A Basic Study for Predicting Dysphagia in Panoramic X-ray Images Using Artificial Intelligence (AI) Part 2: Analysis of the Position of the Hyoid Bone on Panoramic Radiographs

Round 1

Reviewer 1 Report

This paper proposes to perform image diagnosis of the hyoid bone on panoramic radiographs using AI. The idea is good but the necessary details are not found. Specifically, following needs clarification.

1. The originality of the work has to be established as using any AI approach for image diagnosis has nothing new in it and its only a few hours work.

2. What different AI based approaches are used? The details and analytical comments on their selection criteria is not found

3. What is the complexity analysis and how is this better than other classical image classification approaches?

4. The references are not appropriate and needs to be significantly improved. Following papers may be included in Introduction section

https://doi.org/10.3390/healthcare11030347

https://doi.org/10.3390/s23156837

The English language may be improved.

Author Response

Dear　Reviewer 1

Thank you for spending time for me. I am really appreciating it.

1. The originality of the work has to be established as using any AI approach for image diagnosis has nothing new in it and its only a few hours work.

Ans.

General dentists typically do not focus on the position of the hyoid bone. Instead of these dentists, an artificial intelligence (AI) system could automatically check the position of the hyoid bone and alert them to the risk of dysphagia if the hyoid bone is in a low position. We believe that it is possible to prevent the decline in oral function from an earlier stage.

Abstract was changed as follows;

Background: Oral frailty is associated with systemic frailty. The vertical hyoid bone position is important for the risk of dysphagia. However, dentist usually do not to focus to this position. Purpose: To create an AI model for detection of the vertical hyoid bone position.

2. What different AI based approaches are used? The details and analytical comments on their selection criteria is not found

              Ans.

     In this study, we do not use different AI based approaches.

The diagnostic accuracy of the first AI program we originally created was low, so we tried to improve it. The methods were to reduce the number of classifications and to exclude cases in which the entire object could not be understood from the learning targets. In the end, we were able to obtain good results.

3. What is the complexity analysis and how is this better than other classical image classification approaches?

              Ans.

     We understood that the large number of groupings led to complexity and low diagnostic accuracy. Also, by reducing the number of groups, the number of cases in each group increased, which ultimately led to better diagnostic accuracy. If a large number of cases can be collected in each group, it may be possible to increase the diagnostic accuracy even if there are many groups, but there is a limit to how many images can be collected at one facility.

4. The references are not appropriate and needs to be significantly improved. Following papers may be included in Introduction section

              Ans.

Thank you for your advice. I will add the articles which you are suggested. Some of our choose article was old.

Thank you for your patient review.

Author Response File: Author Response.docx

Reviewer 2 Report

OK. You used panoramic radiographs of 915 patients aged 20 to 95. Why you not put this age into the Table 1. Numbers of cases for each type?! Young patients, adults patients and onder patients. Please, clarify how many for each one of this groups.

For me the Exposure parameters are very high for this patients age.

Clarify about the: Panoramic radiographs were taken according to the method of Izzetti et al. [19}. what is the method. It wasn't clear!

About the methods. Also, It wasn't clear about: excluded objects that moved during the exposure. Look, it's very common that the patients move or swallow during the panoramic radiographic view, and thus, the position of the hyoid bone must change. How did you deal with that?! The radiographic technique recommends positioning the tip of the tongue behind the upper anterior teeth or on the anterior palate, among other technical-radiographic procedures. How can you be sure that these images were taken in the most correct way?! Because they were archive images, correct?! What would be the conditions for disposal?! Even in relation to the Frankfurt plane (or infraorbital tragus_foramen line)?!

In the Table 1. Numbers of cases for each type. What are the average ages in the different groups? Dysphagia can affect people of any age; however, it is more common in those over 50 years of age. Have you this date?! So, you need to discuss this one.

Author Response

Dear Reviewer 2,

Thank you for spending time for me. I am really appreciating it.

1. You used panoramic radiographs of 915 patients aged 20 to 95. Why you not put this age into the Table 1. Numbers of cases for each type?! Young patients, adults patients and older patients. Please, clarify how many for each one of this groups.

Ans.

I will add the information of age into the Table 1. We cannot deny the possibility that there is some bias between the age of patients visiting these departments and the frequency of classification.

2. For me the Exposure parameters are very high for this patients age.

Ans.

I agree some exposure parameter is too high for the small body mass patients. However, in this study, exposure parameters are little bit adjusted to the patient size but we did not adjust exposure parameter by age.

