Optimization of Laser-Based Method to Conduct Skin Ablation in Zebrafish and Development of Deep Learning-Based Method for Skin Wound-Size Measurement

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1) There is not enough discussion of the experimental results. In-depth discussion of the results is required.
4) The authors need to better explain the context of this research, including why the research problem is important.
5) The authors should explain clearly what the differences are between the prior work and the solution presented in this paper.
6) The authors should add a table that compares the key characteristics of prior work to highlight their differences and limitations. The authors may also consider adding a line in the table to describe the proposed solution.
7) It is important to clearly explain what is new and what is not in the proposed solution. If some parts are identical, they should be appropriately cited and differences should be highlighted.
8) The experiments should be updated to include some comparison with newer studies.

9)vCost Analysis of Model Training: Please provide a detailed analysis and statistics on the cost of model training, particularly the additional computational costs incurred when using pseudo-labels and regularization terms. Compare the efficacy and cost of this approach with simply increasing the amount of training data, to assess the cost-benefit ratio of your proposed method. Given the relative ease of acquiring image data, analyze whether training with more data is a more effective method and under what circumstances your approach offers more advantages.
10)Error Analysis and Discussion of Limitations: Present failure cases of the model in specific scenarios or data types, analyze the reasons, and explore improvements.
11)Citations for Comparative Methods: In method section where you compare your method with others, please include specific citations for each of the comparative methods discussed. This will not only provide clarity and ease of reference for readers but also establish a solid academic context for your comparisons.
12)Enhance Comparisons with Recent Studies: From the references cited, it appears that comparisons with the very latest studies in the field are limited. I recommend incorporating comparisons with more recent papers, which could further contextualize your work within the current research landscape. This would provide readers with a clearer understanding of how your method stands in relation to the most cutting-edge developments in the field.

 

Comments on the Quality of English Language

The English used is of a moderate level and may benefit from proofreading by a native speaker.

Author Response

Comments and Suggestions for Authors

There is not enough discussion of the experimental results. In-depth discussion of the results is required.

The authors appreciated the comment from the reviewer. Here, despite the numerous results shown in the present study, the authors believed that the results from the validation experiment are the most important issues to be discussed and indeed, in the previous version of the manuscript, these parts were not well-discussed enough. Thus, the authors have added some discussions, especially in the temperature and antioxidant effects to the evaluated wound healing process, since for the AI section, all of the necessary information, such as the differences between the two architectures and their limitations is already described in the text. In the updated manuscript, besides providing more findings from other studies that support the current results, the mechanism of the observed phenomenon was also added to the text. For example, the capability of astaxanthin to enhance a wound healing process might be due to its ability to remove ROS and promote wound healing process by NIH 3T3 cells more effectively without causing cytotoxicity while vitamin C positively affects the early resolution of inflammation and tissue remodeling. In addition, the authors also added a discussion regarding the potencies of the current method to help researchers in studying wound healing in fish.

The authors need to better explain the context of this research, including why the research problem is important.

Thank you for the constructive comment. The authors were aware that the research problem as the background of the current study might not be described well enough in the manuscript. Actually, there are several research problems that were addressed prior to the development of the method. First, rodents are one of the animal models that is generally used in most of the wound healing studies and unfortunately, despite its advantages, the cost of using this animal model is relatively high, thus, another animal model with a lower cost and also suitable with 3R principle is required. Next, regarding the wound-induced method, although some researchers had developed this method in fish, some limitations, such as time-consuming and relatively high-cost apparatus, were still faced. Therefore, by implementing deep-learning algorithms in the wound area measurement processes, the authors believed that the current method will greatly benefit many researchers, especially in the medical field, in evaluating the wound healing process of aquatic animals, especially zebrafish, for instance, in studying the efficacy of some novel medicines. To highlight these issues in the manuscript and the importance of the present findings, the authors have rewritten the manuscript, especially in the abstract and introduction parts.  

The authors should explain clearly what the differences are between the prior work and the solution presented in this paper.

The authors appreciated the suggestion. Actually, the authors have tried their best to explain the differences between the current method and the prior works by summarizing the findings in some prior studies in Table S1 and also mentioning the important points in the text. The differences include the implementation of AI in the wound area size measurement, which also becomes the solution for the tedious and time-consuming process in the regular method using ImageJ from prior studies. However, since these issues have seemed to be quite vague, the authors have revised some parts of the manuscript to help readers understand the findings of this study.

