True Non-Contrast Phase versus Virtual-Non Contrast: “Lights and Shadows” of Dual Energy CT Angiography in Peripheral Arterial Disease

Round 1

Reviewer 1 Report

 

In this manuscript, Chiara Floridi describes the comparative study between "virtual non-contrast" (VNC) and true non-contrast phase (TNC) in the detection of peripheral arterial disease (PAD) implemented by University Politecnica delle Marche. It aimed to evaluate the diagnostic reliability and accuracy of VNC images obtained from arterial scan reprocessing compared to baseline acquisition obtained without intravenous iodinated contrast (TNC). It’s concluded that VNC may be considered as a promising substitute based on 40 consecutive extremities dual-energy CT angiography (DE-CTA) scans.

 

I appreciate the importance of comparing the image quality of TNC to VNC images to reduce the amount of radiation dose in PAD detection. I think the paper deserves to be published in Applied Sciences. However, there is room for improvement in this manuscript on document organization Text. A few suggestions, mainly concerning the important findings, are given below.

 

Major Concerns:

1. First, I am concerned about the limited information contained in the figures and tables in this article.

A) It’s recommended that data presented in a plot or chart shows data distribution clearly, e.g. dot plots, box-and-whisker plots, violin plots. Considering that "consisting of higher values of few HU in VNC", it may be better to overlay the corresponding data points (as dot plots) with bar charts.

B) An error bar can be added to the diagram to visualize data bias. It’s recommended the exact number of biologically independent experiments to be provided in figure legends.

C) It might be better to similarly use the abbreviations directly after they first appear in figures and table legends, e.g. AIC, AFS, DLP.

D) It is suggested that the CT scanning parameters of TNC acquisition and VNC reconstructions described in line 123-141 could be reorganized into a new table.

 

2. Second, there may be a confusing situation "regardless of the double-blinded and independent evaluations" with the following questions.

A) It appears that radiologist Giacomo Agliata (GA) both reconstructed and reviewed the images, which may not be consistent with a blinded observer, according to the Author Contributions: “methodology, G.A.; software, G.A.;”. Please have a check.

B) The same problem also occurs in line 180 "two radiologists blinded to clinical information and previous independently evaluated reports". It may not be sufficient to reliably conclude that "performed by the two blinded observers were illustrated in Figure 4."

 

3. Third, there are concerns about precise reasoning, as some sentences are not clear enough to support conclusions or inconsistent in context.

A) According to the results in Table 2, it’s puzzling to draw a conclusion that “VNC reconstructed images may have resulted in a 54% reduction of the administered mean effective dose” (line 267-268).

B) In addition to qualitative analysis of CT image quality, readers may be curious about the ultimate performance of VNC for clinical diagnosis of PAD. On account of “patients had to be previously submitted to a vascular evaluation for suspected chronic limb ischemia”, what’ s the diagnostic performance on VNC for clinical diagnosis with different blinded observers.

C) Lines 78-86, the previous text is about "several applications of DE technology", and the following text is about problems that algorithms can solve. It didn't seem logical and convincing.

 

4. In my opinion, some statements are unclear or difficult to read. For example:

A) Line 256, "In VNC images did not occur incorrect removals of calcified plaques, which were, at most, slightly attenuated” is difficult to read.

B) Line 143, “were used and in all cases an i.v. injection of 40 mL of saline” is a bit complicated to read.

 

5. There seems to be room for improvement in the fluency and organization of the full text.

A) The paragraph structure of this article is rather casual and lacks consistency between paragraphs. For example, Line 49 -51, Line 78-92, Line 115-141.

B) It’s confused why italics are used on line 112-113.

C) It might be better to similarly use similar abbreviations directly after their first appearance, for instance, Hounsfield unit (HU), multi-planar reconstructions (MPR), field of view (FOV), region-of-interest (ROI).

D) Symbols should be given their full abbreviation, for instance, standard deviation (SD) in line 172,175, when they first appear in text and figure legend.

 

6. I am minimally concerned about the critical novelty of this study, considering that "our results are limited to a DE-CTA dataset and to this vendor-specific Virtual Unenhanced analysis algorithm."

A) Some clinical studies have shown its ability to diagnose the various pathologies discussed in this article. This study is "one of the few to evaluate arterial lower limb vessels from the aortic axis to popliteal segments." The biophysical differences and clinical value of this study from the aortic axis to popliteal segments need to be emphasized.

