Banana 0.9: An open-source, reproducible medical imaging system for low-resource gastric cancer screening

Xiaoqi Hu

doi:10.1371/journal.pone.0339892

Peer Review History

Original SubmissionSeptember 15, 2025
13 Oct 2025 Decision Letter - Siamak Pedrammehr, Editor Dear Dr. hu, 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 Nov 27 2025 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. 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, Siamak Pedrammehr, Ph.D. Academic Editor PLOS ONE Journal Requirements: When submitting your revision, we need you to address these additional requirements. 1. 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, we expect all author-generated code to 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. Please include your tables as part of your main manuscript and remove the individual files. Please note that supplementary tables (should remain/ be uploaded) as separate "supporting information" files. 4. In the online submission form, you indicated that all relevant data are within the manuscript and its Supporting Information files. Additional raw data and analysis scripts are available from the corresponding author upon reasonable request. All PLOS journals now require all data underlying the findings described in their manuscript to be freely available to other researchers, either 1. In a public repository, 2. Within the manuscript itself, or 3. Uploaded as supplementary information. This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If your data cannot be made publicly available for ethical or legal reasons (e.g., public availability would compromise patient privacy), please explain your reasons on resubmission and your exemption request will be escalated for approval. 5. Please amend your list of authors on the manuscript to ensure that each author is linked to an affiliation. Authors’ affiliations should reflect the institution where the work was done (if authors moved subsequently, you can also list the new affiliation stating “current affiliation:….” as necessary). 6. Please remove your figures from within your manuscript file, leaving only the individual TIFF/EPS image files, uploaded separately. These will be automatically included in the reviewers’ PDF. 7. Please include a copy of Table 1 and 2, which you refer to in your text on page 8. 8. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise. [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? Reviewer #1: Yes Reviewer #2: Partly ******** 2. Has the statistical analysis been performed appropriately and rigorously? -->?> Reviewer #1: No Reviewer #2: No ****** 3. Have the authors made all data underlying the findings in their manuscript fully available??> The PLOS Data policy Reviewer #1: Yes Reviewer #2: No ****** 4. Is the manuscript presented in an intelligible fashion and written in standard English??> Reviewer #1: Yes Reviewer #2: Yes ****** Reviewer #1: While the manuscript is technically sound and of high potential impact, several areas could be strengthened: Verification with empirical data Most results are drawn from simulations as well as from the TCGA-STAD dataset. The authors should explain their method of validating their use of Banana 0.9 with larger or future clinical data sets. Such an explanation would increase the assurance in the clinical applicability of the system. Dataset constraints TCGA-STAD required preprocessing and may not fully reflect real-world clinical data. This limitation is acknowledged but could be discussed more explicitly, particularly regarding generalizability. Comparison with Other Tools The comparisons with nnU-Net and DeepMedic are also given briefly. Quoting more quantitative data (e.g., Dice scores, sensitivity/specificity, runtime) would better place the performance trade-offs in context. Biomarker module Since the biomarker module of the TriOx utilizes artificial biomarker data and not real data, the authors should include robust plans of incorporating verified data from biomarkers in future versions. Statistical openness Simulation framework of Monte Carlo is thorough, but additional data on parameter distributions, assumptions of prevalence, and calculation of confidence intervals would sharpen methodological description. Clinical deployment considerations Enlargement of the discussion concerning challenges from implementation-based on actual settings (e.g., staff education, uneven quality of CT images in rural hospitals, and infrastructural requirements for data storage) would enrich the application-related value of the work. Minor presentation problems. Several sentences in the Methods and Discussion section are