing
times/min (n 5 438) (n 5 43) (n 5 291) (n 5 104) P value
\5 245 (55.9) 14 (32.6)a 140 (48.1)a 91 (87.5)b
6-10 115 (26.3) 10 (23.3)a, b 95 (32.6)b 10 (9.6)a \0.0001
11-15 45 (10.3) 10 (23.3)a 33 (11.3)a 2 (1.9)b
.15 33 (7.5) 9 (20.9)a 23 (7.9)b 1 (1.0)c
Note. Values are presented as n (%). Groups with the same superscripted letters (a, b, c) in the same row are not statistically different (P .0.05)
according to post hoc tests.
Fig 1. The sources of AI-related knowledge among respondents (semiopen question, no. 21) (n 5
480).
As one of the most crucial tools for orthodontic diag- developing model that expert detection, machine
nosis and decision-making, cephalometric measure- learning.40-42 In contrast, machine learning is not
ments continue to maintain a unique role.21 optimally consistent; there are differences in landmark
Semiautomatic cephalometric analysis programs have localization from one research and development team to
substantially increased efficiency and usability by auto- another, and they lack an adequate data basis for their
mating the measuring process, largely replacing the applications. Users may frequently not fully be convinced
traditional manual approach involving a viewing box by the results of automated detection because of their
and acetate tracing paper.39 Currently, several commer- own experience. Therefore, sharing, interoperability, and
cial AI automatic cephalometric software or platforms standardization management of huge volumes of data
are available. The results of this study indicated that may be key strategies to leapfrog the landmark detection
around one fifth of participants frequently used the accuracy challenge. The top 3 brands applied by
AI-based cephalometric approach, a more recent inno- respondents (iOrtho, Uceph, Digident) were all locally
vation, whereas the other three fifths were still habitu- owned in China, which might be influenced by racial
ated to semiautomatic cephalometric measures. disparities, marketing access, the accessibility to after-
The results also revealed that time-saving was a big sales services, publicity, and data security.
advantage of using AI technologies for cephalometric Despite the surging of AI-related research in the field
analysis, which was the primary motivation for most users of orthodontics, orthodontic practitioners still lack
to adopt this technology. However, although the overall knowledge of AI and AI-based cephalometric applica-
tendency was satisfactory for the accuracy of AI-assisted tions related to fundamental principles and classifica-
cephalometric applications, there was still potential for tions, which was consistent with the results of other
improvement. Only a tiny fraction (3.9%) of respondents studies,32,43 and high levels of related knowledge were
did not perform manual calibration after employing AI ap- significantly associated with having participated in AI
plications, with the possible reason that the inaccuracy of clinical or scientific research projects. A possible expla-
the AI applications is influenced by the selection of AI nation is that most respondents have only used the ap-
models, the sharpness of medical radiographs and the plications and never acquired a better comprehension of
American Journal of Orthodontics and Dentofacial Orthopedics October 2023 Vol 164 Issue 4
e102 Lin et al
Table III. Results of multivariable GEE regression analyses for self-evaluation and total knowledge score of AI
Self-evaluation of familiarity with AI AI-related knowledge score
AI’s underlying principles and associated information. analysis replacing manual and semiautomatic cephalo-
Despite the proliferating research and application of AI metric measurements soon. This indicates that they
technologies in orthodontics, it is frequently conducted recognized the liberation of productivity from techno-
in partnership between orthodontic practitioners and logical advances and the improved accuracy and stability
programmers. In the future, cultivating multidisciplinary of the progressive development of AI technologies. Or-
orthodontics and computer science abilities may thodontic students were less likely to agree with the
contribute to advancing orthodontic AI. potentially revolutionary effects of AI, possibly because
Previous research, among other medical specialties, of their lack of experience distinguishing the effective-
found that respondents regarded the work of health ness and accuracy of AI-based applications, as evidenced
care workers as irreplaceable by AI technologies.30,32,44 by their less agreement that AI software can reduce the
Medical interactions require trust, communication, and time spent on cephalometric analysis. They might
empathy,44-46 particularly, for a positive patient- require extra time to determine whether the outcomes
physician relationship required in orthodontics, which of automated landmark detection were objectively accu-
needs to be maintained over time. Several aids, such as rate and to make tradeoffs and adjustments if they dis-
cephalometric analysis, may be substituted, which does agreed with the applications.
