Quantitative Assessment of Airway Changes in Fibrotic Interstitial Lung Abnormality Patients by Chest CT According to Cumulative Cigarette Smoking

Round 1

Reviewer 1 Report

Authors quantify the airway changes in 18 healthy patients, and 54 patients with fibrotic interstitial lung abnormalities (fibrotic ILA) using CT chest data. The main parameters quantified the airway changes are “the inner luminal area, airway inner parameter, airway wall thickness, Pi10, skew, and kurtosis”. Statistical analysis used to compare the significant changes of all airway changes parameters quantified. There are some questions to improve the study below.
1. Writing needs more care, e.g., line 68, 85, and 104 are missing the full stop.
2. Line 112, English abbreviation, for example, airway inner (AI) parameter.
3. Line 11, Pi10 is not fully explain, for instance, “the square root of AWT for airways with an internal perimeter of 10 mm [Pi10]” see https://doi.org/10.1513/AnnalsATS.201806-424OC, and https://doi.org/10.1148/radiol.2021210972
4. In section 2.1, the study population was so confusing, it is best to clarify using the workflow diagrams, for example, first, total patients 72(18+54), then healthy 18 (what parameters quantified included/excluded), and altogether with 54 (what parameters quantified included/excluded) and main parameters selected in the conclusion and abstract. 
5. Actually, this work is interesting, however, the comparison of both airway changes of healthy and lung disease. See Table 1 and Table 2 what is the difference, and the Fibrotic ILA in Table 2 was expanded 3 domains: Light, Moderate, and Heavy. What criteria for classifying these 3 domains? Can we group it into 1 Table or Better explained the data analysis in other ways, as they seemed reproduced (also, the control is not equal in Table 1 and Table 2).
6. Definition of abbreviations at Lines 141 and 163, it is better, to create the new table, and explain in a new section.
7. Figure 1 and Figure 2 is repeated the presentation. Can we group it into one picture as Figure 1.  
8. Discussion was not explaining the results and abbreviation, for example, at line 191 on Pi10.
9. Discussion should be improved to compare with the previous research.
10. Conclusion was so short, and it was not support with the results of the findings.
11. Introduction can also be improved by reviewing more recent papers on quantification of ILA found on chest CT.
12. In section, 2.5. Quantitative CT Evaluation, it would be great to give one picture for the authors quantified Pi10, or AI, or AWT using the APOLLO software.
13. Authors uncovered one figure to compare the 3D of normal lung and diseased lung.

Author Response

Thank you very much for your good comments on my paper. We have corrected the points pointed out by the reviewer with our best efforts, and please review the revised paper with a generous heart.

Authors quantify the airway changes in 18 healthy patients, and 54 patients with fibrotic interstitial lung abnormalities (fibrotic ILA) using CT chest data. The main parameters quantified the airway changes are “the inner luminal area, airway inner parameter, airway wall thickness, Pi10, skew, and kurtosis”. Statistical analysis used to compare the significant changes of all airway changes parameters quantified. There are some questions to improve the study below.

1Q. Writing needs more care, e.g., line 68, 85, and 104 are missing the full stop.

Answer) Thank you again for your good comment. I inserted the full stop at line 68, 85, and 104
2Q. Line 112, English abbreviation, for example, airway inner (AI) parameter. A

Answer) I changed AI to AIP

3Q. Line 11, Pi10 is not fully explain, for instance, “the square root of AWT for airways with an internal perimeter of 10 mm [Pi10]” see https://doi.org/10.1513/AnnalsATS.201806-424OC, and https://doi.org/10.1148/radiol.2021210972

Answer) I changed the definition of pi 10 as follows. The Pi10 is derived by plotting the square root of the airway wall area against the internal perimeter of each measured airway, and using the regression line to calculate the square root of the airway wall area for a hypothetical representative airway with an internal perimeter of 10 mm.

4Q. In section 2.1, the study population was so confusing, it is best to clarify using the workflow diagrams, for example, first, total patients 72(18+54), then healthy 18 (what parameters quantified included/excluded), and altogether with 54 (what parameters quantified included/excluded) and main parameters selected in the conclusion and abstract. 

