Integrating Genomics, Radiomics, and Pathomics in Oncology: A Scoping Review and a Framework for AI-Enabled Surgomics

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Reviewer's comments: The paper has some imperfections regarding the manuscript structure as well as displaying items. There are the following remarks,
1. There are six figures and five tables; the total references are 74. There is no PRISMA flow diagram of the systematic search.
2. All the figures and Tables should follow the numerical numbers in the main text. Please follow the author's guidelines for academic writing.
3. Figures: Insufficient image resolution. Copyright issues for some images require more careful handling; if applicable, there are no issues or problems.
4. All tables should be concise and clear for readers and reviewers. MDPI publishes many high-quality papers in the field of bioengineering, which can be shown to the authors.
5. The application of multimodal artificial intelligence in genomics, radiomics, and pathomics represents a promising and innovative integration from preventative screening to precision diagnosis. Hopefully, this manuscript will be even better after revisions for your novelty.

Author Response

 

 

Comments	Response
There are six figures and five tables; the total references are 74. There is no PRISMA flow diagram of the systematic search.	A PRISMA flow diagram of the study selection process has been added and is clearly cited in the Results section.
All the figures and Tables should follow the numerical numbers in the main text. Please follow the author's guidelines for academic writing.	All figures and tables have been renumbered and reordered according to the sequence of first mention in the text, following the journal’s guidelines.
Figures: Insufficient image resolution. Copyright issues for some images require more careful handling; if applicable, there are no issues or problems.	All low-resolution figures have been replaced with high-quality versions. Copyright permissions have been verified, and figure formatting has been standardized.
All tables should be concise and clear for readers and reviewers.	All tables have been revised to improve clarity and conciseness, removing redundancy and enhancing readability.
The application of multimodal artificial intelligence in genomics, radiomics, and pathomics represents a promising and innovative integration from preventative screening to precision diagnosis. Hopefully, this manuscript will be even better after revisions for your novelty	The Introduction and Discussion have been revised to highlight the novelty, key contributions, and significance of the proposed framework in current multimodal AI literature.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The review article entitled “Integrating Genomics, Radiomics, and Pathomics in Oncology: A Systematic Review and a Framework for AI-enabled Surgomics” provides a comprehensive analysis of 11 out of 128 screened publications and discusses the advantages and limitations of artificial intelligence tools—such as machine learning and deep learning algorithms—in unifying diverse data modalities for improved diagnosis, prognosis, and treatment outcome prediction. The authors also propose the concept of surgomics as a framework for precision surgery to enhance clinical outcomes. This is an emerging and compelling field, and the manuscript puts forward an interesting and ambitious hypothesis. However, the manuscript is poorly organized, contains overlapping and redundant sections, and does not clearly highlight its key messages.

Major concerns:

1. It is not practical to generate whole-genome sequencing data from patient samples as part of a multimodal AI model in real-world clinical applications.
2. The idea of building multimodal AI models for surgomics is innovative, and I believe the authors intend to develop such a system. However, the question is that whether the authors have the necessary resources, including funding, domain experts, and computing infrastructure.
3. While integrating genomics, radiomics, and pathomics to support surgical decision-making is highly appealing, it would be more realistic to emphasize the development of an integrated therapeutic team that can unify these data modalities as part of patient care.
4. The Discussion section is overly long and loses focus. It is recommended to add a dedicated Surgomics section before the Discussion to improve clarity and structure.

Minor concerns:

1. The order of the tables is confusing and should be reorganized.
2. Line 207 states, “The eligibility criteria include research that has been published in the last decade.” However, in the Methods section (lines 170–173), the authors write: “A comprehensive literature search was conducted across PubMed, Ovid, Wiley Online Library, and Google Scholar from July 1, 2024, to March 5, 2025.” These statements appear inconsistent and should be clarified.
3. Line 385: What specifically are these “other studies”? Please clarify.
4. Lines 400–401: Please include performance scores from the training dataset as well.
5. Line 529: What is the “N-of-1 therapeutic approach”? Please provide a brief explanation.
6. The quality of figures need to be improved with high resolution.

