A privacy-preserved horizontal federated learning for malignant glioma tumour detection using distributed data-silos

Shagun Sharma; Kalpna Guleria; Ayush Dogra; Deepali Gupta; Sapna Juneja; Swati Kumari; Ali Nauman

doi:10.1371/journal.pone.0316543

Peer Review History

Original SubmissionJuly 16, 2024
13 Aug 2024 Decision Letter - Subramani Neelakandan, Editor PONE-D-24-28560A Privacy-preserved Horizontal Federated Learning for Malignant Glioma Tumour Detection using Distributed Data-silosPLOS ONE Dear Dr. kumari, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please submit your revised manuscript by Sep 27 2024 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript: A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Subramani Neelakandan Academic Editor PLOS ONE Journal Requirements: 1. When submitting your revision, we need you to address these additional requirements. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf 2. Please note that PLOS ONE has specific guidelines on code sharing for submissions in which author-generated code underpins the findings in the manuscript. In these cases, all author-generated code must be made available without restrictions upon publication of the work. Please review our guidelines at https://journals.plos.org/plosone/s/materials-and-software-sharing#loc-sharing-code and ensure that your code is shared in a way that follows best practice and facilitates reproducibility and reuse. 3. We note that your Data Availability Statement is currently as follows: All relevant data are within the manuscript and its Supporting Information files. Please confirm at this time whether or not your submission contains all raw data required to replicate the results of your study. Authors must share the “minimal data set” for their submission. PLOS defines the minimal data set to consist of the data required to replicate all study findings reported in the article, as well as related metadata and methods (https://journals.plos.org/plosone/s/data-availability#loc-minimal-data-set-definition). For example, authors should submit the following data: - The values behind the means, standard deviations and other measures reported; - The values used to build graphs; - The points extracted from images for analysis. Authors do not need to submit their entire data set if only a portion of the data was used in the reported study. If your submission does not contain these data, please either upload them as Supporting Information files or deposit them to a stable, public repository and provide us with the relevant URLs, DOIs, or accession numbers. For a list of recommended repositories, please see https://journals.plos.org/plosone/s/recommended-repositories. If there are ethical or legal restrictions on sharing a de-identified data set, please explain them in detail (e.g., data contain potentially sensitive information, data are owned by a third-party organization, etc.) and who has imposed them (e.g., an ethics committee). Please also provide contact information for a data access committee, ethics committee, or other institutional body to which data requests may be sent. If data are owned by a third party, please indicate how others may request data access. 4. When completing the data availability statement of the submission form, you indicated that you will make your data available on acceptance. We strongly recommend all authors decide on a data sharing plan before acceptance, as the process can be lengthy and hold up publication timelines. Please note that, though access restrictions are acceptable now, your entire data will need to be made freely accessible if your manuscript is accepted for publication. This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If you are unable to adhere to our open data policy, please kindly revise your statement to explain your reasoning and we will seek the editor's input on an exemption. Please be assured that, once you have provided your new statement, the assessment of your exemption will not hold up the peer review process. Additional Editor Comments: Author presented the research work "Privacy-preserved Horizontal Federated Learning for Malignant Glioma Tumour Detection using Distributed Data-silos". I have found some major flaws in the manuscript especially many typo errors and others. Every acronym should be defined the first time it appears for example FL-based, FL should abbreviated first time it appears similarly other abbreviations too. Author should discuss the recent pubished works in the introduction section. In the table.1, author just used symbol for preprocessing and data balancing instead of describing the process like how they done the pre0processing ? whether author used any segmentation or filters or labeling etc. In section 3.1.1 Author discussed about 224X224X3. is the right way to describe the image size ? Similarly the statement "The collected image and pre-processed image sample have been shown in Fig. 1." Author discussing about single image or general process of pre-processing images ?. Many statements are not conveying the exact meaning. Author need to describe properly about privacy-preserved in Fig.2. How this privacy-preserved was executed in Fig.4 too? Also author stated that "the weights from each client are collected and aggregated using the FedAVG technique and passed to the server for global model update". Here