phrased overly long and may best be broken into smaller, more manageable statements. Certain inconsistencies in formatting and spacing, particularly with respect to quotation marks, require correction. Reviewer #2: Dear Authors, thank you for submitting “Banana 0.9: An Open-Source, Reproducible Medical Imaging System for Low-Resource Gastric Cancer Screening.” The topic is timely and the open-source/reproducibility focus is commendable; however, to meet PLOS ONE standards the manuscript needs external/clinical validation beyond simulations and TCGA-STAD, comparative benchmarking against state-of-the-art models using standard metrics with uncertainty, fuller statistical reporting (variance/CI, sample-size justification, error analysis, seeds/runs/stratification), and replication-ready materials (pseudocode, parameters, Docker/Conda). Please ensure complete data/code availability with repository links, versions, licenses, and clarify dataset identifiers, ethics/IRB where applicable, and CC BY-NC 4.0 implications; expand the literature review (reproducible pipelines, federated learning, low-power/edge AI) to position the contribution; and improve presentation (complete captions, pipeline flow diagram, consistent terminology/grammar) alongside a concrete roadmap toward “Banana 1.0.” Integrated recommendation: promising but requires major revision focused on clinical validation, rigorous benchmarking with uncertainty, and reproducible documentation and data/code sharing. ****** 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 Reviewer #2: 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. Attachments Attachment Submitted filename: Review report (v1.1).docx https://doi.org/10.1371/journal.pone.0339892.r001
Revision 1
17 Nov 2025 Author Response Response to Reviewers Manuscript ID: PONE-D-25-49901 Title: Banana 0.9: An Open-Source, Reproducible Medical Imaging System for Low-Resource Gastric Cancer Screening Dear Academic Editor and Reviewers, We sincerely thank the Academic Editor and both reviewers for their detailed, constructive, and insightful comments. Their feedback significantly improved the scientific rigor, structure, and reproducibility of the revised manuscript. We have carefully addressed every point raised, and all changes are incorporated into the revised manuscript. Below we provide a point-by-point response, indicating precisely where modifications were made. Line numbers refer to the clean revised manuscript unless specified otherwise. Response to Academic Editor Requirements 1. PLOS ONE style & formatting All formatting issues have been corrected: Updated file naming Added track-changes version Revised affiliations Relocated tables into the main text Removed figures from the manuscript body (Figure 1 uploaded separately) 2. Code-sharing compliance All author-generated code is: Publicly available on GitHub Released under CC BY-NC 4.0 Includes scripts, configs, and logs Contains a reproducible CPU-only pipeline Added explicit statements in Data & Code Availability (Lines 658–662). 3. Tables merged into main manuscript Tables 1 & 2 now appear in the Results section (Lines 320–337). All standalone table files have been removed. 4. Data availability clarification We explicitly confirm: Full data for TCGA-STAD is publicly available External dataset (AbdomenCT-1K) is publicly available All logs, preprocessing outputs, and hashes are provided No private/personal data were used Updated Data Availability section (Lines 658–662). 5. Affiliation corrections All authors are now assigned correct institutional affiliations. (Current affiliation labels added where needed.) 6. Removal of inline figures All figures have been removed from the manuscript file. Figure files uploaded individually as TIFF. 7. Missing Tables 1 & 2 Tables 1 and 2 are now included and cross-referenced correctly (Lines 320–337). 8. Optional citations from reviewers Relevant literature was added where appropriate (Lines 96–115 and throughout). Response to Reviewer #1 We sincerely appreciate Reviewer #1’s positive evaluation and constructive suggestions. Reviewer #1 Comment 1 – Need for empirical validation“Explain how Banana 0.9 will be validated using larger or future clinical datasets.” Response: We have expanded the Future Work section (Lines 624–642) to include a clear roadmap: Planned multi-center collaborations Integration with real biomarker datasets Prospective validation cohorts External testing on additional cancer types Reviewer #1 Comment 2 – Dataset Constraints / TCGA Limitations“Discuss more explicitly how TCGA preprocessing affects generalizability.” Response: We added a strengthened paragraph