Author Response

 

Response to Reviewer 2 Comments	Response
It is not practical to generate whole-genome sequencing data from patient samples as part of a multimodal AI model in real-world clinical applications.	Statements regarding WGS have been revised to present a realistic view, clarifying that integration is primarily feasible in pre- and post-operative settings.
Whether the authors have the resources (funding, experts, computing infrastructure) to develop a surgomics system.	The manuscript has been updated to clarify that surgomics represents a conceptual roadmap rather than a fully implemented system, with feasibility and resource requirements outlined.
Emphasize an integrated therapeutic team to unify modalities as part of patient care.	The Discussion has been modified to stress the importance of a multidisciplinary team supported by multimodal decision tools.
Discussion is overly long; add a dedicated Surgomics section before Discussion.	A new “Surgomics” section has been added before the Discussion, and the Discussion has been shortened and refocused.
The order of the tables is confusing and should be reorganized.	All tables have been reorganized and renumbered to match their first mention in the text.
Inconsistency between “last decade” vs the stated search dates (July 1, 2024–March 5, 2025).	The eligibility criteria and Methods have been made fully consistent, and the use of older studies has been clarified as background only.
Line 385: “other studies” is unclear.	The term “other studies” has been replaced with specific references for clarity.
Lines 400–401: include performance scores from the training dataset as well.	Training performance metrics have been included where available; where absent, the text explicitly notes this.
Line 529: define “N-of-1 therapeutic approach.”	A brief definition of the “N-of-1 therapeutic approach” has been added at first mention.
Figures need high resolution.	All figures have been upgraded to high-resolution versions with consistent formatting.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript titled "Integrating Genomics, Radiomics, and Pathomics in Oncology: A Systematic Review and a Framework for AI-Enabled Surgomics" provides a comprehensive review of the current evidence regarding the integration of multi-omics technologie (genomics, radiomics, and pathomic) in precision oncology. The authors focus on diagnostic accuracy, predictive modeling, and clinical applicability, and also propose a conceptual framework that combines multi-omics data with intraoperative video analytics (surgomics) to support AI-driven decision-making in surgery.

The authors analyzed the latest studies published from July 1, 2024, to March 5, 2025, using databases such as PubMed, Ovid, Wiley Online Library, and Google Scholar. The review protocol was prospectively registered in PROSPERO (CRD420251009238) on March 11, 2025, before data extraction and synthesis took place. The methodological quality of the studies was assessed using the Newcastle-Ottawa Scale and the Cochrane Risk of Bias Tool. This systematic review adhered to the PRISMA reporting guidelines.

The manuscript thoroughly discusses existing data, highlighting both the strengths and limitations of the approaches employed. It addresses the key challenges and opportunities for AI-driven methods in predicting and prioritizing cancer treatment strategies.

In the Introduction, the authors establish the significance and novelty of the research topic. The methods for a systematic review and analysis of the publications and AI platforms are described in detail. The references included comprise the most relevant publications for the topic. This review is likely to attract the attention of a broad audience including cancer researchers, clinicians, and healthcare specialists.

 

I have the following comments on the manuscript:

1. This review combines two types of scientific literature analysis: a systematic review and an overview of the prospective development of new biomedical technologies, integrating several modern approaches. A significant portion of the manuscript focuses on analyzing existing advances in AI models applied to the integration of genomic and imaging technologies in oncology. It also describes proposed innovations for utilizing AI to tailor surgical approaches and treatments for individual patients by extracting quantitative data from medical images and whole-genome sequencing (WGS).

However, the initial part of the review, which discusses advances in the integration of two- and three-omics data, lacks a thorough analysis of the actual results from the selected publications based on the specified criteria. The Results section offers little beyond a tabular presentation of these publications, with virtually no in-depth analysis or conclusions drawn from them. Additionally, Sections 3.1 to 3.6 are quite brief and uninformative.

2. This review does not fully meet the objectives and requirements of a systematic review. The authors assessed existing studies on the integration of multi-omics technologies, including genomics, radiomics, and pathomics, for precision oncology, adhering to the PRISMA reporting guidelines. However, the manuscript lacks a section dedicated to publications analyzing advances and prospects in surgeomics.
3. The criteria for including or excluding studies in the systematic review, such as "Studies that do not address the promise of genomics, radiomics, and pathomics in medicine/healthcare," are not adequately justified.
4. The Discussion section serves as a separate review focusing on the authors' perspective regarding the challenges and prospects of developing complex AI-driven surgical technologies in oncology, as well as the potential for integrating various healthcare initiatives. While this section is well-written and engaging, it resembles a standalone analysis rather than a discussion of the Results chapter. Therefore, it is essential to align the title, objectives, and all sections of the manuscript.

Author Response

Response to Reviewer 3 Comments Response Results are mostly tabular; Sections 3.1–3.6 are brief and uninformative; needs deeper analysis of included studies. The Results section has been expanded with a richer narrative synthesis, adding discussion of performance trends and validation practices across modalities. Manuscript lacks a section dedicated to advances/prospects in surgomics. A dedicated Surgomics section has been added, clearly separating evidence synthesis from the conceptual framework. Inclusion/exclusion criteria (e.g., “promise”) are not adequately justifi ed. The inclusion and exclusion criteria have been refi ned and justifi ed using clearer operational defi nitions. Discussion reads like a standalone review rather than discussion of Results; align title/objectives/sections. The Discussion has been restructured to directly interpret the Results and ensure full alignment with title and objectives.