what is the initial weight value assigned for each client and what is the base for this ? In section 4.2, all the mathematical formula was dis appeared and showing poor presentation. Author must verify the pdf file or original file before submitting to the journal. In table.3 the performance paramete accuracy alone represented in % where as other parameters presented in normal value , what is the reason ? Also result presented for only 5 clients and what about remaining clients ? Also the values are getting confused to identify the exact performance of each clients, for example C1 achieved the accuracy of 92.79% to 99.76%. Here what is the actual predicted value ? you mean that the predicted range may start from 92.79% to 99.76%. So as a reviewer how we can understand the value claimed by C1? Whether table values presented round wise or epoch? if the values are round wise then why not to represent round basis ? Author should describe the detailed experimental setup information for this result and processing. All the graph showing that after 2nd round all the clients achieved the accuracy 90% what is the real scenario behind this ? Author applied any special features for this or any model evaluation techniques applied ? Author may include the model evaluation or fitness score for this research. Fig.15,16,17 and 18 can be drawn as single image so that the length of the manuscript may be reduced similarly other images too. Table.4 too discuss the same scenarionof table.2. Overall author failed to discuss most important discussion about the privacy preserved result parameters. Instead of this author described too many result which is not required to prove the novelty of the research. Its better to discuss the ablation study and optimization of the proposed model with the state of the art instead of too many result with the different rounds and different clients. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ******** 2. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ****** 3. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes ****** 4. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes ****** 5. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: 1- You need to explain the structure of the proposed system with simplified steps in the abstract 2- Ensure that the scientific term and its abbreviation are included only when they first appear 3-Point 3 listed below is not considered a contribution, but rather results reached by the researcher: The outcomes of the proposed FL model have been analyzed in terms of accuracy, recall, loss, precision, and F1-score. The predictive study of the two data distribution configurations identified that in both IID and non-IID distributions, the better generalization of the accuracy was analysed with 5-client architecture as 99.76% and 99.71%, whereas, for 10 clients architecture, it has been achieved as 99.64% and 99.61%, respectively. 4- Ensure that the research in the section Literature review on tumor detection federated learning frameworks is arranged according to the year of publication (from oldest to newest), in addition to listing the weak points of this research during its review. 5- Insert a detailed table to divide the Dataset. 6- Verify the symbols for the equations mentioned in Results and Discussion section 7- Better clarify future work ****** 6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No ******** [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0316543.r001
Revision 1
30 Sep 2024 Author Response Editor’s Comments 1. Author presented the research work "Privacy-preserved Horizontal Federated Learning for Malignant Glioma Tumour Detection using Distributed Data-silos". I have found some major flaws in the manuscript especially many typo errors and others. Every acronym should be defined the first time it appears for example FL-based, FL should abbreviated first time it appears similarly other abbreviations too. Response: Done! Dear Editor, Thank you for your suggestion. As per the suggestion, the manuscript has been proofread for typo errors. Further, all acronyms are used at their first use, and this change has been applied consistently throughout the manuscript. 2. Author should discuss the recent pubished works in the introduction section. Response: Done! Dear Editor, Thank you for your suggestion. In accordance with the suggestion, recent references relevant to the subject area have been cited in the introduction section as references [7], [9], [22], [25], and [26] on pages 3 and 4, as highlighted in green. 3. In the table.1, author just used symbol for preprocessing and data balancing instead of describing the process like how they done the pre0processing ? whether author used any segmentation or filters or labeling etc. Response: Done! Dear Editor, Thank you for the suggestion! In accordance with the suggestion, the pre-processing and data-balancing techniques used in the existing works have been included in Table 1 on pages 8-15, as highlighted in green. 4. In section 3.1.1 Author discussed about 224X224X3. is the right way to describe the image size ? Response: Done! Dear Editor, Thank you for the suggestion! The grayscale images were used to train the proposed model. These images are resized to 224X224X1; however, the size was mistakenly written as 224X224X3 in the manuscript. This error has now been corrected and updated to 224X224X1 on pages 4, 16, and 23. 