in Limitations (Lines 574–604) describing: TCGA preprocessing variability Scanner heterogeneity Limited demographic diversity Potential HU-range inconsistencies Reviewer #1 Comment 3 – Quantitative comparison with nnU-Net / DeepMedic “Comparisons were brief; add more quantitative benchmarks.” Response: Expanded comparisons in Comparison to Existing Tools (Lines 398–429), including: Dice = 0.83 (95% CI 0.80–0.86) for Banana 0.9 nnU-Net ≈ 0.90 DeepMedic ≈ 0.88–0.91 5× faster CPU-only inference Hardware–efficiency trade-offs Reviewer #1 Comment 4 – Biomarker Module Uses Synthetic Data “Include plans to incorporate verified biomarker data.” Response: Added explicit future directions in TriOx Module description and Future Work (Lines 624–642): Real-world biomarker integration Validation against clinical tests Cross-hospital EHR linkage Reviewer #1 Comment 5 – Statistical Openness “More detail on parameter distributions and Monte Carlo assumptions is needed.” Response: We added statistical detail in Methods (Lines 220–268), including: Parameters and priors Random seeds Bootstrap confidence intervals Prevalence assumptions Stratified sampling Full reproducibility path Reviewer #1 Comment 6 – Clinical Deployment Considerations “Enrich discussion of challenges in real-world environments.” Response: Added a new paragraph in Discussion (Lines 528–566) covering: Staff training Scanner variability Artifact handling Storage constraints Connectivity challenges Reviewer #1 Comment 7 – Minor presentation and formatting issues Response: All noted long sentences, spacing inconsistencies, and quotation formatting have been corrected.The manuscript underwent professional-level English polishing. Response to Reviewer #2 Thank you for your thorough and insightful feedback. We have substantially strengthened the manuscript accordingly. Reviewer #2 Major Comment 1 – External/Clinical Validation “Manuscript lacks external validation beyond TCGA-STAD.” Response: A full External Validation Study using AbdomenCT-1K TumorSubset (n=773) has been added (Lines 342–389), including: Cross-domain testing Scanner/vendor diversity CPU-only reproducibility Log-file hashing Zero-failure rate This addresses generalizability and real-world heterogeneity. Reviewer #2 Major Comment 2 – Benchmarks with State-of-the-art Methods “Claims need benchmarking with standard metrics and uncertainty.” Response: We expanded Comparison to Existing Tools (Lines 398–429) to include: Dice, CI Sensitivity/specificity Runtime Compute cost Interpretability trade-offs This now situates Banana 0.9 within contemporary segmentation literature. Reviewer #2 Major Comment 3 – Statistical Rigor “Report variance, CI, sample-size justification, stratification, seeds, runs.” Response: Substantial updates in Methods (Lines 220–286): Monte-Carlo assumptions clarified CIs added for all metrics Sampling justification Error propagation Reproducibility design Exact seeds reported All metrics now include SD or bootstrap CI. Reviewer #2 Major Comment 4 – Algorithmic Transparency (HU + Bayesian Updating) Response: We added explicit algorithmic detail including: Pseudocode-style descriptions (Lines 187–219) Parameter ranges Threshold calibration YAML configs Deterministic pipeline logic This ensures full reproducibility. Reviewer #2 Major Comment 5 – Positioning Within Prior Work Response: We significantly expanded the Introduction (Lines 71–142), covering: Reproducible AI frameworks Federated learning Edge-AI medical imaging Low-power inference literature Reviewer #2 Major Comment 6 – Ethics & Licensing Response: Clarified IRB requirements and dataset licenses in Ethics Statement (Lines 547–553). Clarified CC BY-NC 4.0 implications in Data & Code Availability (Lines 658–662). Reviewer #2 Major Comment 7 – Reproducibility Evidence Response: We added: Cross-platform reproducibility testing Output-hash verification CPU-only determinism Public logs (773-case validation) Documented in Reproducibility (Lines 429–456). Reviewer #2 Major Comment 8 – Expanded Limitations Response: We added a more critical Limitations section (Lines 574–604), covering: Synthetic biomarkers Low-contrast CT Rural hardware constraints Domain shift Dataset demographic limitations Reviewer #2 Minor Comments All addressed; includes: Abstract restructuring Improved keywords Better figure/table captions Clear pipeline diagram uploaded separately Reduced repetition Added missing references Final Statement We greatly appreciate the reviewers’ and editor’s time and detailed feedback. We believe the revised manuscript is substantially strengthened, with: New external validation (773 cases) Statistically rigorous design Expanded benchmarking Enhanced reproducibility documentation Improved clinical applicability discussion Clearer limitations and future roadmap We hope that the revised submission now meets the standards of PLOS ONE and respectfully submit the manuscript for further consideration. Attachments Attachment Submitted filename: Response_to_Reviewers_PONE-D-25-49901.docx https://doi.org/10.1371/journal.pone.0339892.r002