The authors should add a table that compares the key characteristics of prior work to highlight their differences and limitations. The authors may also consider adding a line in the table to describe the proposed solution.

Thank you for the comment. In Table S1, the authors have listed some important studies in this field to help the readers understand the differences between this study to the prior studies and the advantages of the current method in inducing the wound and measuring the wound area. However, this table still needs improvement, including the limitations of each method as the reviewer suggested. Therefore, the authors have added this matter to the table. However, regarding the proposed solution for each study, the authors felt that it is quite inappropriate since the authors do not test the method from each study and, thus, possess less acknowledgment regarding the details of the methods compared to the original authors themselves and feel that is not the authors’ place to propose a solution for the limitations in each method.

It is important to clearly explain what is new and what is not in the proposed solution. If some parts are identical, they should be appropriately cited and differences should be highlighted.

The authors appreciated the reminders. As mentioned in point (5) above, the authors included some sentences in the abstract, introduction, and discussion sections to highlight the novelty of the proposed method, including the implementation of a deep-learning-based algorithm to measure the size of a wound area in zebrafish caused by the laser-engraving machine. To clearly highlight the differences between the current method and previous methods, Table S1 consists of some crucial information, such as the used animal model and limitations, from several important prior studies in this field and also the current method, which was added to the manuscript so the readers could easily compare the present method with other methods.

The experiments should be updated to include some comparison with newer studies.

Thank you for the suggestion. Without understating the importance of the results from the relatively old studies, the authors are also fully aware of the benefits of including the results from newer studies in comparing the current results. Therefore, to keep relevant with technological advancement, more references from recent studies were added to the manuscript, especially in the discussion section, to compare the obtained results by using the current method to the results shown in the prior studies.

Cost Analysis of Model Training: Please provide a detailed analysis and statistics on the cost of model training, particularly the additional computational costs incurred when using pseudo-labels and regularization terms. Compare the efficacy and cost of this approach with simply increasing the amount of training data, to assess the cost-benefit ratio of your proposed method. Given the relative ease of acquiring image data, analyze whether training with more data is a more effective method and under what circumstances your approach offers more advantages.

The authors appreciated the suggestions. First, regarding the details on the cost of model training, including the additional computational costs, the fixed cost for the computer processing system is approximately 6000 USD while the fixed cost of the wound processing system is approximately USD 7000. However, the approximate price for the computer processing system might be varied that depends on the specification of the hardware itself which would also affect the time required for the wound images processing step, which would affect some of the total time required mentioned in the next part regarding the efficacy of this approach in analyzing image data. Here, various image augmentation was employed to increase the number of training images. Generating a wound and taking the image of the wound takes approximately 60 to 120 seconds. Meanwhile, generating a new wound image through image processing methods from an existing real wound image only requires 0.03 to 0.5 seconds by using the current computer system. Generally, processing one real wound image into 14 training images through image processing is preferable, as generating too many images from a single real image may result in minimal variation between each image, leading to a reduced impact on the training effectiveness. In total, 885, 97, and 89 images were used as training, validation, and testing datasets, respectively, to conduct the U-Net test. To increase image diversity, images were distorted about 14-fold to generate a total of 12,390 images as a training dataset. All of this information was added to the manuscript, especially in the materials and methods section. The authors hoped that the given information was clear and sufficient to provide an illustration regarding the cost and effectiveness of the proposed method.

Error Analysis and Discussion of Limitations: Present failure cases of the model in specific scenarios or data types, analyze the reasons, and explore improvements.

Thank you for the constructive suggestion. The authors fully agreed with the reviewer to present more details regarding the failure cases of the current method in specific scenarios or data types. Therefore, the authors added some scenarios that might not be able to be analyzed by the current study, which become the limitations of the current method. These limitations include a high error rate when dealing with smaller wound area sizes size since they are hard to distinguish. To overcome this issue and obtain optimal results, the contrast between the background and wound area is required to be more clear and thus, easy to be distinguished by the AI. Furthermore, a lower predictive accuracy was also found in the wound site of the golden strain zebrafish compared to other zebrafish strains due to the fewer stripe patterns of this fish, highlighting the pivotal role of skin stripe patterns in the training and learning procedures of both deep learning methodologies. All of the limitations, including the ones mentioned above are extensively discussed in the discussion part, highlighting the drawbacks of the current method, the logical explanation regarding the cause of the drawbacks, and potential solutions, if any, to overcome this issue.

