AI and Face-Driven Orthodontics: A Scoping Review of Digital Advances in Diagnosis and Treatment Planning

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. Can the authors explain why Scopus database is used? Why didn't the authors consider looking at other databases like Web of Science at the same time?

2. The quality of the figures are not up to publication standard. Suggest changing the style and formatting of the graphs to make them look more appealing. Do not use the default style provided by MS words/excel.

3. For each section in the discussion, is it possible to provide some figures to illustrate the scope of discussion?

4. Since this work discussed about artificial intelligence and 3D printing, suggest citing and discussing the following works in the introduction to enrich the literature review.

a. Goh, Guo Liang, et al. "Multi-objective optimization of intense pulsed light sintering process for aerosol jet printed thin film." Mater. Sci. Addit. Manuf 1 (2022).

b. Zhu, Zhijie, et al. "3D-printed multifunctional materials enabled by artificial-intelligence-assisted fabrication technologies." Nature Reviews Materials 6.1 (2021): 27-47.

c. Talaat, Fatma M., and Esraa Hassan. "Artificial Intelligence in 3D Printing." Enabling Machine Learning Applications in Data Science: Proceedings of Arab Conference for Emerging Technologies 2020. Springer Singapore, 2021.

5. Also, since this work also discussed artificial intelligence in biomedical field. suggest citing and discussing the following works in the introduction to enrich the literature review.

a. Elbadawi, M., McCoubrey, L. E., Gavins, F. K., Ong, J. J., Goyanes, A., Gaisford, S., & Basit, A. W. (2021). Harnessing artificial intelligence for the next generation of 3D printed medicines. Advanced Drug Delivery Reviews, 175, 113805.

b. Goh, Guo Dong, et al. "Machine learning for bioelectronics on wearable and implantable devices: challenges and potential." Tissue Engineering Part A 29.1-2 (2023): 20-46.

c. Athanasopoulou, Konstantina, et al. "Artificial intelligence: the milestone in modern biomedical research." BioMedInformatics 2.4 (2022): 727-744.

 

Comments on the Quality of English Language

nil

Author Response

Reviewer 1

1. Can the authors explain why Scopus database is used? Why didn't the authors consider looking at other databases like Web of Science at the same time?
2. The quality of the figures are not up to publication standard. Suggest changing the style and formatting of the graphs to make them look more appealing. Do not use the default style provided by MS words/excel.
3. For each section in the discussion, is it possible to provide some figures to illustrate the scope of discussion?
4. Since this work discussed about artificial intelligence and 3D printing, suggest citing and discussing the following works in the introduction to enrich the literature review.
5. Goh, Guo Liang, et al. "Multi-objective optimization of intense pulsed light sintering process for aerosol jet printed thin film." Mater. Sci. Addit. Manuf1 (2022).
6. Zhu, Zhijie, et al. "3D-printed multifunctional materials enabled by artificial-intelligence-assisted fabrication technologies." Nature Reviews Materials6.1 (2021): 27-47.
7. Talaat, Fatma M., and Esraa Hassan. "Artificial Intelligence in 3D Printing." Enabling Machine Learning Applications in Data Science: Proceedings of Arab Conference for Emerging Technologies 2020. Springer Singapore, 2021.
8. Also, since this work also discussed artificial intelligence in biomedical field. suggest citing and discussing the following works in the introduction to enrich the literature review.
9. Elbadawi, M., McCoubrey, L. E., Gavins, F. K., Ong, J. J., Goyanes, A., Gaisford, S., & Basit, A. W. (2021). Harnessing artificial intelligence for the next generation of 3D printed medicines. Advanced Drug Delivery Reviews, 175, 113805.
10. Goh, Guo Dong, et al. "Machine learning for bioelectronics on wearable and implantable devices: challenges and potential." Tissue Engineering Part A29.1-2 (2023): 20-46.
11. Athanasopoulou, Konstantina, et al. "Artificial intelligence: the milestone in modern biomedical research." BioMedInformatics2.4 (2022): 727-744.

Dear reviewer, thank you for your constructive comments. Here are my responses:

1. Scopus database was used because it conveniently provides the FWCI score, which was used to rank searched articles. This score provides a good comparison of citations within a specific research area. FWCI is part of Scopus metrics.

 

2. Thank you. All figures were reformatted as recommended.

 

 

3. Thank you for your suggestion. We have considered this, although we find it unnecessary and difficult to cover the subsections of AI, 3D printing, facial scanning, and limitations for appropriate scientific illustration. Such a solution would likely distort the focus of the manuscript and may not be scientifically appropriate as it is difficult to provide such illustrations that would define the scope of corresponding subsections.

