Artificial Intelligence in Cardiovascular Pathology: Toward a Diagnostic Revolution
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsReviewer Comments
The review “Artificial Intelligence in Cardiovascular Pathology: Toward a Diagnostic Revolution” gives an overview of how artificial intelligence is being applied in cardiovascular pathology, with a focus on clinical areas such as heart transplant rejection, cardiomyopathies, myocarditis, atherosclerotic disease, and valvular disease. The authors also discuss methodological considerations, major limitations like data availability, generalisability, interpretability, and ethical-legal aspects, and include a dedicated section on regulatory perspectives. Overall, the topic is timely and clinically relevant, and the manuscript is generally well-organised and readable. The main weaknesses relate to: clarifying the review methodology and strengthening the critical evaluation of included studies (dataset size, external validation, and clinical readiness), particularly in the summary tables. Therefore, before being considered for publication, I suggest that the authors address the following comments.
Major comments
- The authors explicitly state that the covered studies are not intended to be exhaustive, but rather a curated set of “representative examples”. However, the manuscript would benefit from a clearer description in the Methods section of: how studies were identified (databases/search engines used, keywords, time window) and high-level inclusion/exclusion criteria
- I would suggest strengthening the critical evaluation. Several sections provide performance metrics (for example accuracy or AUC), but it is often unclear whether these results come from single-centre datasets, external validation, or multicentre cohorts. When summarising key studies, consistently report sample size, number of centres, staining/scanner variability, and whether external validation was performed would be really helpful.
- Table 1 is a useful entry point (heart transplant rejection, myocarditis, cardiomyopathies, etc.) and includes selected metrics (for example “AUC up to 0.962; accuracy >90%”). However, as currently structured, the table is difficult to compare across rows because essential context is missing.I suggest adding one or two extra columns, such as: data type (for example histology WSI, cardiac magnetic resonance imaging, echocardiography), cohort/validation (single-centre vs multicentre; internal vs external test set), optionally task type (classification, segmentation, prediction).
- The manuscript uses multiple technical abbreviations (for example whole-slide images (WSIs), electrocardiograms (ECGs), cardiac magnetic resonance imaging (CMR), software as a medical device (SaMD), point-of-care ultrasound (POCUS), convolutional neural networks (CNNs)). Please ensure that every abbreviation is defined at first use (including in tables). This is a minor editorial point but improves readability.
In summary, the manuscript is timely, generally well written, and already includes key limitations and regulatory considerations. Addressing the points above especially clarifying the review scope/methodology, would substantially improve clarity and usefulness without requiring major restructuring.
Author Response
Dear Reviewer 1,
We would like to thank the reviewer for his careful reading of the manuscript and for his constructive comments, which helped improve its clarity, methodological rigor, and clinical utility. Below, we respond in detail to the main comments.
Comment 1. The authors explicitly state that the covered studies are not intended to be exhaustive, but rather a curated set of "representative examples." However, the manuscript would benefit from a clearer description in the Methods section of how studies were identified (databases/search engines used, keywords, time window) and high-level inclusion/exclusion criteria.
Response 1: We thank the reviewer for this observation, which we believe to be particularly pertinent. Although the manuscript was not intended as a systematic review, we felt it appropriate to more clearly explain the methodological approach adopted for study selection. To this end, we have included the Materials and Methods section (highlighted in light blue).
Comment 2. I would suggest strengthening the critical evaluation. Several sections provide performance metrics (for example accuracy or AUC), but it is often unclear whether these results come from single-center datasets, external validation, or multicenter cohorts. When summarizing key studies, consistently report sample size, number of centers, staining/scanner variability, and whether external validation was performed would be really helpful.
Answer 2: We fully agree with the Reviewer on the need to strengthen the critical evaluation and have therefore integrated the clinical subsections of paragraph 4 (highlighted light blue) according to the indications.
Comment 3. Table 1 is a useful entry point (heart transplant rejection, myocarditis, cardiomyopathies, etc.) and includes selected metrics (for example “AUC up to 0.962; accuracy >90%”). However, as currently structured, the table is difficult to compare across rows because essential context is missing.I suggest adding one or two extra columns, such as: data type (for example histology WSI, cardiac magnetic resonance imaging, echocardiography), cohort/validation (single-centre vs multicentre; internal vs external test set), optionally task type (classification, segmentation, prediction).
Answer 3: We have reformulated table 1 according to the indications.
Comment 4. The manuscript uses multiple technical abbreviations (for example whole-slide images (WSIs), electrocardiograms (ECGs), cardiac magnetic resonance imaging (CMR), software as a medical device (SaMD), point-of-care ultrasound (POCUS), convolutional neural networks (CNNs)). Please ensure that every abbreviation is defined at first use (including in tables). This is a minor editorial point but improves readability.
Answer 4: We have inserted the list of abbreviations before the references (highlighted in blue) and corrected the first citation in the manuscript.
We thank the reviewer again for his suggestions that improved the article.
Best regards,
The Authors
Reviewer 2 Report
Comments and Suggestions for Authors Once the manuscript has been read, I request the following from the authors in order to consider it for acceptance, as it requires major revisions.- Reorganize the manuscript to avoid repeated ideas and improve the overall flow between sections.
- Shorten the longest sections, especially where concepts are reiterated without adding new insight.