B) It is recommended that the modified calcium subtraction algorithm (or other updated algorithm) be added to post-processing and image reconstruction on PAD evaluation.

C)     The hypothesis on “reduced by the preference for contrast media with lower iodine concentrations” is worth verifying in future work.

 

7. The manuscript would be improved if the same information could be reduced:

A) Line 215-228, The detailed description of five arterial segments is repeated with Table 2 on comparison between TNC and VNC images. The detailed description of the attenuation difference in line 234-241 is also repeated with Figure 3.

B) The same information is repeated in different paragraphs, e.g. line 166-167 "avoiding vessel wall, vessel defects, plaques, and stent materials " is repeated in line 223 "ROIs were made as large as possible while avoiding calcifications, plaques, and stent material.”

C) The same statement on line 174 "mean Hounsfield unit (HU) and corresponding standard deviation (SD)" is also repeated in line 172.

Minor Concerns:

1.         Line 52, "CTA of the lower extremity runoff, which is nowadays the preferred choice” should be improved.

2.         Line 54-55, "CTA is also highly accurate in assessing… also evaluating… " may be better in "CTA is also highly accurate in assessing… and evaluating… "

3.         Line 270, “Table 2. RTa” ?

4.         Line 172-173, “Mean attenuation and image noise (one standard deviation) in HU were registered” means “were measured”?

 

 

Some statements are unclear or difficult to read. For example:

A) Line 256, "In VNC images did not occur incorrect removals of calcified plaques, which were, at most, slightly attenuated” is difficult to read.

B) Line 143, “were used and in all cases an i.v. injection of 40 mL of saline” is a bit complicated to read.

Minor Concerns:

1.         Line 52, "CTA of the lower extremity runoff, which is nowadays the preferred choice” should be improved.

2.         Line 54-55, "CTA is also highly accurate in assessing… also evaluating… " may be better in "CTA is also highly accurate in assessing… and evaluating… "

3.         Line 270, “Table 2. RTa” ?

4.         Line 172-173, “Mean attenuation and image noise (one standard deviation) in HU were registered” means “were measured”?

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

 

The aim of the study was to evaluate the diagnostic accuracy of virtual non-contrast imaging of DE-CTA compared to true non-contrast phase.

- Avoid using acronyms in abstract. Also, it is not necessary to put the words background, methods, results, and conclusion.

- In the initial section of methodology, make a general methodological diagram, so that the process can be better understood. And add a paragraph describing in a general manner.

-  Lines 112 and 113 should not be in italics.

- In quantitative image analysis, how do you ensure that the region of interest is drawn in the same position in both datasets? This is to avoid bias in the information.

- The first paragraph of the results section (lines 205 – 209) must be in section 2.1. Including table 1.

- Figures 2 - 4 must be made in a professional software.

- Line 324, put the reference number after the author Sauter et al. Similarly, on line 332.

 

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 3 Report

The paper has several weaknesses, and among them are listed below:

1) The title has the terms of "lights and shadows", but this terms are not been discussed in the text.  What does this terms means? 

2) The writing needs to be significantly improved.  Currently, the paper seems like a note, with most of the paragraphs are a single-sentence paragraph (i.e., paragraphs with only one sentence).  A good paragraph should has at least 3 sentences.

3) The abbreviations should be defined properly.  For examples, what is IRB in line 20? 

4) It is not clear whether the authors proposed new method, or just use the currently available method.  The contribution is not clear.  For example, the main thing in this paper is about "virtual-non contrast".  However, this approach is not been introduced properly.  Statements about virtual non-contrast (VNC) at lines 84 is too brief.  Thus, the paper is not a self-contained paper.  More description on how VNC works should be provided.

5)  At line 20, "... 40 patients (28 men; ...".   How about the remaining 12 persons?

6) At the end of Section 1, better to add one paragraph to describe the structure of the paper.

7) Better to have at least one paragraph in between Section 2 and Subsection 2.1.

8) Why lines 112 and 113 need to be typed in Italic?

9) In Section 2.2, describe why and how those values for the scanning parameters are selected.

10) Line 270, what is Table 2 RTa?

11) Figure 2 are presenting the same results presented in Table 2.  Better to remove this figure.

12) What is 0% above the pie-chart in Figure 3?