 

4. Thank you. One of the sources you suggested was included and cited.

 

 

5. Thank you. We tried to strictly keep the focus on orthodontics and dental medicine (and marginally medicine). Albeit very interesting to read, we decided not to use these resources in the paper except of comparison to bioelectronics of author Guo Dong Goh et al. Perhaps we will cite other sources in our future works on Artificial Intelligence. We have added text and reference on the lines 332-346 “This paper highlighted the potential of AI to revolutionize domains to which it is applied. The analysis is demonstrating the versatility and adaptability of this technology. For example, in the case of bioelectronics, AI is helping to overcome the challenges associated with material development, fabrication processes, and system integration. Similarly, in orthodontics, AI is enabling facial analysis to go beyond mere symmetry and proportionality, providing a more comprehensive understanding of facial structure and its impact on dental alignment. AI empowers to tailor treatment strategies to individual patient needs. AI can personalize device design and selection based on patient-specific characteristics in bioelectronics. In orthodontics, AI-driven facial analysis can identify unique facial features and optimize treatment plans accordingly. Data-Driven Decision Making is fundamental for guiding AI-based decision-making processes. In bioelectronics, AI algorithms analyse vast amounts of data to identify patterns and optimize device performance. Similarly, facial analysis tools in orthodontics rely on patient data, such as 2D or 3D scans, to generate insights for treatment planning [103].”

Reviewer 2 Report

Comments and Suggestions for Authors

Dear Authors,

thank you very much for this interesting and relevant review focusing on recent developments regarding AI. The limitations (data base) are discussed and named in the paper. I think this in a really interesting paper.

Some suggestions: Please give a clear aim of your review at the end of the introduction section and in the abstract. Maybe the used data base could be included in the title to clarify the limits of this review.

Some inclusion and exclusion criteria were mentioned. I guess, regarding the discussion and limitations section there were further inclusion and exclusion criteria. They should be mentioned in the Materials and Methods section as well.

Table 1 is sorted by the FWCI. This is fine. I think to improve reader´s interest and benefit, the main outcome of the different included studies could be mentioned in a separate column. Furthermore, the type of study corresponding to figure 3. This might improve understanding and the association between the table and the figures becomes visible. Just a recommendation- may it is possible to realize.

Please adjust the numbering of the figures. There is a mistake.

Please discuss also the limitations, current problems and risks of AI in orthodontics beside all the benefits and chances. The challenges should also be included in this review. Could you compare the possibilities of AI with current treatment concepts - some points should be included in the discussion and also in the conclusion. 

I think at this time, the conclusion should be weakened. We are standing at the beginning of the inclusion of AI in daily practice- this should be considered.

Comments on the Quality of English Language

Minor corrections are needed.

Author Response

Reviewer 2

Dear Authors,

thank you very much for this interesting and relevant review focusing on recent developments regarding AI. The limitations (data base) are discussed and named in the paper. I think this in a really interesting paper.

Some suggestions: Please give a clear aim of your review at the end of the introduction section and in the abstract. Maybe the used data base could be included in the title to clarify the limits of this review.

Some inclusion and exclusion criteria were mentioned. I guess, regarding the discussion and limitations section there were further inclusion and exclusion criteria. They should be mentioned in the Materials and Methods section as well.

Table 1 is sorted by the FWCI. This is fine. I think to improve reader´s interest and benefit, the main outcome of the different included studies could be mentioned in a separate column. Furthermore, the type of study corresponding to figure 3. This might improve understanding and the association between the table and the figures becomes visible. Just a recommendation- may it is possible to realize.

Please adjust the numbering of the figures. There is a mistake.

Please discuss also the limitations, current problems and risks of AI in orthodontics beside all the benefits and chances. The challenges should also be included in this review. Could you compare the possibilities of AI with current treatment concepts - some points should be included in the discussion and also in the conclusion. 

I think at this time, the conclusion should be weakened. We are standing at the beginning of the inclusion of AI in daily practice- this should be considered.

Dear reviewer,

Thank you for your constructive comments and suggestions. We added a clearly formulated aim of this paper at the very end of the introduction.

In the Materials and Methods section, the search algorithm is explained. The inclusion criteria are stated there as well. We are not aware of any new inclusion criteria stated in the Limitations section.

Another column was added to Table 1 to display the main focus of the searched articles.

The numbering of the figures was corrected.

As you have suggested, we have discussed the limitations current problems and risks of AI in orthodontics beside all the benefits and chances. Lines 377-454

We have also compared the possibilities of AI with current treatment concepts and for this we have added a new table 2 – from line 453

As suggested, we have rewritten the conclusion.

Reviewer 3 Report

Comments and Suggestions for Authors

Here are my responses:

 I carefully read through the research paper presented on the use of digital technologies and artificial intelligence in orthodontics. I would offer the following respectful suggestions to strengthen this paper:

 

First, expand the discussion on validating the accuracy of AI algorithms before implementing them clinically. More research quantifying precision and reproducibility would reinforce confidence. 