27 Nov 2025 Decision Letter - Siamak Pedrammehr, Editor Dear Dr. hu, 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 Jan 11 2026 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. 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, Siamak Pedrammehr, Ph.D. Academic Editor PLOS ONE Journal Requirements: 1. If the reviewer comments include a recommendation to cite specific previously published works, please review and evaluate these publications to determine whether they are relevant and should be cited. There is no requirement to cite these works unless the editor has indicated otherwise. [Note: HTML markup is below. Please do not edit.] Reviewers' comments: Reviewer's Responses to Questions Comments to the Author Reviewer #3: (No Response) Reviewer #4: All comments have been addressed ******** 2. Is the manuscript technically sound, and do the data support the conclusions??> Reviewer #3: Partly Reviewer #4: Yes ****** 3. Has the statistical analysis been performed appropriately and rigorously? -->?> Reviewer #3: Yes Reviewer #4: Yes ****** 4. Have the authors made all data underlying the findings in their manuscript fully available??> The PLOS Data policy Reviewer #3: Yes Reviewer #4: Yes ****** 5. Is the manuscript presented in an intelligible fashion and written in standard English??> Reviewer #3: Yes Reviewer #4: Yes ****** Reviewer #3: The manuscript introduces Banana 0.9, an open-source, GPU-free CT imaging framework designed for low-resource gastric cancer screening. The work is clearly written, technically transparent, and well aligned with reproducibility principles. Public code, modular design, and full configuration files are strong strengths. However, a few points require clarification before acceptance: 1. External validation: Since the AbdomenCT-1K subset lacks segmentation ground truth, the external experiment demonstrates reproducibility rather than accuracy. The authors should clarify this more explicitly and consider adding an indirect accuracy check or an additional dataset with annotations. 2. Biomarker & clinical-risk modules: All risk parameters are literature-based rather than clinically measured. This should be more clearly framed as conceptual modeling, not validated clinical prediction. 3. Cost-reduction claim: The ~60% cost savings need clearer assumptions (hardware pricing, lifespan, maintenance, electricity). A brief sensitivity analysis would strengthen this claim. 4. HU-based segmentation limitations: The discussion should more directly acknowledge potential issues: low-contrast areas, scanner variability, and motion artifacts. Overall, the manuscript is well presented and contributes meaningfully to open, reproducible medical imaging. With the above clarifications, it would be suitable for publication. Reviewer #4: (No Response) ****** 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 #3: No Reviewer #4: Yes: kimia shirini ******** [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.] To ensure your figures meet our technical requirements, please review our figure guidelines: https://journals.plos.org/plosone/s/figures You may also use PLOS’s free figure tool, NAAS, to help you prepare publication quality figures: https://journals.plos.org/plosone/s/figures#loc-tools-for-figure-preparation. NAAS will assess whether your figures meet our technical requirements by comparing each figure against our figure specifications. Attachments Attachment Submitted filename: The manuscript introduces Banana 0.docx https://doi.org/10.1371/journal.pone.0339892.r003
Revision 2