5. Similarly the statement "The collected image and pre-processed image sample have been shown in Fig. 1." Author discussing about single image or general process of pre-processing images ?. Response: Dear Editor, The statement “The collected image and pre-processed image sample have been shown in Fig. 1” is intended to illustrate examples of the original and pre-processed images from the entire dataset. This represents the general pre-processing technique that has been applied to the complete dataset. 6. Many statements are not conveying the exact meaning. Author need to describe properly about privacy-preserved in Fig.2. How this privacy-preserved was executed in Fig.4 too? Response: Done! Dear Editor, Thank you for the suggestion! In accordance with the suggestion, the detailed description of privacy preservation in Fig. 2 and Fig. 4 has been added on pages 18 and 23, as highlighted in green. Healthcare providers follow data privacy preservation acts, including HIPAA in the United States, the General Data Protection Regulation (GDPR) in the European Union, the Personal Information Protection and Electronic Documents Act (PIPEDA) in Canada, and the Digital Personal Data Protection Act. (DPDT 2023) in India. The proposed federated framework also follows privacy preservation making it technically suitable for disease prediction, without sharing the data to the server or any other peer clients. The proposed model preserves privacy in such a way that this framework contains numerous hospitals and a central server where each hospital contains its own private dataset at a personalized location. The data contained within each client is neither shared with the central server nor with the other participating clients. The federated process starts from the central server containing the base model, which is distributed to each participating clients to train using a personalized dataset. The clients use their private dataset for training the base model and share the model updates with the central server, which then performs the updates aggregation and updates the global model with the aggregated data. Further, this process continues until the federated framework achieves the most optimal accuracy for glioma detection in a privacy-preserving architecture. This framework shows the model training process in a decentralized manner, where, the client’s data is private and neither shared with the central server nor with other clients. 7. Also author stated that "the weights from each client are collected and aggregated using the FedAVG technique and passed to the server for global model update". Here what is the initial weight value assigned for each client and what is the base for this ? Response: Done! Dear Editor, Thank you for the suggestion! In accordance with the comment, the description regarding the base of the initial weights has been addressed in section 3.2 on pages 17 and 18, as highlighted in green color. 8. In section 4.2, all the mathematical formula was disappeared and showing poor presentation. Author must verify the pdf file or original file before submitting to the journal. Response: Done! Dear Editor, Thank you for your valuable suggestion! In accordance with the suggestion, the equations have been updated throughout the manuscript on pages 21, 22, and 26, as highlighted in green. 9. In table.3 the performance paramete accuracy alone represented in % where as other parameters presented in normal value , what is the reason ? Response: Dear Editor, Thank you for your valuable analysis! In Tables 3 and 4 (now updated to Table. 4 and Table. 5), accuracy is represented as a percentage, while other parameters are shown as decimal values. In the proposed work, the accuracy parameter is expressed as a percentage to indicate the proportion of correct predictions, reflecting the model’s performance. It is a standard convention in the area of deep learning to depict accuracy as a percentage. In contrast, Precision, Recall, and F1-Score are presented as decimal values to align with their inherent scale and calculation methods, ensuring consistency. Therefore, the accuracy of the proposed model has been depicted as a percentage and other parameters as decimal values. 10. Also result presented for only 5 clients and what about remaining clients ? Response: Done! Dear Editor, Thank you for your valuable analysis! The proposed FL-based MobileNetV2 model was trained using both 5-client and 10-client architectures. Results for the 5-client setup are presented in Tables 4 and 6, while results for the 10-client framework are shown in Tables 5 and 7, located on pages 29, 36, 37, 44, 45, 51, 52, and 53, as highlighted in green. Additionally, performance graphs for the 5-client and 10-client architectures are provided in sections 4.1 and 4.2 for IID data and in sections 4.3 and 4.4 for non-IID data. 11. Also the values are getting confused to identify the exact performance of each clients, for example C1 achieved the accuracy of 92.79% to 99.76%. Here what is the actual predicted value ? you mean that the predicted range may start from 92.79% to 99.76%. So as a reviewer how we can understand the value claimed by C1? Whether table values presented round wise or epoch? if the values are round wise then why not to represent round basis? Response: Done! Dear Editor, Thank you for your valuable suggestion! As per the suggestion, the accuracy as well as other performance parameters, have been updated in the round-wise in Tables 4 on pages 28 and 29, Table 5 on pages 35, 36, and 37, Table 6 on pages 44 and 45, and Table 7 on pages 51, 52, and 53, as highlighted in green. 