Citations for Comparative Methods: In method section where you compare your method with others, please include specific citations for each of the comparative methods discussed. This will not only provide clarity and ease of reference for readers but also establish a solid academic context for your comparisons.

The authors thanked the reviewer for the reminders. The authors are fully aware of the importance of the specific citations to provide clarity for the readers and establish a solid academic context for the comparison in a research article. As the reviewer suggested, the authors have included the specific citations for each of the comparative methods discussed, whether they are in the main body of the manuscript or in Table S1 to enhance the quality of the current writing.

Enhance Comparisons with Recent Studies: From the references cited, it appears that comparisons with the very latest studies in the field are limited. I recommend incorporating comparisons with more recent papers, which could further contextualize your work within the current research landscape. This would provide readers with a clearer understanding of how your method stands in relation to the most cutting-edge developments in the field.

Thank you for the detailed suggestions and the authors strongly agreed with the reviewer. As mentioned in point (8) above, the authors have tried their best to include some of the latest studies in this field despite the limited number of those studies to provide readers with comprehensive information regarding what has been done in this field and how the current method is still relevant to the most cutting-edge developments in the field. As already mentioned in the text, the current study implements an AI-based algorithm, which is becoming a broad and current interest in many fields of research nowadays, to analyze the wound area induced by the machine. The authors believed that the proposed method is standing in relation to the technological advancements in the research field and is able to stand out even in the following years.

Comments on the Quality of English Language

The English used is of a moderate level and may benefit from proofreading by a native speaker.

We have asked native speaker on editing this revised manuscript and hope can reach high standard.

 

Reviewer 2 Report

Comments and Suggestions for Authors

How AI can validate the experimental data?

Pls add some information about image J? 

Introduction section should be improved

Aims znd goals of paper xhould be clarify.

Pls add numerical results in conclusion to make more  strong

Comments on the Quality of English Language

Quality needs to be improved

Author Response

Comments and Suggestions for Authors

How AI can validate the experimental data?

The authors thanked the reviewer for the question. Here, after the fish skin was ablated by using the laser-engraving machine, the size of the wound area was measured every 5 days to evaluate the wound healing process by the newly developed AI tool that is based on a deep learning algorithm instead of a manual calculation by ImageJ. However, since ImageJ has been used as one of the most commonly used methods to measure the wound area of animals, this method was also applied in some of the images to validate the data generated by AI. In addition, due to the high contrast between the background and wound area in the majority of the images, all of the processed images by AI were also validated by a manual observation since the wound areas were easy to be distinguished by unaided eyes. Therefore, in the present study, the performance and validity of the AI tool were validated by the mentioned methods, instead of the AI that validated the experimental data.

Pls add some information about image J? 

Thank you for the question. ImageJ is an image processing program that is based on Java programming language and has been widely used for processing images in various fields. Here, prior to the manual selection of the wound area, the scale function in ImageJ was set to convert the pixel into centimeters as one of the standard units for length. Later, by using the draw function in ImageJ, the outer part of the wound was manually selected and ImageJ would calculate the drawn area size. The authors agreed that this information is essential to the manuscript, thus, this information was added to the text, specifically in the Material and Methods (2.4 Skin wound healing assay), Results (3.3.1 Image collection and pre-processing), and Discussion parts to provide more information regarding ImageJ to the readers.

Introduction section should be improved

The authors appreciated the suggestion. The authors have tried their best to improve the Introduction part by adding some important findings from previous publications, such as various methods applied in wound healing studies in fish including scratch assay and dermal laser along with the limitations of each study, and more literature that became the background of the present study, thus, the authors believed that this newly added information could help in emphasizing the importance of the current study.

Aims and goals of paper should be clarify.

