The Feasibility of Combining 3D Cine bSSFP and 4D Flow MRI for the Assessment of Local Aortic Pulse Wave Velocity

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Congratulations for your interesting work and especially for the fact that you present a negative result, meaning the fact you found that PWVQA maps show high variability. This is very important to reduce the bias of publication of positive findings only. Some suggestions for improvement:

• Although the abbreviations are presented in a separate page, please describe all abbreviations with the whole term mentioned for the first time in the manuscript. For example, line 83 terms TR, TE, etc.
• You provide a thorough description of the techniques in the methodology and probably this is suitable, as this data are too specialized. However, please mention some parts in the methodology, which are necessary in a separate paragraph with a separate subtitle. Make clearer the outcomes of your study, how you chose the subjects studied (was there any matching for their demographic data, etc?)
• It would be preferable to present the figures after the paragraph of the results you mention the respective findings, as this is more convenient for readers instead of scrolling up and down the manuscript to find the related data.
• In the discussion I suppose that you could use more citations of studies comparing their results to your findings.
• Regarding any future perspectives, would not it be a possible suggestion of conducting the same study on the abdominal aorta or iliac arteries, which are more distally to the heart compared to aortic root and descending aorta? Maybe, in these parts the heart beats do not affect the findings and PWV includes also abdominal aorta and iliac arteries.

Author Response

Reviewer 1

Congratulations for your interesting work and especially for the fact that you present a negative result, meaning the fact you found that PWVQA maps show high variability. This is very important to reduce the bias of publication of positive findings only.

We thank the reviewer for appreciating our work and agreeing with us that is important to publish negative results to reduce the bias of publications of positive findings only. 

Some suggestions for improvement:

• Although the abbreviations are presented in a separate page, please describe all abbreviations with the whole term mentioned for the first time in the manuscript. For example, line 83 terms TR, TE, etc. 

R1.1 We have adjusted the manuscript accordingly.

• You provide a thorough description of the techniques in the methodology and probably this is suitable, as this data are too specialized. However, please mention some parts in the methodology, which are necessary in a separate paragraph with a separate subtitle.

R1.2 Thank you for this suggestion. We have adjusted the manuscript accordingly

•  Make clearer the outcomes of your study, how you chose the subjects studied (was there any matching for their demographic data, etc?)

R1.3 We have added in the statistical paragraph: “The outcomes of the study are the area, velocity, flow and PWV-QA and the correspondence between scan and rescan values. Furthermore, the correspondence between PWV-QA and PWV-WC are assessed.” The control and Marfan subjects were chosen from a larger database [15] such that sex was matched. This sentence was added to section 2.1. 

• It would be preferable to present the figures after the paragraph of the results you mention the respective findings, as this is more convenient for readers instead of scrolling up and down the manuscript to find the related data.

R1.4 We have adjusted the manuscript accordingly (not annotated).

• In the discussion I suppose that you could use more citations of studies comparing their results to your findings.

  R1.5 We have adjusted the manuscript accordingly and added more 4D flow PWV studies to the discussion for comparison: “PWV measurements for the entire thoracic aorta based on 4D flow MRI have been reported to increase with age from 5.4 m/s in young adults (45 years) to 7.2 m/s in midlife (45-65 years) and 9.4 in later life (>65 years) [12]. Another study found PWV values of 4.9 m/s in 20-30 year-olds and 8.1 m/s in 70-80 year-olds [11].”   

• Regarding any future perspectives, would not it be a possible suggestion of conducting the same study on the abdominal aorta or iliac arteries, which are more distally to the heart compared to aortic root and descending aorta? Maybe, in these parts the heart beats do not affect the findings and PWV includes also abdominal aorta and iliac arteries. 

