Write a comment

This manuscript addresses an important and timely topic but has significant limitations that affect its suitability for publication in its current form. The primary concerns include: inadequate methodological description for a review article, lack of systematic approach to literature synthesis, insufficient critical appraisal of evidence, and imbalanced coverage of topics. The manuscript reads more as an educational primer than a scholarly review that advances understanding of the field.

The writing quality is acceptable though occasionally informal. The authors appear to have relevant expertise based on their affiliations, though the large author list (15 authors) relative to the manuscript length and scope is notable.

I do not recommend acceptance in the current form. With major revisions addressing the methodological concerns, narrowing of scope, and deeper critical analysis, the manuscript could potentially make a contribution to the literature.

Major issues

• Methodological Concerns with Literature Search Strategy

The methods section indicates that articles were screened "since inception till March 25th" and 44 articles were included "after careful consideration." This description is insufficiently detailed for a review article. The authors should provide explicit inclusion and exclusion criteria, describe the screening process (e.g., how many reviewers, how disagreements were resolved), report the total number of articles identified and excluded at each stage, and ideally present a PRISMA flow diagram. Without this information, it is impossible to assess whether the literature search was comprehensive or whether selection bias may have influenced the findings. The term "careful consideration" does not constitute a reproducible methodology.

• Scope and Depth Imbalance

The manuscript attempts to cover an extremely broad range of topics, including basic AI definitions, ECG interpretation, echocardiography, cardiac CT/MRI, wearable devices, CRISPR gene editing, and bioprinting. This breadth comes at the expense of depth. For instance, the discussion of AI in echocardiography is limited to a single sentence referencing one study (Reference 28), despite this being an area with extensive published literature. The authors should either narrow the scope to allow for more thorough treatment of each topic or explicitly frame this as a high-level overview with appropriate acknowledgment of its limitations.

• Critical Appraisal of Evidence is Lacking

The review presents AI applications in an overwhelmingly positive light without sufficient critical evaluation. For example, the statement that "AI algorithms...have demonstrated remarkable success in interpreting and analyzing medical imaging data" (page 5) is not accompanied by discussion of sensitivity/specificity metrics, comparison with clinician performance, validation study designs, or potential failures. A high-quality review should address questions such as: What is the current level of evidence for clinical benefit? How many AI tools in cardiology have received FDA clearance? What are the documented failure modes? Have any prospective randomized trials demonstrated improved patient outcomes?

• Table 1 Requires Revision

Table 1 (Fundamentals of AI) lists basic definitions with citations [1-10] for each entry. This citation format is problematic because it suggests all 10 references support each definition, which is unlikely. More importantly, the definitions provided are elementary and could be found in any introductory textbook. The authors should either provide more cardiology-specific definitions with appropriate singular citations or consider whether this table adds value to the manuscript.

• Missing Discussion of Regulatory and Implementation Frameworks

While the future directions section briefly mentions "FDA approval," the manuscript does not adequately address the current regulatory landscape for AI in cardiology. The FDA has cleared numerous AI-enabled cardiac devices (e.g., for ECG interpretation, coronary artery calcium scoring), and discussion of these approvals, their regulatory pathways, and post-market surveillance requirements would strengthen the clinical relevance of this review.

• Limited Discussion of Bias and Health Equity

The manuscript mentions that "quality and representativeness of healthcare data are crucial for ensuring that AI models are accurate and unbiased" (page 6), but does not elaborate on this critical issue. Published studies have demonstrated that AI algorithms can perpetuate or amplify healthcare disparities. Given the importance of health equity in contemporary cardiovascular medicine, a more substantive discussion of algorithmic bias, underrepresentation of certain populations in training datasets, and strategies for ensuring equitable AI deployment is warranted.

• Figures Require Improvement

Figure 1 attempts to outline the AI training process but combines multiple concepts (learning types, data sources, functions, outputs) in a manner that may confuse rather than clarify for the intended clinical audience. The flow of information is not intuitive, and the relationship between elements is unclear. Figure 2 is more successful but would benefit from a clearer legend explaining each preprocessing step.

Minor issues

• The abstract states this is "the initial review in a series of reviews regarding A.I in Cardiology." If subsequent reviews are planned, the authors should clarify what topics this review covers versus what will be addressed in future publications.

• Inconsistent abbreviation usage: "A.I" versus "AI" appears throughout the manuscript. The authors should select one format and use it consistently.

• Page 2: The phrase "we as physicians and cardiologists have tried to put our perspective" is informal for scientific writing. Consider revising to more objective language.

• Page 3: The sentence "A.I, often perceived as human-like intelligence exhibited by a machine" conflates AI with artificial general intelligence. Most AI applications in cardiology are narrow AI performing specific tasks, and this distinction should be clarified.

• The discussion of CRISPR and bioprinting in the future directions section feels disconnected from the main AI focus of the review. While AI may have applications in these areas, the connection is not clearly articulated.

• Reference 38 (Lundberg and Lee) is cited as "ArXiv. 2017, May:" with incomplete information. This should be corrected with full citation details.

• The acknowledgments state "None," yet the author contribution section describes extensive collaborative work. Consider whether any individuals, institutions, or funding sources should be acknowledged.

• Several references are to editorials or opinion pieces rather than primary research or systematic reviews. The authors should ensure the evidence base is as strong as possible.

Competing interests

The author declares that they have no competing interests.

Use of Artificial Intelligence (AI)

The author declares that they did not use generative AI to come up with new ideas for their review.