Thank you for the constructive comment. The authors have tried their best to highlight the aims and goals of the present study, which were divided into several parts. The primary goal of the current experiment was to establish a novel method for skin ablation in zebrafish as an alternative way to study the wound healing process in fish. In achieving this purpose, the authors have developed a new methodology for inducing a wound in the fish skin by using a laser-engraving machine and an AI tool that is based on a deep-learning algorithm to automatically measure the wound area with relatively high validity and efficacy. The authors believed that by developing this method, researchers could have a more affordable option in studying the wound healing process in fish and provide a complete methodology from upstream to downstream of a wound healing experiment, from skin ablation to the wound area measurement. Finally, the authors had rewritten some parts of the manuscript, especially in the Introduction part, to emphasize the aims and goals of this study.

Pls add numerical results in conclusion to make more strong

The authors thanked the reviewer for the suggestion and strongly agreed to add numerical results in the Conclusions part to provide a stronger conclusion. Therefore, some values from the results, including the time required for the healing process of the wound until reaches some specific threshold levels were included in the Conclusions section to highlight the effectiveness and efficacy of the tested compounds or how the temperatures could play a major role in the wound healing process of zebrafish.

Comments on the Quality of English Language

Quality needs to be improved

We have asked native speaker on editing this revised manuscript and hope can reach high standard.

Reviewer 3 Report

Comments and Suggestions for Authors

Dear Authors,

I have read the manuscript with interest and some questions raised. Enlisted please find my comments.

Overall. General English grammar revision (Minor spelling errors).

Key words. “medicine” could be added in my opinion.

Abstract. Please add the names of the statistical tests in this section.

Introduction. Authors stated “However, due to economic reasons like high costs for daily maintenance as well as for the animal welfare issue, the scientific community continuously searches for other alternatives that able to fit the 3R principle for replace, reduce and refine on experimental animal usage”. Please add a reference for this statement.

Materials and Methods. Authors stated “The astaxanthin and vitamin C were purchased from Sigma-Aldrich and then prepared as 1000 ppm stock concentration and kept in the 4°C until the time of exposure. After skin laser ablation, fish was transferred to a 3L tank filled with either astaxanthin or vitamin C at dose from 2 to 20 ppm with n number of 12 for each group.”. Please add if and how sample size calculation has been performed.

Materials and Methods. Please add a reference for each method.

Materials and Methods. For each material used, please add details about commercial name manufacturer, City and State.

Materials and Methods. For each machinery used, please add details about commercial name manufacturer, City and State.

Materials and Methods. Please add details about software used, version, Manufacturer, City and State.

Materials and Methods. ANOVA is used for gaussian distributions. Please explain how normality of data was tested.

Materials and Methods. Statistics. Please add significance level for P values (0.05? 0.01?).

Results. Please add P values all along this section.

Discussion. Authors stated “Based on temperature, Astaxanthin and Vitamin C treatment results gave a consistent and similar result with previous studies using these factors in wound healing. This fact proves that the laser-engraving machine and the protocol used in this study was indeed works properly as an alternative for wound healing investigation. Consistent and repeatable result could be achieve using this new established method”. Provide a general interpretation of the results in the context of other evidence, and implications for future research. It could be added that “Additionally it could be interesting to test Laser in combination with other adjuvant treatments such as Ozonized gel  (Scribante A, Gallo S, Pascadopoli M, Frani M, et al. Ozonized gels vs chlorhexidine in non-surgical periodontal treatment: A randomized clinical trial. Oral Diseases 2023 [in press]. DOI: 10.1111/odi.14829) and water (Talasani RR, Potharaju SP, Vijaya Lakshmi B, Durga Bai Y, Chintala RK, Mahankali V, Koppolu P, AlGhamdi ARS. Efficacy of ozonated water over chlorhexidine mouth rinse in chronic gingivitis patients - A comparative clinical study. Saudi Dent J. 2022 Dec;34(8):738-743. doi: 10.1016/j.sdentj.2022.09.004. Epub 2022 Oct 7), or probiotics (Butera A, Maiorani C, Gallo S, Pascadopoli M, Venugopal A, Marya A, et al. Evaluation of Adjuvant Systems in Non-Surgical Peri-Implant Treatment: A Literature Review. Healthcare (Basel). 2022 May 11;10(5):886. doi: 10.3390/healthcare10050886. PMID: 35628025; PMCID: PMC9140356.) in order to understand their mutual effect on wound healing and their possible interaction with Deep Learning-based Methods for Size Measurement.”. These concerns should be added to Discussion section.