6 Dec 2025 Author Response Response to Reviewers — Revised Manuscript (R2 Final) Manuscript ID: PONE-D-25-49901R1 Title: Banana 0.9: An Open-Source, Reproducible Medical Imaging System for Low-Resource Gastric Cancer Screening Journal: PLOS ONE Reviewer #3 — Point-by-Point Response Comment 1 — External validation clarity Reviewer comment: “Since the AbdomenCT-1K subset lacks segmentation ground truth, the external experiment demonstrates reproducibility rather than accuracy. The authors should clarify this more explicitly and consider adding an indirect accuracy check or an additional dataset with annotations.” Response: We thank the reviewer for this important clarification. We fully agree that because AbdomenCT-1K does not contain voxel-level segmentation ground truth, the external experiment evaluates procedural reproducibility, not accuracy. The manuscript has been revised accordingly: Methods — External Dataset Validation “Because the AbdomenCT-1K dataset does not include segmentation ground truth, this external validation evaluates procedural reproducibility and cross-platform determinism rather than absolute segmentation accuracy.” Discussion — External Generalization “As the dataset lacks voxel-level annotations, the external experiment is not an accuracy benchmark but a reproducibility stress test across heterogeneous scanners and acquisition protocols.” Conclusions “The external evaluation should be interpreted as a reproducibility assessment rather than a measure of segmentation accuracy, given the absence of ground-truth annotations in AbdomenCT-1K.” These revisions ensure transparent interpretation of the external validation results. Comment 2 — Biomarker & clinical-risk modules Reviewer comment: “All risk parameters are literature-based rather than clinically measured. This should be more clearly framed as conceptual modeling, not validated clinical prediction.” Response: We appreciate this suggestion and agree fully. We have added explicit statements describing the TriOx biomarker simulation and clinical risk-factor module as conceptual modeling tools, not validated predictors. Revisions include: Methods — Biomarker Module “The TriOx biomarker and clinical-risk parameters represent a conceptual modeling framework derived from published literature, rather than a validated clinical prediction system.” Limitations “Because all risk parameters are literature-derived, the biomarker module should be interpreted as conceptual modeling rather than a clinically validated prognostic tool.” Comment 3 — Cost-reduction claim Reviewer comment: “The ~60% cost savings need clearer assumptions (hardware pricing, lifespan, maintenance, electricity). A brief sensitivity analysis would strengthen this claim.” Response: Thank you for highlighting this point. We clarified the assumptions underlying the Monte Carlo cost estimations: Results — Cost Simulation “The estimated ~60% per-patient cost reduction assumes standard mid-range CPU hardware, a 3–5 year device lifespan, typical outpatient electricity costs, and no specialized GPU maintenance.” Low-Resource Deployment “The cost comparison reflects hardware-class assumptions (CPU vs. GPU acquisition cost, expected lifespan, and operating energy consumption) rather than absolute economic projections.” This clarification ensures transparency regarding economic assumptions. Comment 4 — HU-based segmentation limitations Reviewer comment: “The discussion should more directly acknowledge potential issues: low-contrast areas, scanner variability, and motion artifacts.” Response: We appreciate this comment. We expanded the manuscript to explicitly acknowledge these well-known limitations of deterministic HU-thresholding: Comparison to Existing Tools “Deterministic HU-based segmentation is sensitive to low-contrast lesions, motion artifacts, and scanner-specific noise patterns, which may reduce performance in certain clinical scenarios.” Limitations “HU-threshold segmentation may be affected by low-contrast boundaries, acquisition variability across institutions, and patient motion, underscoring the need for future hybrid HU + deep-learning approaches.” Reviewer #4 Reviewer #4 stated: “All comments have been addressed.” No further action was required. General Manuscript Revisions To improve clarity and compliance with PLOS ONE guidelines, we also implemented the following updates: Structural Improvements Strengthened transitions between Results → Discussion → Conclusions. Standardized section titles and terminology (e.g., “procedural reproducibility,” “deterministic HU segmentation”). Language Polishing Improved readability by shortening long sentences. Ensured consistency in tense, voice, and formatting. Technical Consistency Verified internal consistency of numerical values, confidence intervals, dataset names, module descriptions, and table references. Ensured Table 1 and Table 2 appear in the main manuscript as required. Reproducibility Enhancements Added clarifications regarding deterministic execution, hashing, fixed random seeds, runtime logs, and cross-platform reproducibility. Confirmed that all code, configuration files, and documentation are publicly available under a CC BY-NC 4.0 license. Required Editorial Compliance Statements 1. Data Availability Compliance All data used in this study (TCGA-STAD, AbdomenCT-1K, and synthetic simulations) are publicly available, and no restricted-access data were used. Code, configuration files, and logs are openly accessible. 