3. Clarify about the: Panoramic radiographs were taken according to the method of Izzetti et al. [19]. what is the method. It wasn't clear!

Ans.

It was unclear the method. In our clinic, we take standard protocol such as FH plane is parallel to the floor and tongue should have to touch to the palate. As you know, it was so difficult. In our clinic, the patients were also instructed to relax their tongues. So we change article as follows;

Panoramic radiographs were taken as a standardized protocol. When taking the radiographs, the Frankfort horizontal (FH）plane of the patient was set parallel to the floor. During the panoramic radiograph, the patient bite down on a cotton roll to prevent infection. The patients were also instructed to relax their tongues. The excluded objects that hyoid bone moved during the exposure.

4. About the methods. Also, It wasn't clear about: excluded objects that moved during the exposure. Look, it's very common that the patients move or swallow during the panoramic radiographic view, and thus, the position of the hyoid bone must change. How did you deal with that?! The radiographic technique recommends positioning the tip of the tongue behind the upper anterior teeth or on the anterior palate, among other technical-radiographic procedures. How can you be sure that these images were taken in the most correct way?! Because they were archive images, correct?! What would be the conditions for disposal?! Even in relation to the Frankfurt plane (or infraorbital tragus_foramen line)?!

Ans.

The image quality of panoramic radiograph was assessed by Izetti et al [19]. Symmetry, inclination of the occlusal plane, localization of mandibular condyles, aspect of upper teeth root apexes, position of the cervical spine was assessed.

Thank you for your patient review.

Reviewer 3 Report

The purpose of this study was to perform image diagnosis of the hyoid bone on panoramic radiographs using AI.

Method; “The vertical position of the hyoid bone was classified into 6 types based on the classification method described in our previous paper [3].” Add the figure on each type.

The classification was evaluated by external validation? I think that it will occur measurement bias using non-popular classification..

Method; “judged based on the AUC value, as follows: “ , add the citation.

Method; Why did you perform cross validation?

Discussion: too long and not academic scientific writing. Re-write it based on the results.

Discussion:  write your limitation.

Discussion:  please write your strength points and clinical and research implication clearly.

 Minor editing of English language required

Author Response

Dear Reviewer 3,

Thank you for spending time for me. I am really appreciating it.

1. Method; “The vertical position of the hyoid bone was classified into 6 types based on the classification method described in our previous paper [3].” Add the figure on each type.

Ans.

I was added the figure which was cited from our past article.

2. The classification was evaluated by external validation? I think that it will occur measurement bias using non-popular classification.

Ans.

Before this study, we measured the position of the hyoid bone relative to the lower border of the mandible in analysis using lateral cephalometric radiographs [1]. And we have also used this measurement method in our previous research [3]. However, there were no papers that analyzed the position of the hyoid bone in panoramic X-ray photographs. We consider this point to be a limitation of our research.

3. Method; “judged based on the AUC value, as follows: “ , add the citation.

Ans.

I was replaced following sentences as follows;

An AUC value of 0.9 or higher means high accuracy

AUC values above 0.7 and below 0.9 mean moderate accuracy

AUC value greater than or equal to 0.5 and less than 0.7 means low accuracy

4. Method; Why did you perform cross validation?

Ans.

We add the section which named “Cross validation”, add following sentences;

In this study, we performed cross validation. Table 2 shows the number of training sets, validation sets, and test sets for each plan. The learning parameter, the number of epochs, was set to 100, and the batch size was set to 2. Confidence interval was examined under various conditions, and the value with the highest average F value was adopted.

And add Table 2 the number of training sets, validation sets, and test sets for each plan.

5. Discussion: too long and not academic scientific writing. Re-write it based on the results.

Ans.

I was re-writing and delete several sentences.

6. Discussion: write your limitation.

Ans.

I was added the limitation of this study.

7. Discussion: please write your strength points and clinical and research implication clearly.

              Ans.

              I was changed following sentences;

In our previous study, we concluded that the 43 patients diagnosed with dysphagia, 28 patients had a Type 0 or Type 1 hyoid bone position. We found that dysphagia was observed when the hyoid bone was positioned below the mandibular border line [2]. However, the hyoid bone is an anatomical structure that has received little attention in the literature concerning panoramic radiographs. In light of this result, in the present study, we aimed to detect the position of the hyoid bone on panoramic radiographs using AI.

In particular, we found that suspected cases of dysphagia in which the hyoid bone was not visible or only partially visible could be accurately assessed.

Thank you for your patient review.

Round 2

Reviewer 1 Report

I am satisfied with the revised version.

Reviewer 3 Report

No comments

Minor editing of English language required