Comments	Response
Results are mostly tabular; Sections 3.1–3.6 are brief and uninformative; needs deeper analysis of included studies.	The Results section has been expanded with a richer narrative synthesis, adding discussion of performance trends and validation practices across modalities.
Manuscript lacks a section dedicated to advances/prospects in surgomics.	A dedicated Surgomics section has been added, clearly separating evidence synthesis from the conceptual framework.
Inclusion/exclusion criteria (e.g., “promise”) are not adequately justified.	The inclusion and exclusion criteria have been refined and justified using clearer operational definitions.
Discussion reads like a standalone review rather than discussion of Results; align title/objectives/sections.	The Discussion has been restructured to directly interpret the Results and ensure full alignment with title and objectives.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

This review focused on integrating genomics, radiomics, and pathomics toward AI-enabled precision medicine and the new concept of “surgomics”. The manuscript has provided a useful synthesis of recent work and an ambitious conceptual framework (AiRGOS) for intraoperative multi-omics decision support. However, several substantive issues must be addressed to improve rigor before publication. 

1. The manuscript indicated to restrict inclusion to English-language studies published between July 1, 2024 and March 5, 2025, but it still cited and discussed older seminal works outside that window. I suggest the authors to justify more fully why such a narrow and recent window was chosen for the systematic synthesis and discuss how this affects conclusions. 
2. The review indicated that PROSPERO registration was stated as March 11, 2025, after the end of the search window, March 5, 2025. Was the protocol developed prior to screening and data extraction? If registration occurred after data extraction began, please explain timing and potential risk for bias.
3. I think the authors should provide complete search syntaxes for all databases including PubMed, Ovid, Wiley,and Google Scholar in a supplementary file rather than only listing the main keywords/MeSH terms in the main text.
4. The inclusion criterion “requiring ≥2 integrated omics layers and quantitative clinical outcomes” seemed What was the precise operational definition of “integrated” like statistical/model fusion, cross-scale correlation, or joint-model training? How were quantitative clinical outcomes defined, including OS, PFS, response, and diagnostic AUC?
5. Although the NOS thresholds and recommendations have been described,the methods for scoring were not sufficient enough. Please describe inter-rater reliability on quality scoring. For the one study with RCT elements (Braman et al.), RoB2 was used. Please explain the rationale and ensure consistent application across eligible study designs.
6. The authorsmentioned sensitivity analyses but no results are shown. How high-risk studies affected the conclusions?
7. The manuscript suggested that intraoperative real-time genomics and radiomics will be feasible shortly. It seemed too I recommend the authors for balanced discussion of current technical barriers to avoid overstatement, such as time-to-result for WGS, intraoperative sample processing and sequencing limits, and regulatory concerns of AI-guided intraoperative changes.

Author Response

 

Response to Reviewer 4 Comments	Response
Narrow inclusion window (July 1, 2024–March 5, 2025) but older works are cited; justify and discuss impact on conclusions.	The time window for systematic synthesis has been justified, and its implications are discussed in the Methods and Discussion.
PROSPERO registration date is after the search end date; clarify timing and risk of bias.	The PROSPERO timeline has been clarified, indicating when protocol finalization and screening occurred, addressing potential bias.
Provide complete search syntaxes in a supplementary file.	The manuscript now includes clearly reported database names, search dates, and core terms. Full syntaxes are referenced in supplementary materials.
Define “integrated” and “quantitative clinical outcomes” precisely.	Operational definitions for “integrated” and “quantitative clinical outcomes” have been added to the Methods section.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have adequately addressed my concerns and incorporated the suggested revisions. I have no further comments, and I recommend the manuscript for publication.

Author Response

We sincerely thank the reviewers and the Editorial Board for their constructive comments and valuable feedback. Thank you for your constructive feedback. 

Reviewer 3 Report

Comments and Suggestions for Authors

The revised manuscript was reclassified and renamed as a scoping review. The Results section has been expanded to include a discussion on performance trends. A dedicated section on Surgomics has been added to ensure alignment with the title and objectives. The inclusion and exclusion criteria have been refined and clearly justified. The Discussion has been reorganized to directly interpret the Results, significantly improving readability. Additionally, the prospects of the proposed approaches and technologies are clearly and distinctly outlined.

Author Response

We sincerely thank the reviewers and the Editorial Board for their constructive comments and valuable feedback. Thank you for your constructive feedback.