Discussion. Please add a paragraph showing the limitations of the present report.

Conclusions. This section is very long and it could be reduced to a couple of paragraphs

References. Some references are quite old (1994;1997;1990;1995;1998;1996;1999;1997;1988;1986;1989;1981;1986;1989;1982). If possible, please switch with some more modern research. Some recent studies have been suggested in the sections above.

Table 1. Please use a letter based representation for intergroup and intragroup significances.

Figure 8. Please enlarge the figures in order to increase readability.

 

Comments on the Quality of English Language

General English grammar revision needed (Minor spelling errors).

Author Response

Comments and Suggestions for Authors

Dear Authors, I have read the manuscript with interest and some questions raised. Enlisted please find my comments. Overall. General English grammar revision (Minor spelling errors). Key words. “medicine” could be added in my opinion.

The authors thanked the reviewer for the suggestion and strongly agreed with the suggestion. “Medicine’ would be a significant addition to the keywords since here, the tested compounds were proven to induce the wound healing process of zebrafish, indicating a potency of this compound to be used as a medicine in similar cases in the future. Thus, the suggested word was included in the keywords, accordingly.

Abstract. Please add the names of the statistical tests in this section.

Thank you for the suggestion. There are two datasets that are mentioned in the abstract and both of them were analyzed by two different statistical tests since their data structure was different from each other. First, for the dataset from the behavior test to obtain the optimal laser power for the wound induction, the Kruskal-Wallis test continued with Dunn’s multiple comparisons test was applied to evaluate the differences in each behavior endpoint between untreated and laser-treated fish groups. Meanwhile, in calculating the statistical differences between control and treated groups in wound healing process datasets that included 25%, 50%, and 75% of wound closure percentage, two-way ANOVA with the Geisser-Greenhouse correction followed by Sidak’s multiple comparison test was applied. The names of these statistical tests were added to the Abstract as the reviewer suggested.

Introduction. Authors stated “However, due to economic reasons like high costs for daily maintenance as well as for the animal welfare issue, the scientific community continuously searches for other alternatives that able to fit the 3R principle for replace, reduce and refine on experimental animal usage”. Please add a reference for this statement.

The authors appreciated the correction and admitted that there was a mistake regarding the absence of references for the mentioned statement. Therefore, in the updated manuscript, the intended references were added to the text, supporting the necessity of the 3R principle in the usage of experimental animals by using other animal models or developing a novel methodology that is more effective and efficient than previous methods. The newly added references are also mentioned below.

Díaz, L., Zambrano, E., Flores, M.E., Contreras, M., Crispín, J.C., Alemán, G., Bravo, C., Armenta, A., Valdés, V.J., Tovar, A. and Gamba, G., 2021. Ethical considerations in animal research: the principle of 3R's. Revista de investigacion clinica, 73(4), pp.199-209.

Franco, N.H. and Olsson, I.A.S., 2014. Scientists and the 3Rs: attitudes to animal use in biomedical research and the effect of mandatory training in laboratory animal science. Laboratory animals, 48(1), pp.50-60.

Kousholt, B.S., Præstegaard, K.F., Stone, J.C., Thomsen, A.F., Johansen, T.T., Ritskes-Hoitinga, M. and Wegener, G., 2023. Reporting of 3Rs Approaches in Preclinical Animal Experimental Studies—A Nationwide Study. Animals, 13(19), p.3005.

Smith, A.J. and Lilley, E., 2019. The role of the three Rs in improving the planning and reproducibility of animal experiments. Animals, 9(11), p.975.

Materials and Methods. Authors stated “The astaxanthin and vitamin C were purchased from Sigma-Aldrich and then prepared as 1000 ppm stock concentration and kept in the 4°C until the time of exposure. After skin laser ablation, fish was transferred to a 3L tank filled with either astaxanthin or vitamin C at dose from 2 to 20 ppm with n number of 12 for each group.”. Please add if and how sample size calculation has been performed.