12. Author should describe the detailed experimental setup information for this result and processing. Response: Done! Dear Editor, Thank you for your valuable suggestion! In accordance with the suggestion, the experimental setup information for results and processing has been included in Section 3.2 on page 19, highlighted in green. Additionally, a detailed hyperparameter configuration is provided in Table 3 on page 25, also highlighted in green. 13. All the graph showing that after 2nd round all the clients achieved the accuracy 90% what is the real scenario behind this ? Author applied any special features for this or any model evaluation techniques applied ? Response: Dear Editor, Thank you for your valuable comment! Yes, in the proposed federated learning framework, we have implemented an optimized version of the FedAvg algorithm with the Adam adaptive learning rate to accelerate convergence during the model's development. Further, the base MobileNetV2 model is also initialized with pre-trained weights from the ImageNet dataset, which contributes to achieving over 90% accuracy after the second round. 14. Author may include the model evaluation or fitness score for this research. Fig.15,16,17 and 18 can be drawn as single image so that the length of the manuscript may be reduced similarly other images too. Response: Done! Dear Editor, Thank you for your valuable suggestion! The fitness score for the proposed model has been calculated based on accuracy, precision, recall, loss, and F1-score, as detailed in Section 4 ("Results and Discussion"). This section includes fitness scores for each client and for the entire FL framework, covering both IID and non-IID data distributions across 5-client and 10-client architectures. In response to the suggestion, the graphs have been consolidated into single images. Specifically, accuracy has been represented separately in Fig. 5, Fig. 15, Fig. 25, and Fig. 35 (now updated to Fig. 5, ● Fig. 13, Fig. 21, and Fig. 29), which is shown as a percentage. Additionally: ● Fig. 6, 7, and 8 have been combined into a single image and replaced with Fig. 6 on page 27. ● Fig. 16, 17, and 18 have been merged into a single image and replaced with Fig. 14 on page 34. ● Fig. 26, 27, and 28 have been combined into one image and replaced with Fig. 22 on page 43. ● Fig. 36, 37, and 38 have been merged into one image and replaced with Fig. 30 on page 50. 15. Table.4 too discuss the same scenarionof table.2. Response: Done! Dear Editor, Thank you for sparing your valuable time in reviewing the manuscript. We would like to clarify that Table 2 (now updated to Table. 3) and Table 4(now updated to Table. 5) serve different purposes: Table 2(now updated to Table. 3) details the hyperparameter configurations for developing the proposed federated learning-based glioma brain tumor detection framework, while Table 4(now updated to Table. 5) presents a comprehensive analysis of the round-wise performance outcomes for the 10-client federated framework with IID data distribution. 16. Overall author failed to discuss most important discussion about the privacy preserved result parameters. Instead of this author described too many result which is not required to prove the novelty of the research. Response: Dear Editor, Thank you for the suggestion! In the context of federated architecture, Data privacy refers to the protection of personal or sensitive information from unauthorized access or misuse, ensuring that individuals' data is handled in compliance with privacy laws and regulations, whereas, In the proposed work, federated learning has been used for data privacy, where the data at each hospital is neither shared with the server nor with peer clients. Only the updates have been shared with the central server instead of the actual data. This shows that the proposed model follows strict data privacy preservation acts, including HIPAA in the United States, the General Data Protection Regulation (GDPR) in the European Union, the Personal Information Protection and Electronic Documents Act (PIPEDA) in Canada, and the Digital Personal Data Protection Act. (DPDT 2023) in India. As per the suggestion, the discussion about the proposed privacy-preserved federated learning architecture has been presented in section 3 on pages 18 and 23, as highlighted in green. 17. Its better to discuss the ablation study and optimization of the proposed model with the state of the art instead of too many result with the different rounds and different clients. Response: Done! Dear Editor, Thank you for your valuable suggestion! In response to your valuable suggestion, the ablation study and optimization of the proposed model compared to state-of-the-art models have been included in Section 4.6 on pages 60 and 61, as highlighted in green. Additionally, to reduce the number of images in the manuscript, several figures have been combined. The updated figure details are as follows: ● Fig. 6, 7, and 8 have been combined into a single image and replaced with Fig. 6 on page 27. ● Fig. 16, 17, and 18 have been merged into a single image and replaced with Fig. 14 on page 34. ● Fig. 26, 27, and 28 have been combined into one image and replaced with Fig. 22 on page 43. ● Fig. 36, 37, and 38 have been merged into one image and replaced with Fig. 30 on page 50. #Reviewer’s Comments 1. You need to explain the structure of the proposed system with simplified steps in the abstract Response: Done! Dear reviewer, Thank you for your suggestion. Following the suggestion, the comment has been addressed in the abstract on pages 1 and 2, as highlighted in green color. 