not require physicians to use them in the presence of pa- Subjects who had participated in AI-related clinical or
tients, rather than clinical practice from start to finish. scientific research and those whose workplace setting
The results of this survey show that a large percentage was privately rated themselves as more familiar with AI
of orthodontic practitioners were optimistic about the technologies. The high self-perception of those groups
possibility of AI-assisted automatic cephalometric might be influenced by curiosity and interest in new
October 2023 Vol 164 Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
Lin et al e103
Table IV. Respondents’ attitudes and behaviors toward AI and AI cephalometric applications by specialty
Questions about attitude and behaviors Total Orthodontists (n 5 298) Orthodontic students (n 5 182) P value
24. I am interested in the future application of AI in 5 (4-5) 5 (4-5) 4 (4-5) 0.002
orthodontics
29. AI cephalometric applications reduce the time 4 (4-5) 5 (4-5) 4 (4-5) 0.001
required to perform cephalometric analysis
30. AI cephalometric applications can help me to get more 4 (3-5) 4 (3-5) 4 (3-4) 0.041
accurate analysis results
37. AI cephalometric applications will replace manual and 4 (3-5) 4 (4-5) 4 (3-4) \0.001
semiautomatic cephalometric analysis soon
Note. Values are presented as median (25th-75th percentile). Likert scales ranged from 1 (completely disagree) to 5 (completely agree).
Fig 2. The most important strengths that respondents consider an AI-assisted cephalometric applica-
tion should possess (semiopen question, no. 36) (n 5 480).
technologies, a more flexible private enterprise model, and personalized orthodontic treatment plans, esthetic di-
comparatively less stringent access rules. Respondents’ versity, and various operating technique levels. The
main sources of AI information were lectures and training design of applications that match these needs and are
programs, followed by academic conferences, whereas very widely applicable would face the challenge of robust-
few came from undergraduate and graduate courses. This ness, comparability, and generalizability of the results,
differed from the results in other surveys, which showed which might necessitate even more data support, stan-
the main source was social media.32,44 Most respondents dardization of medical data administration, and perhaps
were enthusiastic about the future application of AI in or- international cooperation.27 The survey reveals that re-
thodontics, as indicated by the findings. Therefore, further spondents were willing to pay for using AI-assisted
academic presentations and lectures focused on AI tech- cephalometric applications, a positive sign from a health
nologies can be considered to assist practitioners in com- economics perspective for future AI research and devel-
prehending the most recent industry developments. opment, leading to clinical translation and commercial-
Meanwhile, related content could be incorporated into un- ization. The development prospect of AI technologies in
dergraduate and graduate education to stimulate interest, orthodontics is still broad, and this survey undoubtedly
cultivate interdisciplinary talents, and advance the indus- provides a valuable reference for research direction.
try’s growth, as previous studies suggested.2,30,47 There were also limitations in this survey. First, as with
The route AI technologies in orthodontics would take previous surveys, although the number of valid question-
in the future was a topic of increasing interest to many naires returned in this survey was sufficient for data anal-
respondents. This survey was conducted from the view- ysis, selection bias should still be considered, and the
point of orthodontic practitioners or users of AI-based results be interpreted with caution. Second, the sample
applications, and their expectations mainly focused on size of certain population groups might not fully repre-
assisting treatment decisions and design and orthodon- sent the aggregate in which they are located. However,
tic result prediction and presentation. However, details the sample of this study covered almost all provincial
of orthodontic clinical practice are influenced by administrative regions in China, which was
American Journal of Orthodontics and Dentofacial Orthopedics October 2023 Vol 164 Issue 4
e104 Lin et al
October 2023 Vol 164 Issue 4 American Journal of Orthodontics and Dentofacial Orthopedics
Lin et al e105
the diagnosis of orthognathic surgery. J Craniofac Surg 2019;30: and opinions gleaned from a nationwide online survey. Radiol Med
1986-9. 2021;126:63-71.
19. Tanikawa C, Yamashiro T. Development of novel artificial intelli- 34. Eysenbach G. Improving the quality of Web surveys: the Checklist
gence systems to predict facial morphology after orthognathic sur- for Reporting Results of Internet E-Surveys (CHERRIES). J Med
gery and orthodontic treatment in Japanese patients. Sci Rep Internet Res 2004;6:e34.
2021;11:15853. 35. Tyrer S, Heyman B. Sampling in epidemiological research: issues,
20. Topol EJ. High-performance medicine: the convergence of human hazards and pitfalls. BJPsych Bull 2016;40:57-60.
and artificial intelligence. Nat Med 2019;25:44-56. 36. Hua F, Qin D, Yan J, Zhao T, He H. COVID-19 related experience,
21. Broadbent BH. A new x-ray technique and its application to ortho- knowledge, attitude, and behaviors among 2,669 orthodontists,
dontia: the introduction of cephalometric radiography. Angle Or- orthodontic residents, and nurses in China: a cross-sectional sur-
thod 1981;51:93-114. vey. Front Med (Lausanne) 2020;7:481.