R1.6 We thank the reviewer for this interesting suggestion. Indeed, the motion of the abdominal aorta and iliac arteries is substantially reduced compared to the thoracic aorta as these vessels are located more distally and are less affected by heart motion. We added a future work section in the Discussion where we state that it would be interesting to investigate local PWV in healthy abdominal aortas and those with aneurysms: “Also, it would be interesting to evaluate the performance of the local PWV methodology in abdominal aortas, where the effect of cardiac motion will be reduced. However, minimal area change and reflective waves are expected to also have an effect on reproducibility in this region. Both would warrant even higher resolution 4D bSSFP and 4D flow MRI to capture these changes”

Reviewer 2 Report

Comments and Suggestions for Authors

The article proposes a novel imaging method that combines 3D cine bSSFP with 4D flow MRI to assess the local aortic pulse wave velocity. Preliminary verification was conducted in healthy volunteers and patients with Marfan syndrome. This study provides new ideas for the quantitative analysis of local aortic function. However, after carefully reading and evaluating your research results, I found that there is still room for further improvement in some aspects of the paper. Here are some suggestions and opinions that I hope will be helpful for you to further optimize the paper.

 

1. In the introduction section of the article, the background of the existing PWV measurement methods was reviewed. However, there is still a lack of further summarization of the shortcomings of these current methods in various aspects. If the technical bottlenecks existing in this research field could be further elaborated, it would be more conducive to highlighting the value of this study.

 

2. The sample size of this study is relatively small, consisting of only 6 healthy volunteers and 6 Marfan syndrome patients. This may result in a relatively weak generalizability of the statistical results. In future studies, it would be advisable to increase the sample size to enhance the accuracy of the research conclusions.

 

3. The article used the QA method to estimate the local PWV, but the results showed that there was high variability in the measurements at the aortic root and the descending aorta region. In future studies, it would be better to further analyze the reasons for these high fluctuations and explore possible improvements.

 

4. In Section 2.3 of the article, it is mentioned that "The fit was considered sufficiently accurate if the R²exceeded 0.85". However, the article has not yet explained how the threshold of 0.85 was determined.It would be preferable to revise the methodology with explicit justification for the R²> 0.85 fitting threshold, thereby enhancing the study's methodological rigor..

 

5. Some of the images in the text (such as Figure 1, Figure 6, etc.) have relatively low resolution, resulting in blurriness when enlarged. The details in the images are not clear enough, which may affect the readers' accurate understanding of the results and their overall reading experience. It is suggested that you consider providing higher-resolution image versions in the revised manuscript to address this issue.

 

6. The article has not clearly stated whether a correction mechanism for the subject's body movement was implemented between the 4D bSSFP and 4D flow scans. Although the authors mentioned image registration and alignment processing, it would be better to provide further clarification on whether the potential errors caused by such body movements were taken into account, as this would help enhance the credibility of the results.

 

7. In Figures 4 and 5, there are some measurement points with extremely high PWV values. These abnormal values may mislead the subsequent statistical analysis and spatial distribution map. If you can conduct further analysis on these abnormal values, it will significantly enhance the rationality of your research results.

Author Response

Reviewer 2

The article proposes a novel imaging method that combines 3D cine bSSFP with 4D flow MRI to assess the local aortic pulse wave velocity. Preliminary verification was conducted in healthy volunteers and patients with Marfan syndrome. This study provides new ideas for the quantitative analysis of local aortic function. However, after carefully reading and evaluating your research results, I found that there is still room for further improvement in some aspects of the paper. Here are some suggestions and opinions that I hope will be helpful for you to further optimize the paper.

We thank the reviewer for the useful comments that further improve our work. 

1. In the introduction section of the article, the background of the existing PWV measurement methods was reviewed. However, there is still a lack of further summarization of the shortcomings of these current methods in various aspects. If the technical bottlenecks existing in this research field could be further elaborated, it would be more conducive to highlighting the value of this study.

R2.1 Thank you for this comment. The main issue for existing methods is that these measurements are for a larger part of the aorta at once instead of locally, which is what we are interested in. Additionally, in transit time based methods a high temporal resolution is very important, which makes it challenging for 4D flow. In our method we have both high spatial and temporal resolution, and in theory a method for measuring a PWV measure per slice through the aorta. We have further emphasized this in the introduction: “For accurate and regionally or locally specific PWV measurements, high spatial and temporal resolution anatomical and 4D flow MRI need to be acquired, and the dynamic nature of the aorta needs to be taken into account.”  

2. The sample size of this study is relatively small, consisting of only 6 healthy volunteers and 6 Marfan syndrome patients. This may result in a relatively weak generalizability of the statistical results. In future studies, it would be advisable to increase the sample size to enhance the accuracy of the research conclusions.