2. Ensure that the scientific term and its abbreviation are included only when they first appear Response: Done! Dear reviewer, Thank you! In accordance with the suggestion, scientific terms have been abbreviated before their first use throughout the manuscript. 3. Point 3 listed below is not considered a contribution, but rather results reached by the researcher: The outcomes of the proposed FL model have been analyzed in terms of accuracy, recall, loss, precision, and F1-score. The predictive study of the two data distribution configurations identified that in both IID and non-IID distributions, the better generalization of the accuracy was analysed with 5-client architecture as 99.76% and 99.71%, whereas, for 10 clients architecture, it has been achieved as 99.64% and 99.61%, respectively. Response: Done! Dear Reviewer, Thank you for your valuable suggestion! In response to the suggestion, the contribution of the proposed work has been updated in section 1 on page 4, as highlighted in green color. 4. Ensure that the research in the section Literature review on tumor detection federated learning frameworks is arranged according to the year of publication (from oldest to newest), in addition to listing the weak points of this research during its review. Response: Done! Dear Reviewer, Thank you for the valuable suggestion. As suggested, the changes have been incorporated into Section 2 on pages 6-15. Additionally, the weaknesses and research gaps in existing models have been detailed in Section 2, following Table 1 on page 15, as highlighted in green. These research gaps are also summarized in the last column, "Future Scope," of Table 1. 5. Insert a detailed table to divide the Dataset. Response: Done! Dear Reviewer, Thank you for the valuable suggestion. In accordance with the suggestion, a detailed description of the dataset distribution has been added to Table. 2 in section 3.1 on page 16 as highlighted in green color. 6. Verify the symbols for the equations mentioned in Results and Discussion section Response: Done! Dear reviewer, Thank you for the suggestion! In accordance with the suggestion, the symbols and equations presented in the results and discussion section have been verified and updated in section 4 on page 26, as highlighted in green color. 7. Better clarify future work Response: Done! Dear reviewer, Thank you for the suggestion! In accordance with the suggestion, future work has been added to the conclusion section on page 62, as highlighted in green. Submitted for your kind information and action. We will be highly obliged if we are provided with the opportunity to publish our research work through your esteemed journal. Attachments Attachment Submitted filename: Reviewer Comments [PONE-D-24-28560] (1).docx https://doi.org/10.1371/journal.pone.0316543.r002
29 Oct 2024 Decision Letter - Subramani Neelakandan, Editor PONE-D-24-28560R1A Privacy-preserved Horizontal Federated Learning for Malignant Glioma Tumour Detection using Distributed Data-silosPLOS ONE Dear Dr. kumari, Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process. Please submit your revised manuscript by Dec 13 2024 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file. Please include the following items when submitting your revised manuscript: A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'. A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'. An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'. If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter. If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols. We look forward to receiving your revised manuscript. Kind regards, Subramani Neelakandan Academic Editor PLOS ONE Journal Requirements: Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice. Additional Editor Comments: Many of my comments not addressed properly with the required information so please address the below comments carefully with the valid information without fail. In the table.1, author just used symbol for preprocessing and data balancing instead of describing the process like how they done the preprocessing ? whether author used any segmentation or filters or labeling etc. In section 3.1.1 Author discussed about 224X224X3. is the right way to describe the image size ? Similarly the statement "The collected image and pre-processed image sample have been shown in Fig. 1." Author discussing about single image or general process of pre-processing images ?. Many statements are not conveying the exact meaning. Author need to describe properly about privacy-preserved in Fig.2. How this privacy-preserved was executed in Fig.4 too? Also author stated that "the weights from each client are collected and aggregated using the FedAVG technique and passed to the server for global model update". Here what is the initial weight value assigned for each client and