Preliminary Results in the Investigation of In Vivo Iliac and Coronary Flow Collision, Vortex Formation, and Disorganized Flow Degeneration: Insights from Invasive Cardiology Based on Fluid Mechanics Principles and Practices

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This manuscript presents the preliminary results of a study focusing on the hemodynamics in the coronary arteries. Using a new dynamic angiographic protocol, the flow regime was observed and analyzed, including laminar, collision, and retrograde flow conditions. An AI approach is also introduced to detect retrograde flow from the image to aid analysis. There is no doubt about the amount of work and the novelty of this study. However, the manuscript can be improved, and there are a few issues. Please see my comments below:

1. The title should be revised. The specific vessel (coronary) should be included. I don't understand why "invasive" is in the title as I don't see open surgeries. Is the contrast injection invasive?

2. In the abstract, the sentence about the AI sub-study (55-56) seems to be out-of-context and confusing. This should probably be moved up to methods as it seems to be the method that helps with the analyses. 

3. The statement "Without retrograde flow, collisions and turbulence do not occur" on lines 73-74 is wrong. As mentioned below, the Reynolds number tells whether of the flow can be laminar or turbulent. So it is possible that a very strong flow with a high Reynolds number can be turbulent even if there is no retrograde flow or flow reversal. 

4. The Reynolds number is introduced from line 183 in the Methods. However, I don't see any calculation or results about this number in this study. If not, it's unnecessary to include it in the methods.

5. The "velocity and pressure measurements" method is introduced from line 196. However, I don't see these methods are used in this study. If not, please remove them.

6. Are any of the methods in "Practical Measurements" used in the study? If so, please specify it. If not, this paragraph can be removed or move up to the introduction. 

7. More literature about the study of flow and hemodynamics should be included in the introduction section. There are abundant of studies focusing on this topic, including in vivo, in vitro, and in silicon studies using CFD models.

8. In the results of the AI prediction from 301-304 seems to be much lower than the accuracy reported in the paragraph above. Please explain why this happens. 

9. A lot of the analyses in the study are based on visual observation as described from line 205. However, this approach can be very subjective, and it appears to me that the flow regimes can be hard to tell based on the angiogram images as included in the results. Without some guide of the text, I can not recognize the flow regimes solely from the images. This raises concerns about the repeatability of the method and the variances among the different observers. The authors should address this concern. 

Author Response

Please see attached

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This article presents a compelling study that integrates fluid mechanics with in vivo vascular imaging observations, offering new insights into arterial flow phenomena. The findings reveal various flow patterns, including laminar flow, retrograde flow, and flow collisions within blood vessels, and explore their potential impact on endothelial injury and atherosclerosis. Additionally, the study highlights the important correlation between these flow phenomena and blood pressure, underscoring the critical role of hypertension management in preventing adverse hemodynamic events. Before considering acceptance, there are several aspects that the authors should address or elaborate on with more detailed explanations.

1. Line 248 to 255. The article should clarify the dimensions of the segmented images and provide justification for selecting a sliding window size of 20x20 pixels. This might involve discussing the size of the features being analyzed that influenced by this choice.

2.Line 455 to 456. The author could elaborate on physiological and anatomical factors that make retrograde flow more common in the iliac artery compared to the coronary artery system. This could include differences in vessel geometry, pressure gradients, and the presence of bifurcations or branching patterns that influence flow dynamics differently in these arteries.

3.Line 455 to 456. How to prevent the retrograde flow?

4. Line 475 to 484 and Figure 12. The author should consider how vascular compliance, or the ability of blood vessels to expand and contract with pressure changes, differs from rigid pump and pipeline models. Discussing how compliance affects blood flow dynamics, particularly in how it might dampen or amplify wave reflections within the arterial system, could provide valuable insights. 

Author Response

Please see attached

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors have addressed some of my concerns and improved the manuscript. I would like to follow up with the comment 8 of my last review. I don't think the response answered my question about why the AI model performed significantly worse on the validation than the training set. I would like to see a detailed breakdown, e.g., the confusion matrices of both training and validation. I'm worried that the AL model is over-fitted, which may explain the deficit. Also, the accuracy, sensitivity, specificity should be reported for both training and validation. 

Author Response

Please see attached. Our team appreciates the comments in your review

The new changes in the manuscript are on lines 285-288

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Accept

Author Response

Please see attached

Author Response File: Author Response.pdf