Answer) Thank you for your comment. I added the workflow diagram.

5Q. Actually, this work is interesting, however, the comparison of both airway changes of healthy and lung disease. See Table 1 and Table 2 what is the difference, and the Fibrotic ILA in Table 2 was expanded 3 domains: Light, Moderate, and Heavy. What criteria for classifying these 3 domains? Can we group it into 1 Table or Better explained the data analysis in other ways, as they seemed reproduced (also, the control is not equal in Table 1 and Table 2).

Answer) Thank you again for your good comment.

• As you know, fibrotic ILA is a very complicated CT findings. Symptoms also may vary depending on the extent of fibrosis on CT in Fibrotic ILA patients. This study is to find out whether there is a change in airways depending on the amount of smoking in fibrotic ILA patients. Table 1 attempts to determine whether there is a difference in PFT or QCT from the normal group and the group diagnosed with fibrotic ILA, and we know that, in this study, DLco decreased in fibrotic ILA and changed in skewness and Kutosis in QCT. However, Pi 10, the biomarker of the change of airway change, is normal. In Table 2, we analyzed fibrotic ILA patients again according to the amount of smoking, and Pi 10 only showed a change in heavy smoker with more than 20 pack-years. Therefore, in the case of smoker, fibrosis occurs in lung parenchyma first, and airway change can be measured only when the amount of smoking is above a certain level. This is the difference between table 1 and 2.
• For the classification of the smoking state, see reference 16.
• The control group in Tables 1 and 2 was corrected because the numbers were incorrectly entered.

6Q. Definition of abbreviations at Lines 141 and 163, it is better, to create the new table, and explain in a new section.

      Answer) The abbreviations at lines 141 and 163 have already been explained in M&M, and tables 1 and 2 have already been described.

7Q. Figure 1 and Figure 2 is repeated the presentation. Can we group it into one picture as Figure 1.  

      Answer) Figure shows the difference in Pi 10 according to the difference in smoking amount. Figure 2 shows that Pi10 is normal in fibrotic ILA patients, and Figure 3 shows different cases of increased Pi10 in fibrotic ILA patients.

8Q. Discussion was not explaining the results and abbreviation, for example, at line 191 on Pi10. And 9Q. Discussion should be improved to compare with the previous research.

      Answer) The first paragraph of Discussion describes the interpretation of the results, and the abrevation for Pi10 has already been explained in M&M. In the second paragraph of the discussion, we described the usefulness of the results of this study compared to other papers.

9Q. Conclusion was so short, and it was not support with the results of the findings.

      Answer) As you requested, I have added the conclusion in detail based on the results data.

10Q. Introduction can also be improved by reviewing more recent papers on quantification of ILA found on chest CT.

      Answer) There are few recent papers on quantification of ILA. We have a QCT for COVID-19pnumonia's ILA findings, but we regret that this paper is different from our research in study population and QCT methods, so we cannot add the latest paper to be cited in introduction.

11Q. In section, 2.5. Quantitative CT Evaluation, it would be great to give one picture for the authors quantified Pi10, or AI, or AWT using the APOLLO software. 13. Authors uncovered one figure to compare the 3D of normal lung and diseased lung.

   Answer) I added a figure for quantified Pi 10 in fibrotic ILA group

Author Response File: Author Response.docx

Reviewer 2 Report

The results of this study are interesting, but there are major concerns about study design.

The extent of interstitial lesions in each patient has not been assessed in this study. It is also unclear what kind of pattern the interstitial lung disease is, such as the UIP pattern and the probable UIP pattern. Fibrosis causes thickening of the peribronchial bundle and traction bronchiectasis, which has a non-negligible effect on the evaluation of the airway by QCT. Therefore, it is necessary to evaluate how the degree of fibrotic lesions, honeycombing lungs, ground glass lesions, etc. affects various QCT parameters, and after excluding those effects, the association with smoking should be evaluated. A major revision of the study design is required.