2. Figure Upload Policy All figures have been removed from the manuscript body and are provided as separate high-resolution files as required by PLOS ONE. 3. Author Affiliation The single author is now properly affiliated as Independent Researcher. Recommended Citation Evaluation Both reviewers provided methodological suggestions, but neither reviewer recommended any specific publications for citation. We carefully reviewed all reviewer comments and confirmed that no additional references were explicitly requested. Therefore: No new citations were added in response to reviewer recommendations. Existing citations remain appropriate and sufficient to contextualize the Banana 0.9 framework. Closing Statement We thank the reviewers and the academic editor for their constructive comments. We believe the revisions have improved the clarity, transparency, and reproducibility of the manuscript, while maintaining its core contribution as an interpretable, GPU-free framework for low-resource medical imaging. We respectfully submit this revised manuscript for further consideration by PLOS ONE. Attachments Attachment Submitted filename: Response_to_Reviewers_PONE-D-25-49901_auresp_2.docx https://doi.org/10.1371/journal.pone.0339892.r004
15 Dec 2025 Decision Letter - Siamak Pedrammehr, Editor Banana 0.9: An Open-Source, Reproducible Medical Imaging System for Low-Resource Gastric Cancer Screening PONE-D-25-49901R2 Dear Dr. hu, 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. For questions related to billing, please contact billing support . 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, Siamak Pedrammehr, Ph.D. Academic Editor PLOS One Additional Editor Comments (optional): Reviewers' comments: Reviewer's Responses to Questions Comments to the Author Reviewer #3: All comments have been addressed ******** 2. Is the manuscript technically sound, and do the data support the conclusions??> Reviewer #3: Yes ****** 3. Has the statistical analysis been performed appropriately and rigorously? -->?> Reviewer #3: Yes ****** 4. Have the authors made all data underlying the findings in their manuscript fully available??> The PLOS Data policy Reviewer #3: Yes ****** 5. Is the manuscript presented in an intelligible fashion and written in standard English??> Reviewer #3: Yes ****** Reviewer #3: Dear Authors, Thank you for submitting the revised version of your manuscript and the detailed point-by-point response. I have reviewed the revision and I’m pleased to confirm that you have addressed all of my previous comments. In particular, the revision clearly (1) reframes and explains the external dataset evaluation in light of the absence of ground-truth annotations, (2) explicitly states that the biomarker and clinical risk-factor components are literature-informed and presented as conceptual/illustrative modules rather than clinically validated predictive tools, (3) clarifies the assumptions underlying the reported cost-reduction estimate, and (4) strengthens the discussion of limitations of HU-based segmentation, including issues related to low-contrast anatomy, scanner variability, and motion artifacts. These changes improve clarity, transparency, and appropriate interpretation of the work. Overall, I believe the manuscript is now suitable for publication, and I support its acceptance. Best regards, ****** 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 #3: Yes: Mohsen mokhtari Keshavar ******** https://doi.org/10.1371/journal.pone.0339892.r005
Formally Accepted
Acceptance Letter - Siamak Pedrammehr, Editor PONE-D-25-49901R2 PLOS One Dear Dr. hu, 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 You will receive further instructions from the production team, including instructions on how to review your proof when it is ready. Please keep in mind that we are working through a large volume of accepted articles, so please give us a few days 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. You will receive an invoice from PLOS for your publication fee after your manuscript has reached the completed accept phase. If you receive an email requesting payment before acceptance or for any other service, this may be a phishing scheme. Learn how to identify phishing emails and protect your accounts at https://explore.plos.org/phishing. 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. Siamak Pedrammehr Academic Editor PLOS One https://doi.org/10.1371/journal.pone.0339892.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 .