Thank you for the suggestion. The authors were not sure regarding the reference of the “sample size” that was mentioned by the reviewer since that could refer to the calculation of the tested compounds of the calculation regarding the applied n number of fish in the present study. However, the authors tried their best to elaborate on these issues in this response. Regarding the preparation of the used chemicals, which were astaxanthin and vitamin C, both compounds in a powder form were dissolved in distilled water until they reached a 1000 ppm concentration to be used as the stock solution. Later this stock solution was diluted again in filtered tank water until the concentrations were 2 and 20 ppm to be used as the incubation water for the tested fish. These concentrations were chosen considering the previously used dose of astaxanthin in mice that was 400-1000 ppm, thus, the applied dose in the fish should be lower than those concentrations in mice. In addition, by using the selected doses, significant beneficial effects already could be seen in this study. Next, regarding the applied n number of fish, which was twelve, it was based on the previous publication that also used the same number of fish in studying wound healing in zebrafish.

Seo, S.B., et al., Silver nanoparticles enhance wound healing in zebrafish (Danio rerio). Fish & shellfish immunology, 2017. 68: p. 536-545.

 

Materials and Methods. Please add a reference for each method.

The authors appreciated the suggestion and admitted that there was a mistake in not including the reference for the cited method. Therefore, the authors added the reference for the 3D locomotion assay that was applied in the present study to evaluate the fish swimming performance after laser ablation. In addition, since some methods from prior studies in evaluating the wound healing process of animals were also described in Table S1, the respective references for each study were also added accordingly.

Materials and Methods. For each material used, please add details about commercial name manufacturer, City and State.

Thank you for the reminders. The details about the used materials, such as astaxanthin and vitamin C, including their commercial name and the city and state of their manufacturer were added to the text to provide clearer information regarding the origin of the materials as the reviewer suggested.

Materials and Methods. For each machinery used, please add details about commercial name manufacturer, City and State.

The authors appreciated the reminder. Similar to the response above, the details about the used machinery, including the laser-engraving machine and a digital dissecting microscope, including the city and state of their manufacturer were also included in the manuscript according to the reviewer’s suggestion.

Materials and Methods. Please add details about software used, version, Manufacturer, City and State.

Thank you again for the reminder. The details of the used software, including the mini engraving software and statistical analysis software, were added to the manuscript, including the city and state of the manufacturer of these software as the reviewer suggested.

Materials and Methods. ANOVA is used for gaussian distributions. Please explain how normality of data was tested.

The authors thanked the reviewer for the comment. Prior to the statistical analysis, the distribution normality of the dataset was determined by using a normality test to compute how likely it is for a random variable underlying the data set to be normally distributed (Gaussian distribution). Here, although Graphpad Prism is able to perform several normality tests, the results from a D'Agostino-Pearson test were chosen as the reference for the dataset distribution since this test is more recommended by Graphpad Prism itself than other normality tests. After the dataset was found to follow Gaussian distributions, two-way ANOVA was chosen as the statistical test. Meanwhile, if the data distribution was not following Gaussian distributions, Geisser-Greenhouse correction was implemented in the statistical test. The information regarding the data normality test was added to the manuscript.

Materials and Methods. Statistics. Please add significance level for P values (0.05? 0.01?).

Thank you for the suggestion. The authors admitted that there was a mistake regarding the significance level for P values in the Materials and Methods section, especially in the Statistics part. Therefore, additional information regarding the asterisk symbol shown in very statistical differences was added in the manuscript as the reviewer suggested. The used asterisk and its p-value are described as * p < 0.05; ** p < 0.01; *** p < 0.001; **** p < 0.0001.

Results. Please add P values all along this section.

The authors appreciated the comment. Similar to the response above, the p values in every figure, especially the ones with the wound healing data, in the Results section were added and the values were also added in the figure caption. In addition, to help the readers understand the statistical differences in the data, the p-values were also added to the text. The used asterisk and its p-value are similar to the ones mentioned above.