 

Other minor comments are listed below.

 

“Skew” should be revised to “skewness”.

 

LAA cannot distinguish between emphysema lesions and honeycombing. How did you distinguish between the these lesions?

 

You excluded “non-specific interstitial pneumonia (NSIP) without emphysema. Please describe the reason.

Author Response

Thank you very much for your good comments on my paper. We have corrected the points pointed out by the reviewer with our best efforts, and please review the revised paper with a generous heart.

The results of this study are interesting, but there are major concerns about study design.

1Q. The extent of interstitial lesions in each patient has not been assessed in this study. It is also unclear what kind of pattern the interstitial lung disease is, such as the UIP pattern and the probable UIP pattern. Fibrosis causes thickening of the peribronchial bundle and traction bronchiectasis, which has a non-negligible effect on the evaluation of the airway by QCT. Therefore, it is necessary to evaluate how the degree of fibrotic lesions, honeycombing lungs, ground glass lesions, etc. affects various QCT parameters, and after excluding those effects, the association with smoking should be evaluated. A major revision of the study design is required.

Answer) Thank you again for your good comment. I think that supplementary explanation is necessary for the study design. Everything you said is correct. This study was conducted on patients with ILA findings on CT, and whether these patients were smoking-related fibrosis, probably UIP, or UIP was not separately analyzed. Also, the extent of fibrosis was not analyzed separately. However, after excluding fibrotic lesions such as honeycomb or GGA as you asserted, the association between smoking and Pi10 has already been published in my research in the past (CT Quantification of Lungs and Airways in Normal Korean Subjects. Kim SS, Jin GY, Li YZ, Lee JE, Shin HS.Korean J Radiol. 2017 Jul-Aug;18(4):739-748; In Korean subjects with normal spirometry and visually normal chest CT, there may be significant differences in QCT parameters according to sex, age, and smoking history). Based on the experience of this study, this study conducted a study on Pi 10 after correcting for age and gender as a control group. Also, although some patients were histologically diagnosed with UIP, because most patients had mild interstitial lung abnormality, so no histological diagnosis was made. As a preliminary study, analysis was performed by dividing it into SRIF, UIP, and probably UIP, but the extent of interstitial lung abnormality did not affect airway remodeling in our study as it was a small number. Miller's paper reported (ref 13#) that COPD with ILA or IPF had an increased Pi10 level compared to patients without ILA. However, it is common to have no COPD even with fibrotic ILA, and I have clinically experienced cases where Pi 10 is normal even with fibrotic ILA. Looking at the results of our study, even if there is fibrotic ILA on CT, airway remodeling is highly likely to occur only in heavy smokers, so I think these results are helpful research results for fibrotic ILA management in the future. When studying the association between smoking and fibrosis in ILA or IPF, it is very easy to rule out smoking factors. However, the extent of fibrosis, including honeycomb, cannot ultimately resolve the differences between observers. This is the reason why classification of fibrotic ILA was not performed in this study. As you know, fibrotic ILA cannot be diagnosed simply with probably UIP and UIP. Regarding the point of study design, I will describe the limitation.

 2.“Skew” should be revised to “skewness”.

Answer) thank you for your point. I corrected all skew to skewness

 3. LAA cannot distinguish between emphysema lesions and honeycombing. How did you distinguish between these lesions?

Answer) The limitation is that LAA using the APOLLO software cannot be distinguished from emphysema and honeycomb when performing QCT for ILD or fibrotic ILA disease. In this study, LAA was an insignificant factor, so the contents of LAA were deleted.

 4. You excluded “non-specific interstitial pneumonia (NSIP) without emphysema. Please describe the reason.

Answer) Thank you again for your good comment. I added the reason for exclusion. “Idiopathic non-specific interstitial pneumonia (NSIP) without emphysema because ground-glass without reticular abnormalities, lung distortion, traction bronchiectasis, honeycombing (n = 3)”

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

This manuscript has been sufficiently improved to warrant publication in Tomography.