 

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 4 Report

 

In this manuscript, the diagnostic reliability of DE-CTA virtual non-contrast (VNC) images for the detection of peripheral arterial disease (PAD) has been evaluated and compared with the traditional true non-contrast phase (TNC). The study demonstrates that the VNC images have lower attenuation values than the TNC and it provides satisfactory diagnostic quality for PAD detection. The study presented is compelling and the manuscript is well-structured. However, there appear to be some crucial areas that require further clarification. The authors should address the following concerns for the manuscript to be considered for publication.

1.    Please clarify whether any form of respiratory or cardiac gating was used during image acquisition. Please discuss how motion artifacts were managed or mitigated in the study.

2.   Please elaborate on the rationale behind choosing the specified trigger (300 HU) and delay (8 s) for bolus tracking acquisition.

3.  Please clarify how the comparability of the readings by the two radiologists (Likert scale) has been ensured. Did they use any predefined criteria when assessing the quality of CT images? Or it is subjective. Furthermore, it has been stated that the radiologists were blinded to clinical information. Could you please clarify whether they were also blinded to each other’s evaluations to avoid bias?

4.   In the statistical analysis section, the author mentioned that the intra-class correlation coefficient (ICC) was employed for assessing the agreement between the two radiologists. Please provide more detail on which form of ICC was used and why it was chosen.

5.  Please discuss the modified algorithm for calcium subtraction mentioned in the study. How does this compare to the post-processing algorithm employed?

6.  The study demonstrates that the TNC images have higher attenuation values than the VNC images. Please clarify whether this difference influences the diagnostic applicability, could this potential difference in attenuation values alter the sensitivity or specificity of the diagnostic process?

7.  Please clarify how the results inform clinical practice. Can VNC imaging be used routinely instead of TNC, or are there specific scenarios where one is preferable to the other?

 

 

 

 

 

 

 

 

Author Response

Please see the attachments

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

- Thank you for presenting the general methodological diagram, however, the font is too small and it does not look good. Please modify the figure.

- Lines 220-224 and table 2 must be in section 2 in patien population, because you are describing the population. They are not results.

 

Author Response

2^nd Round Point-by-point Response

- Thank you for presenting the general methodological diagram, however, the font is too small and it does not look good. Please modify the figure.

The general methodological diagram was modified as kindly suggested by the reviewer.

- Lines 220-224 and table 2 must be in section 2 in patient population, because you are describing the population. They are not results.

Thank you for the kind comment. As suggested, patient general characteristics and the corresponding Table 2 were moved from Results section to Materials and Methods (2.1-Patient population and study design). Please see highlighted lines 130-133. Table numbers have been adjusted accordingly.

Author Response File: Author Response.docx

Reviewer 3 Report

The authors still not addressed some of my previous concerns.  For examples:

1) There are still some single-sentence paragraphs.  For examples, the paragraphs starting at line 51, line 61, line 79, line 97, and there are more...

2) The description of VNC is still not adequate.  For example, in line 88, "analyzed with specific algorithms", the authors should mention and describe on how this algorithm works.  Currently, the description is on what the VNC can do, but not on how the method can do it.

 

 

 

Author Response

Reviewer 3

2^nd Round Point-by-point Response

The authors still not addressed some of my previous concerns.  For examples:

1) There are still some single-sentence paragraphs. For examples, the paragraphs starting at line 51, line 61, line 79, line 97, and there are more...

We are sincerely sorry for this inattention. Another careful general revision of the text and of the indicated paragraphs has been made by the Authors. Please see highlighted lines 51-52, 58-60, 61-64, 72-73,80-83, 89-91, 92-95, 314-317, 334-337, 340-343, 354-357, 364-368, 381-384, 400-404.

2) The description of VNC is still not adequate.  For example, in line 88, "analyzed with specific algorithms", the authors should mention and describe on how this algorithm works.  Currently, the description is on what the VNC can do, but not on how the method can do it.

We sincerely thank the reviewer for the kind suggestion. We added a more precise description of VNC reconstruction technique. Please see highlighted lines 87-99.

Author Response File: Author Response.docx

Reviewer 4 Report

The author has addressed all my comments and concerns. 

Author Response

We sincerely thank the Reviewer for the consideration given to our study.

Round 3

Reviewer 3 Report

There are still paragraphs with only one sentence.  For examples, on line 100, 128, 131, 137, 152, 154, and many more.  Please check the manuscript carefully.