- Clearly describe how the reviewed studies were selected, including search strategy and inclusion criteria.
- Strengthen the critical analysis by addressing study quality, biases, sample size limitations, and clinical applicability.
- Add a comparative summary table showing key studies by pathology, including datasets, model type, and main performance metrics.
- Review the reference list to correct numbering inconsistencies, duplicates, or misplaced citations.
- Ensure that the figures included (e.g., workflow diagrams) contribute meaningful information and are fully integrated into the text.
- Streamline the regulatory section, focusing only on what is most relevant for cardiovascular pathology.
- Expand the discussion on dataset biases and the challenges associated with single‑center cohorts.
- Adjust the note regarding the use of digital tools to align better with journal standards and maintain a purely academic tone.
- Review the manuscript for clarity, tighten the writing, and improve transitions to enhance readability.
Author Response
Dear Reviewer 2,
We thank the reviewer for their careful reading of the manuscript and their valuable constructive comments. We have carefully revised the manuscript, considering all the observations and have made the requested changes to improve the clarity, structure, and critical analysis of the work. Below, we respond to the comments point by point.
Comment 1) Reorganize the manuscript to avoid repeated ideas and improve the overall flow between sections.
Response 1: We partly reorganized the manuscript by shortening some obvious repetitions and adding inserts, as suggested by the reviewers, to better harmonise different paragraphs.
Comment 2) Shorten the longest sections, especially where concepts are reiterated without adding new insight.
Response 2: When possible, we tried to shorten the longest sections to avoid repetitions without modifying the structure of the manuscript.
Comment 3) Clearly describe how the reviewed studies were selected, including the search strategy and inclusion criteria.
Response 3: We have expanded the Materials and Methods section to clarify the search strategy and study selection process (highlighted in light blue).
Comment 4) Strengthen the critical analysis by addressing study quality, biases, sample size limitations, and clinical applicability.
Response 4: We fully agree with the reviewer and have strengthened the critical analysis of the literature. In particular, a new subsection entitled "4.7 Critical Appraisal of Available Evidence" has been added (highlighted in purple).
Comment 5) Add a comparative summary table showing key studies by pathology, including datasets, model type, and main performance metrics.
Response 5: We have reworded Table 1 according to the suggestions of both reviewers.
Comment 6) Review the reference list to correct numbering inconsistencies, duplicates, or misplaced citations.
Response 6: We have carefully reviewed the entire reference list and in-text citations to correct any inconsistencies.
Comment 7) Ensure that the figures included (e.g., workflow diagrams) contribute meaningful information and are fully integrated into the text.
Response 7: We have improved the integration of figures into the manuscript text. Specifically, explicit references have been added to Figure 1 and Figure 2 (highlighted in purple).
Comment 8) Streamline the regulatory section, focusing only on what is most relevant to cardiovascular pathology.
Response 8: We thank the reviewer for the suggestion. The regulatory section has been revised and simplified, reducing general descriptions of regulatory frameworks and focusing on the aspects most relevant to the application of artificial intelligence in cardiovascular pathology, particularly validation requirements, monitoring algorithm performance, and implications for clinical implementation.
Comment 9) Expand the discussion on dataset biases and the challenges associated with single-center cohorts.
Response 9: In paragraph 4.7, we expanded the discussion on dataset biases and the challenges associated with single-center cohorts.
Comment 10) Adjust the note regarding the use of digital tools to better align with journal standards and maintain a purely academic tone.
Response 10: We have modified the statement regarding the use of digital tools to maintain a more scholarly tone and conform to editorial guidelines (highlighted in purple).
Comment 11: Review the manuscript for clarity, tighten the writing, and improve transitions to enhance readability.
Response 11: we reviewed the manuscript with particular attention to improving transitions and clarity of the text.
We thank the reviewer again for his suggestions that improved the article.
Kind regards,
The Authors
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsIn light of the implemented improvements, I believe the manuscript has progressed significantly, but it still requires a further minor revision to refine the structure, eliminate redundancies, and ensure a more concise read. Therefore, my recommendation is: Acceptance after Minor Revisions. I believe that, once these final adjustments are completed, the work will adequately meet the journal's standards and offer a solid and well-organized contribution to the field of AI applied to cardiovascular pathology.
The revised version (V2) incorporates relevant improvements and addresses most of the requests. In particular, the improvements in the methodological description, the expansion of the critical analysis, and the focus of the regulatory section are noteworthy. However, the following would still be advisable:
- More significantly reducing the length of some sections.
- Avoiding repetitions that hinder reading speed.
- Strengthening the cohesion between sections to achieve an even clearer flow.
Author Response
Dear Reviewer 2,
We thank you for your constructive comments, which have significantly improved the quality of the manuscript. In response to your comments, we have made the following revisions:
- we eliminated redundancies and simplified some wording in Section 3 (changes highlighted in red);
- we substantially shortened Section 7.1, making it more concise and focused;
- we also shortened and simplified Section 6.1, eliminating repetition and improving its clarity;
- we improved the cohesion of the manuscript by inserting transition sentences (highlighted in red) at the end of the Introduction, at the beginning of Sections 5 and 6;
- we strengthened the Conclusions section.
Additionally, we would like to note that Section 4 has been expanded in response to Reviewer 1's comments.
We thank you again for your time and attention to our manuscript.
Kind regards,
The Authors