22. Miloro M, Borba AM, Ribeiro-Junior O, Naclerio-Homem MG, 37. Oh K, Oh IS, Le VNT, Lee DW. Deep anatomical context feature
Jungner M. Is there consistency in cephalometric landmark identi- learning for cephalometric landmark detection. IEEE J Biomed
fication amongst oral and maxillofacial surgeons? Int J Oral Max- Health Inform 2021;25:806-17.
illofac Surg 2014;43:445-53. 38. Jianbin MA, Chaoran XUE, Ding BAI. Progress on application of
23. Yu HJ, Cho SR, Kim MJ, Kim WH, Kim JW, Choi J. Automated skel- artificial intelligence technology in orthodontic diagnosis and
etal classification with lateral cephalometry based on artificial in- treatment. J Prev Treat Stomatol Dis 2022;30:278-82.
telligence. J Dent Res 2020;99:249-56. 39. Erkan M, Gurel HG, Nur M, Demirel B. Reliability of four different
24. Nishimoto S, Sotsuka Y, Kawai K, Ishise H, Kakibuchi M. Personal computerized cephalometric analysis programs. Eur J Orthod
computer-based cephalometric landmark detection with deep 2012;34:318-21.
learning, using cephalograms on the Internet. J Craniofac Surg 40. Sayinsu K, Isik F, Trakyali G, Arun T. An evaluation of the errors in
2019;30:91-5. cephalometric measurements on scanned cephalometric images
25. Lee JH, Yu HJ, Kim MJ, Kim JW, Choi J. Automated cephalometric and conventional tracings. Eur J Orthod 2007;29:105-8.
landmark detection with confidence regions using Bayesian con- 41. Hwang HW, Moon JH, Kim MG, Donatelli RE, Lee SJ. Evaluation of
volutional neural networks. BMC Oral Health 2020;20:270. automated cephalometric analysis based on the latest deep
26. Hwang HW, Park JH, Moon JH, Yu Y, Kim H, Her SB, et al. Auto- learning method. Angle Orthod 2021;91:329-35.
mated identification of cephalometric landmarks: part 2-Might it 42. Song Y, Qiao X, Iwamoto Y, Chen Y-w. Automatic cephalometric
be better than human? Angle Orthod 2020;90:69-76. landmark detection on x-ray images using a deep-learning
27. Shan T, Tay FR, Gu L. Application of artificial intelligence in method. Appl Sci 2020;10:2547.
dentistry. J Dent Res 2021;100:232-44. 43. Waymel Q, Badr S, Demondion X, Cotten A, Jacques T. Impact of
28. Hosny A, Parmar C, Quackenbush J, Schwartz LH, Aerts HJWL. the rise of artificial intelligence in radiology: what do radiologists
Artificial intelligence in radiology. Nat Rev Cancer 2018;18: think? Diagn Interv Imaging 2019;100:327-36.
500-10. 44. Bisdas S, Topriceanu CC, Zakrzewska Z, Irimia AV, Shakallis L,
29. Schwendicke F, Golla T, Dreher M, Krois J. Convolutional neural Subhash J, et al. Artificial intelligence in medicine: a multinational
networks for dental image diagnostics: a scoping review. J Dent multi-center survey on the medical and dental students’ percep-
2019;91:103226. tion. Front Public Health 2021;9:795284.
30. Pinto Dos Santos D, Giese D, Brodehl S, Chon SH, Staab W, 45. Inkster B, Sarda S, Subramanian V. An empathy-driven, conversa-
Kleinert R, et al. Medical students’ attitude towards artificial intel- tional artificial intelligence agent (Wysa) for digital mental well-
ligence: a multicentre survey. Eur Radiol 2019;29:1640-6. being: real-world data evaluation mixed-methods study. JMIR
31. Huisman M, Ranschaert E, Parker W, Mastrodicasa D, Koci M, MHealth UHealth 2018;6:e12106.
Pinto de Santos D, et al. An international survey on AI in radiology 46. Blease C, Kaptchuk TJ, Bernstein MH, Mandl KD, Halamka JD,
in 1,041 radiologists and radiology residents part 1: fear of DesRoches CM. Artificial intelligence and the future of primary
replacement, knowledge, and attitude. Eur Radiol 2021;31: care: exploratory qualitative study of UK general practitioners’
7058-66. views. J Med Internet Res 2019;21:e12802.
32. Y€uzbaşıo
glu E. Attitudes and perceptions of dental students to- 47. Pauwels R, Del Rey YC. Attitude of Brazilian dentists and dental
wards artificial intelligence. J Dent Educ 2021;85:60-8. students regarding the future role of artificial intelligence in oral
33. Coppola F, Faggioni L, Regge D, Giovagnoni A, Golfieri R, radiology: a multicenter survey. Dentomaxillofac Radiol 2021;
Bibbolino C, et al. Artificial intelligence: radiologists’ expectations 50:20200461.
American Journal of Orthodontics and Dentofacial Orthopedics October 2023 Vol 164 Issue 4