Discussion. Authors stated “Based on temperature, Astaxanthin and Vitamin C treatment results gave a consistent and similar result with previous studies using these factors in wound healing. This fact proves that the laser-engraving machine and the protocol used in this study was indeed works properly as an alternative for wound healing investigation. Consistent and repeatable result could be achieve using this new established method”. Provide a general interpretation of the results in the context of other evidence, and implications for future research. It could be added that “Additionally it could be interesting to test Laser in combination with other adjuvant treatments such as Ozonized gel  (Scribante A, Gallo S, Pascadopoli M, Frani M, et al. Ozonized gels vs chlorhexidine in non-surgical periodontal treatment: A randomized clinical trial. Oral Diseases 2023 [in press]. DOI: 10.1111/odi.14829) and water (Talasani RR, Potharaju SP, Vijaya Lakshmi B, Durga Bai Y, Chintala RK, Mahankali V, Koppolu P, AlGhamdi ARS. Efficacy of ozonated water over chlorhexidine mouth rinse in chronic gingivitis patients - A comparative clinical study. Saudi Dent J. 2022 Dec;34(8):738-743. doi: 10.1016/j.sdentj.2022.09.004. Epub 2022 Oct 7), or probiotics (Butera A, Maiorani C, Gallo S, Pascadopoli M, Venugopal A, Marya A, et al. Evaluation of Adjuvant Systems in Non-Surgical Peri-Implant Treatment: A Literature Review. Healthcare (Basel). 2022 May 11;10(5):886. doi: 10.3390/healthcare10050886. PMID: 35628025; PMCID: PMC9140356.) in order to understand their mutual effect on wound healing and their possible interaction with Deep Learning-based Methods for Size Measurement.”. These concerns should be added to Discussion section.

Discussion. Please add a paragraph showing the limitations of the present report.

Thank you for the constructive suggestion and for providing several essential references to enhance the quality of the manuscript. In the updated manuscript, the authors added some references to support the drawn conclusion of the current study based on the obtained results. As an example, the benefit effects of vitamin C in fish wound healing process that was observed in the present study was also demonstrated by a prior study in mice and later, it was suggested that this phenomenon was due to the ability of vitamin C in increasing the angiogenesis. The authors believed that the addition of the new references could provide a general interpretation of the results in the context of other evidence. Meanwhile, regarding the implications for future research, the authors strongly agreed with the reviewer's point of view as using this method, researchers could examine any potential compound and material to evaluate the effect of those substances to the wound recovery effect in fish. Thus, the authors have included the mentioned references to the text, especially in the Discussion section. Finally, as the reviewer also suggested, the limitations of this method, including the difficulty in measuring the small wound area, were also included in the manuscript.

Conclusions. This section is very long and it could be reduced to a couple of paragraphs

The authors understood the reviewer’s point of view and also agreed that the Conclusions section might be too long. Therefore, to avoid lengthy writing, the authors have rewritten the paragraph and reduced this section to a couple of paragraphs. In addition, the numeric values from some of the results were added to this part according to the reviewer’s suggestion. The authors hoped that these changes would result in better-written text for the readers to understand the article.

References. Some references are quite old (1994;1997;1990;1995;1998;1996;1999;1997;1988;1986;1989;1981;1986;1989;1982). If possible, please switch with some more modern research. Some recent studies have been suggested in the sections above.

Thank you for the suggestion. Again, the authors understood the reviewer’s point of view and without understating the importance of the results from the relatively old studies, the authors also believed that the addition of more recent references could enhance the reliability and validity of the text itself. Therefore, new citations were added to the manuscript, including the references from recent studies that were suggested in the sections above, especially in the Introduction and Discussion parts as the reviewer suggested.

 

Table 1. Please use a letter based representation for intergroup and intragroup significances.

Thank you for the suggestion and the constructive comment. Prior to the statistics of the Table 1, authors make a changes by incorporated a letter based representation for significance in specificity and sensitivity since author think that these 2 are the most important in the AI wound measurement. Author used One-Way ANOVA followed by Tukey’s multiple comparisons for letter based significances.

Figure 8. Please enlarge the figures in order to increase readability.

The authors thanked the reviewer for the reminder. After reading the comment, the authors were fully aware that the size of Figure 8 was relatively small, hence, providing a low readability to the readers. Therefore, the authors have revised the mentioned figure by increasing its font size to increase its readability according to the reviewer’s suggestion.

Comments on the Quality of English Language

General English grammar revision needed (Minor spelling errors).

We have asked native speaker on editing this revised manuscript and hope can reach high standard.

 

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

All comments have been answered thank you