 

Additionally, include perspectives from orthodontic practitioners on integrating these emerging technologies into daily workflows. Their insights would reveal practical challenges to adoption.  

Finally, given rapid advances in this field, providing periodic updates showcasing the latest capabilities and limitations would keep the analyses current. Research lags technology, so frequent reviews are prudent.

Author Response

Reviewer 3

 

Here are my responses:

I carefully read through the research paper presented on the use of digital technologies and artificial intelligence in orthodontics. I would offer the following respectful suggestions to strengthen this paper:

First, expand the discussion on validating the accuracy of AI algorithms before implementing them clinically. More research quantifying precision and reproducibility would reinforce confidence. 

Additionally, include perspectives from orthodontic practitioners on integrating these emerging technologies into daily workflows. Their insights would reveal practical challenges to adoption.  

Finally, given rapid advances in this field, providing periodic updates showcasing the latest capabilities and limitations would keep the analyses current. Research lags technology, so frequent reviews are prudent.

 

Dear Reviewer,

Thank you very much for your constructive suggestions.

As you suggested, we have expanded the discussion and elaborated on the limitations of the current problems and risks of AI in orthodontics along with all the benefits and opportunities. We agree that validating the accuracy of AI algorithms is a critical step prior to their clinical implementation. Lines 377-454

We have also addressed your other comments and compared the possibilities of AI with current treatment concepts and added a new Table 2 - from line 453 onwards.

We have also rewritten the Conclusions chapter.

Reviewer 4 Report

Comments and Suggestions for Authors

The paper is a survey on a very important topic which is face-driven orthodontics with ai. In general, such a review is needed, but i cannot agree that this paper is valuable for readers. The authors analyze the state of the art by (in general) a few years ago. The survey should focus mainly on the last 2-3 years to show the current state of art and the actual knowledge. This paper missed it. Therefore, i recommend rejection of it due to outdated analysis. Moreover, the authors should discuss different approaches of ai that are used in this area. For instance, the last two years brought many attention-based models or even hybrid solutions. There is no such information or even comparison. This survey should be extended to the analysis of feature extraction (which paper analyzes local or global ones? why?). The same with future directions etc. 

Author Response

Reviewer 4

The paper is a survey on a very important topic which is face-driven orthodontics with ai. In general, such a review is needed, but i cannot agree that this paper is valuable for readers. The authors analyze the state of the art by (in general) a few years ago. The survey should focus mainly on the last 2-3 years to show the current state of art and the actual knowledge. This paper missed it. Therefore, i recommend rejection of it due to outdated analysis. Moreover, the authors should discuss different approaches of ai that are used in this area. For instance, the last two years brought many attention-based models or even hybrid solutions. There is no such information or even comparison. This survey should be extended to the analysis of feature extraction (which paper analyzes local or global ones? why?). The same with future directions etc. 

 

Dear reviewer.

Thank you for your insightful feedback. We appreciate your concern about the paper's timeliness and the need to provide a comprehensive overview of AI-powered facial-driven orthodontics. We have carefully considered your suggestions and have made the following modifications to the manuscript:

• Expanded coverage of recent advancements: We have significantly expanded our discussion of recent advancements in AI for facial-driven orthodontics, particularly those published within the last two to three years. We have incorporated the latest relevant literature from Scopus, even though it may not be fully indexed yet. We understand that the indexing process can take time, but we have made every effort to include the most current information available. Lines 377-492
• Attention-based models and hybrid solutions: We have added a new subchapter 4.5 in Discussion specifically addressing attention-based models and hybrid solutions in facial-driven orthodontics. This section includes five new references that delve into these emerging approaches.
• Feature extraction analysis: We have incorporated a detailed analysis of feature extraction techniques used in facial-driven orthodontics, highlighting the distinction between local and global feature extraction methods and their associated benefits and drawbacks. We have also provided explanations for the rationale behind selecting particular feature extraction methods in specific studies.
• Future directions: We have expanded our discussion of future directions for AI in facial-driven orthodontics, considering the potential of AI to personalize treatment plans, optimize monitoring and prediction, and streamline treatment workflows. We have also discussed the need for further research on ethical considerations and data security in the context of AI-powered orthodontics.

 

We believe that these modifications have substantially enhanced the manuscript's timeliness, comprehensiveness, and clinical relevance. We are confident that the revised manuscript will provide readers with a valuable and up-to-date overview of AI in facial-driven orthodontics.