R2.2 We agree with the reviewer that it would be interesting to evaluate our method in larger cohorts. We added a future work paragraph that contains the sentence: “Future work will focus on application in larger cohorts, as the poor reproducibility of local PWV can be reduced by averaging the 3D maps as has been shown in previous work for wall shear stress [37] such that differences between patient groups and healthy volunteers can be shown at group-level.”

3. The article used the QA method to estimate the local PWV, but the results showed that there was high variability in the measurements at the aortic root and the descending aorta region. In future studies, it would be better to further analyze the reasons for these high fluctuations and explore possible improvements.

R2.3 Indeed, future research will focus on improving the method. We expect that an advanced 4D bSSFP sequence that encompasses the left ventricle and outflow tract and has appropriate segmentations would better enable local PWV measurements of the aortic root. For the descending aorta, measurements can potentially be improved by increasing the isotropic spatial resolution of 4D bSSFP MRI to better detect the area changes.  We have now further emphasized this in the final paragraph of the Discussion.  

 

4. In Section 2.3 of the article, it is mentioned that "The fit was considered sufficiently accurate if the R²exceeded 0.85". However, the article has not yet explained how the threshold of 0.85 was determined. It would be preferable to revise the methodology with explicit justification for the R²> 0.85 fitting threshold, thereby enhancing the study's methodological rigor..

R2.4 Thank you for this remark, we agree this can be expanded upon. We have added an explanation as to how the threshold of 0.85 was determined.  “Given that 3-4 points on the systolic upstroke were used for fitting, and to ensure a balance between goodness of fit and keeping as many PWV values as possible, the fit was considered sufficiently accurate if R² >0.85”

5. Some of the images in the text (such as Figure 1, Figure 6, etc.) have relatively low resolution, resulting in blurriness when enlarged. The details in the images are not clear enough, which may affect the readers' accurate understanding of the results and their overall reading experience. It is suggested that you consider providing higher-resolution image versions in the revised manuscript to address this issue.

R2.5 We believe this is a result of the format that Applied Sciences provides for review. We can assure the reviewer that images are of high resolution and will be printed as such upon publication.

6. The article has not clearly stated whether a correction mechanism for the subject's body movement was implemented between the 4D bSSFP and 4D flow scans. Although the authors mentioned image registration and alignment processing, it would be better to provide further clarification on whether the potential errors caused by such body movements were taken into account, as this would help enhance the credibility of the results.

R2.6 We stated that bulk motion between 4D bSSFP and 4D flow was qualitatively assessed on through-plane velocity images (4D flow) overlaid with segmentation planes (4D bSSFP). We did not find any bulk motion in this small cohort. Potential errors were thus not taken into account as there weren’t any. In future work when the methodology will be applied to larger cohorts, qualitative assessment is not feasible and a registration step is needed to automatically align the 4D bSSFP and 4D flow scans. This has been added to the manuscript in the limitations section: “Future studies in larger cohorts should incorporate registration of the bSSFP and 4D flow segmentations before further quantification of PWV_QA”.

7. In Figures 4 and 5, there are some measurement points with extremely high PWV values. These abnormal values may mislead the subsequent statistical analysis and spatial distribution map. If you can conduct further analysis on these abnormal values, it will significantly enhance the rationality of your research results.

R2.7 We agree that the abnormal values are unwanted, but inherent to the current quality of this method. This is also part of the ‘negative result’ which we are conveying. We are not sure what the reviewer means by “further analysis”. We proposed a method of deriving PWV_QA using 4D bSSFP and 4D flow MRI, and we reported earnestly that it performs sub-optimally for reasons mentioned in the paper. Supplementary Video S2 shows four example planes with corresponding fit plots, which also contains one “abnormal” high value. Probable causes for this abnormal value are given in the discussion of the current work. It is unfeasible to visually assess all the fit plots and in some way adjust these (how?) if they provide high slopes. We believe there is no reason to investigate each extremely high PWV value as the cause and the solution are unknown but based on how the fits look we have suggested some potential causes.

 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Authors have made the changes the reviewers asked. Congratulations for your work again.