what is the base for this ? [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author 1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation. Reviewer #1: All comments have been addressed ******** 2. Is the manuscript technically sound, and do the data support the conclusions? The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented. Reviewer #1: Yes ****** 3. Has the statistical analysis been performed appropriately and rigorously? Reviewer #1: Yes ****** 4. Have the authors made all data underlying the findings in their manuscript fully available? The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified. Reviewer #1: Yes ****** 5. Is the manuscript presented in an intelligible fashion and written in standard English? PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here. Reviewer #1: Yes ****** 6. Review Comments to the Author Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters) Reviewer #1: (No Response) ****** 7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files. If you choose “no”, your identity will remain anonymous but your review may still be made public. Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy. Reviewer #1: No ******** [NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.] While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step. https://doi.org/10.1371/journal.pone.0316543.r003
Revision 2
19 Nov 2024 Author Response Comments Response In the table.1, author just used symbol for preprocessing and data balancing instead of describing the process like how they done the pre0processing ? whether author used any segmentation or filters or labeling etc. Done! Dear Editor, Thank you for the valuable suggestion! In accordance with your suggestion, we have thoroughly reviewed the existing literature and incorporated the relevant pre-processing methods and data-balancing techniques utilized in the existing works at the end of section 2.2 before Table 1. The same has been highlighted in green color on pages 8-16. In section 3.1.1 Author discussed about 224X224X3. is the right way to describe the image size ? Done! Dear Editor, Thank you for the suggestion! As per the suggestion provided in the previous revision, we have updated the image dimensions to 224X224X1, which is intended to show the dimensions of an image tensor used as input to the EfficientNetB3 model. However, in section 3.1.1, we want to show the spatial dimensions of the image, hence, it has been updated to 224X224 pixels for height and width, respectively, on pages 4, 18, and 25, as highlighted in green color. Similarly the statement "The collected image and pre-processed image sample have been shown in Fig. 1." Author discussing about single image or general process of pre-processing images ?. Done! Dear Editor, The statement “The collected image and pre-processed image sample have been shown in Fig. 1” is intended to illustrate examples of the original and pre-processed images from the entire dataset. However, for simplification and better understanding, the statement has been updated to “An example of the initially collected images and the resized images has been illustrated in Fig. 1. The dataset has 2 classes, where the example of resizing the images have been shown as; sample 1 and sample 2 represent the resizing of actual no-tumour image to 224X224 pixels and sample 3 and sample 4 shows the resizing of actual glioma image to 224X224 pixels.” This represents the general pre-processing technique that has been applied to the complete dataset. The same has been updated in section 3.1.1 and in Fig 1 as highlighted in green on page 18. Many statements are not conveying the exact meaning. Author need to describe properly about privacy-preserved in Fig.2. How this privacy-preserved was executed in Fig.4 too? Done! Dear Editor, Thank you for the suggestion! In accordance with the suggestion, the detailed description of privacy preservation in Fig. 2 and Fig. 4 has been added on pages 20 and 25, as highlighted in green. Further, an additional Figure, “Fig 5 Flow diagram of the proposed FL-based glioma detection model” has also been added in section 3.2.2 for a better understanding of the proposed privacy-preserved federated learning architecture, as highlighted in green on pages 27 and 28. The proposed model preserves privacy in such a way that this framework contains numerous hospitals and a central server where each hospital contains its own private dataset at a personalized location. The data contained within each client is neither shared with the central server nor with the other participating clients. The federated process starts from the central server containing the base model, which is distributed to each participating client to train using a personalized dataset. The clients use their private dataset to train the base model and share the model updates with the central server instead of sharing the actual data. The central server uses a FedAVG aggregation for update aggregation. This aggregated value is used to update the global model. Further, this process continues until the federated framework achieves the most optimal accuracy for glioma detection in a privacy-preserving architecture. This framework shows the model training process in a decentralized manner, where the client’s data is private and is not shared with the central server or with other clients. Also author stated