Thank you again for your feedback and for your point of view. We value your input and are committed to providing high-quality and informative research reviews. The purpose of this paper was to explore the scope of digital technologies used in orthodontics, identify the articles with the highest citations, and provide a framework for further studies. The FWCI ratio was used as a more objective citation criterion, which is rarely calculated for articles published in recent months, and when it is, the values tend to be lower (as discussed in the limitations section). All of the technologies discussed in the paper have already found their clinical applications in daily practice, which we believe is an added value for readers who are not up to date with digital technologies that can be used on a daily basis in clinical settings. If you expect us to elaborate more on attention based models, perhaps another review could explore these latest digital approaches in more detail, albeit we believe a systematic review would be more appropriate in this case.

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

The paper was improved but not sure, if the revisions are sufficient for proper analysis of state of art and insight into discussed topic. 

In my opinion, the paper is a very shallow survey without any proper evaluation of existing solutions. The main idea of the paper was to analyze digital advances used in selected areas of application. However, the paper (even after revisions) does not propose a detailed discussion on the used tools, the differences, or even an analysis of different datasets and reached metrics by those solutions.   After reading the paper, i am not sure what exactly is now used in terms of machine learning and why. Moreover, future directions should be discussed based on the latest solutions and existing modifications - this issue is still very poorly explained and shown.    I hope that these comments will help you understand my position as a reviewer on this paper.

 

Author Response

Review Report (Round 1) 

The paper is a survey on a very important topic which is face-driven orthodontics with ai. In general, such a review is needed, but i cannot agree that this paper is valuable for readers. The authors analyze the state of the art by (in general) a few years ago. The survey should focus mainly on the last 2-3 years to show the current state of art and the actual knowledge. This paper missed it. Therefore, i recommend rejection of it due to outdated analysis. Moreover, the authors should discuss different approaches of ai that are used in this area. For instance, the last two years brought many attention-based models or even hybrid solutions. There is no such information or even comparison. This survey should be extended to the analysis of feature extraction (which paper analyzes local or global ones? why?). The same with future directions etc. 

Review Report (Round 2)

In my opinion, the paper is a very shallow survey without any proper evaluation of existing solutions. The main idea of the paper was to analyze digital advances used in selected areas of application. However, the paper (even after revisions) does not propose a detailed discussion on the used tools, the differences, or even an analysis of different datasets and reached metrics by those solutions.   After reading the paper, i am not sure what exactly is now used in terms of machine learning and why. Moreover, future directions should be discussed based on the latest solutions and existing modifications - this issue is still very poorly explained and shown.    I hope that these comments will help you understand my position as a reviewer on this paper.

 

Dear Reviewer,

We have carefully read your comments from both rounds of this review process. We did our best trying to understand your position and your reservations about our manuscript which was difficult as there were fundamental differences between your first and second round of remarks, albeit we did our best to address them. In the first round you had recommend rejection due to outdated analysis, which we have managed to refute as a misapprehended assertion. In the second round you have rejected dominantly due to the shallow discussion which you had considered not detailed enough. So now, we had addressed this.

The purpose of this manuscript, intended for clinicians linked with orthodontic field, is to determine what clinical applications of AI have been most studied (as evaluated using FWCI indicator), and after that, to outline these technologies to make the reader familiar with intersection of artificial intelligence, CBCT and face scanning with clinical practice. This paper by no means aims to provide detailed information on underlying technical processes running in the background of data processing. Your request to immerse into details of machine learning would render the paper unreadable to the majority of the intended readership. As mentioned, the focus of the paper is orthodontics and how it benefits from digital advances as outlined in cited literature, which we also hope does not give the impression of a shallow survey.

From your last review we have extracted these two particular remarks:

• does not propose a detailed discussion on the used tools, the differences, or even an analysis of different datasets and reached metrics by those…
• future directions should be discussed based on the latest solutions and existing modifications..

We sincerely appreciate your thorough review of our manuscript, and we value every feedback you have provided. As we have carefully considered your comments, we have made revisions to address your concerns. We acknowledge that the original manuscript may have lacked specificity in its evaluation of existing solutions in the selected areas of application. To address this, we have incorporated more detailed discussions about the various digital tools implemented including AI. We have also expanded our analysis to include a more comprehensive assessment of the latest solutions and existing modifications, albeit we find our Discussion chapter too long. Added in Lines 274-323 and 178-187. According to your recommendations, we have also provided insights into the current trends and future directions of the field (Lines 563-599). Paper`s focus is rather clinical so the elaboration on “the used tools, the differences, or even an analysis of different datasets and reached metrics by those” stays intentionally limited. We have also edited the conclusion chapter accordingly.

We understand that the revised manuscript may still require further refinement, and we are committed to addressing any remaining concerns. Please do not hesitate to provide additional feedback as we work towards improving the manuscript.

Thank you again for your time and expertise. We value your insights and appreciate your willingness to assist us in enhancing the quality of our work.

Sincerely,

The team of authors

Author Response File: Author Response.pdf

Round 3

Reviewer 4 Report

Comments and Suggestions for Authors

The paper was improved according to all my comments.