that "the weights from each client are collected and aggregated using the FedAVG technique and passed to the server for global model update". Here what is the initial weight value assigned for each client and what is the base for this ? Done! Dear Editor, Thank you for the suggestion! We have updated the manuscript to specify that each client in the federated architecture is provided with the initial global model’s weights, which were trained on the ImageNet dataset on the server. Additionally, as per your suggestion, we have revised the description regarding the initial weight values in Section 3.2 on page 19 and highlighted these updates in green color. The same has been mentioned below for your kind information: The proposed work introduces a FL-based MobileNetV2 model for brain tumour glioma classification from non-tumour MRI scans. The base MobileNetV2 model has been initially trained with the ImageNet dataset on the central server and distributed to the participating client in the FL framework. Each layer in the MobileNetV2 has been initialized with the pre-trained weights of the ImageNet dataset, which contains 1 million images and 1000 different classes. The MobileNetV2 model training using the ImageNet dataset is responsible for configuring the initial weights to the server and participating clients. The range of these weights is between min = -1.061939001083374 to max = 1.4804697036743164. These weights obtained from the ImageNet dataset provide better model training and faster convergence. In the proposed FL-based model, the initial weight values assigned to each client remain the same as in the global model. These weight values to each client typically come from the central server, where global model initialization has been performed using a pre-trained MobileNetV2 model trained with the ImageNet dataset. In the proposed FL-model, the initial weights distributed to each client are the same as the global MobileNetV2 model trained using ImageNet dataset. It represents that each client receives and starts training with the same weights of the initial global model developed on the server side. These weight values provide the model with information regarding the classification task and lead to optimal outcomes as the model is already trained using a large dataset. https://doi.org/10.1371/journal.pone.0316543.r004
13 Dec 2024 Decision Letter - Subramani Neelakandan, Editor A Privacy-preserved Horizontal Federated Learning for Malignant Glioma Tumour Detection using Distributed Data-silos PONE-D-24-28560R2 Dear Dr. kumari, We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements. Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication. An invoice will be generated when your article is formally accepted. Please note, if your institution has a publishing partnership with PLOS and your article meets the relevant criteria, all or part of your publication costs will be covered. Please make sure your user information is up-to-date by logging into Editorial Manager at Editorial Manager® and clicking the ‘Update My Information' link at the top of the page. If you have any questions relating to publication charges, please contact our Author Billing department directly at authorbilling@plos.org. If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. Kind regards, Subramani Neelakandan Academic Editor PLOS ONE Additional Editor Comments (optional): Reviewers' comments: https://doi.org/10.1371/journal.pone.0316543.r005
Formally Accepted
10 Jan 2025 Acceptance Letter - Subramani Neelakandan, Editor PONE-D-24-28560R2 PLOS ONE Dear Dr. kumari, I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now being handed over to our production team. At this stage, our production department will prepare your paper for publication. This includes ensuring the following: * All references, tables, and figures are properly cited * All relevant supporting information is included in the manuscript submission, * There are no issues that prevent the paper from being properly typeset If revisions are needed, the production department will contact you directly to resolve them. If no revisions are needed, you will receive an email when the publication date has been set. At this time, we do not offer pre-publication proofs to authors during production of the accepted work. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few weeks to review your paper and let you know the next and final steps. Lastly, if your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org. If we can help with anything else, please email us at customercare@plos.org. Thank you for submitting your work to PLOS ONE and supporting open access. Kind regards, PLOS ONE Editorial Office Staff on behalf of Dr. Subramani Neelakandan Academic Editor PLOS ONE https://doi.org/10.1371/journal.pone.0316543.r006

Open letter on the publication of peer review reports

PLOS recognizes the benefits of transparency in the peer review process. Therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. Reviewers remain anonymous, unless they choose to reveal their names.

We encourage other journals to join us in this initiative. We hope that our action inspires the community, including researchers, research funders, and research institutions, to recognize the benefits of published peer review reports for all parts of the research system.

Learn more at ASAPbio .