Collagen 1 Fiber Volume Predicts for Recurrence of Stage 1 Non-Small Cell Lung Cancer

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Please show a schematic of SHG microscope along with full specifications of microscope and laser.

 

What test was performed in figure 2 to get the p value. It does not look like it is <0.05. Please give full statistical equation and anlysis in suplementary. 

Author Response

Reviewer 1

 

Comment 1: Please show a schematic of SHG microscope along with full specifications of microscope and laser.

Response: We have provided a reference [1] (reference # 43 in the revised manuscript) that contains a schematic, in the revised manuscript.  Additionally, we have provided details of the microscope and laser in Methods.  Please see pages 3-4, lines 120-128 in the revised manuscript.

 

Comment 2: What test was performed in figure 2 to get the p value. It does not look like it is <0.05. Please give full statistical equation and analysis in supplementary.

            Response: Each point in Figure 2 represents the mean fiber volume for each patient in the non-recurrent and recurrent group obtained from an average of 6-18 randomly selected fields of view (FOVs).  We performed a one-tailed unpaired Student’s t-test and obtained a p value of 0.051.  We used a one-tailed unpaired Student t-test as we hypothesized that the recurrent tumors would have a denser Col1 fiber distribution compared to the non-recurrent tumor patient groups based on previously published data with other cancers including a recently published study with lung cancer that showed increased Col1 fibers in early stage lung cancer compared to normal tissue[2] (reference # 26 in the revised manuscript).  We used a t-test since the data did not seriously violate normality distribution assumption. This is clarified in the revised manuscript.

We have included Supplementary Figure 1 where the percent fiber volume was calculated from 6-18 randomly selected fields of view (FOVs) from each tissue block, and one block per patient was analyzed. To compare the percent fiber volume between patients with and without recurrence, we employed a random effects model. In this model, the tissue block served as the random effect (random intercept), while the cancer recurrence status was the fixed effect. Although the mean percent fiber volume from all FOVs from the same tissue block is approximately normally distributed, the individual percent fiber volumes across all FOVs are skewed. We thus log-transformed the individual FOV percent fiber volume to reduce skewness before fitting the random effects model. The model output indicates that recurrent patients exhibited a higher percent fiber volume than non-recurrent patients (two-sided test p-value = 0.078, one-sided test p-value = 0.039). Please see pages 4-5, lines 169-184, and Supplementary Figure 1 and corresponding figure legend in the revised manuscript.

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript by Kakkad et al. presents an exploration of type I collagen SHG signal as a means to predict recurrence of non-small cell lung carcinoma (NSCLC). The potential role of collagen as a biomarker for NSCLC recurrence is demonstrated. The authors aimed to validate their findings using gene expression from RNA sequencing data. I believe this topic is of great interest considering the high incidence and prevalence of lung cancer. The manuscript is well written and well structured. However, I have some concerns regarding data analysis which I outline below.

The investigation is not exhaustive. The manuscript presents a superficial analysis of collagen SHG signature for a limited number of cases. The analysis of SHG signal is limited to fiber volume. But why stop here, instead of carrying out a more in-depth geometric and textural analysis? Can alternative collagen SHG parameters also be predictive of recurrence? This would be worth showing, irrespective of trends.

Further, I have concerns regarding the RNA sequencing data presented by the authors. I believe it lacks relevance for this study. I understand the need for validation and understanding the origins of the morphological alterations observed in the SHG signals. However, as far as I understood, the SHG and RNA datasets are of different origins, i.e. they are not from the same group of patients. This raises the question of how representative the RNA dataset is of this cohort of patients?

Another key point refers to the description of statistical methods, which is very brief and lacks some details, including justification for test choice (why one sided and what was the directional hypothesis?), sample size, assumptions for the t-test (are the data homogenous and normally distributed?), correction methods, and effect size. This should be addressed by the authors.

In addition, it is also not clear how the values were obtained for Figure 2. Does each data point correspond to the average fiber volume for each patient? Was this obtained from a single or many regions of interest?

Additional questions/minor comments:

·         Have the authors considered polarization-resolved measurements (PIPO SHG)?

·         Abbreviations or acronyms must be defined on their first occurrence in the abstract or main body of the manuscript (e.g. NSCLC, CAF, etc).

In summary, I think this investigation is quite interesting, but, as it stands, not significant enough. The work can be significantly improved. Considering the lack of relevance of RNA data, the manuscript is left with fiber volume data obtained from SHG images, which I believe is not sufficiently compelling for publication.

Author Response

Reviewer 2

 

Comments and Suggestions for Authors

 

Comment 1: The manuscript by Kakkad et al. presents an exploration of type I collagen SHG signal as a means to predict recurrence of non-small cell lung carcinoma (NSCLC). The potential role of collagen as a biomarker for NSCLC recurrence is demonstrated. The authors aimed to validate their findings using gene expression from RNA sequencing data. I believe this topic is of great interest considering the high incidence and prevalence of lung cancer. The manuscript is well written and well structured. However, I have some concerns regarding data analysis which I outline below.

Response: We thank the Reviewer for these comments

 

Comment 2: The investigation is not exhaustive. The manuscript presents a superficial analysis of collagen SHG signature for a limited number of cases. The analysis of SHG signal is limited to fiber volume. But why stop here, instead of carrying out a more in-depth geometric and textural analysis? Can alternative collagen SHG parameters also be predictive of recurrence? This would be worth showing, irrespective of trends.

            Response: We agree with the Reviewer that this was an exploratory study.  We have mentioned this in the Discussion of the revised manuscript.  We did carry out textural analyses but these were not highly predictive of recurrence in the limited sample size study performed here.  We did not comment on this in the current exploratory study as an expanded study would provide greater confidence in this observation.  Please refer to page 9, line 280 and page 10, lines 339- 343 in the revised manuscript.

 

Comment 3: Further, I have concerns regarding the RNA sequencing data presented by the authors. I believe it lacks relevance for this study. I understand the need for validation and understanding the origins of the morphological alterations observed in the SHG signals. However, as far as I understood, the SHG and RNA datasets are of different origins, i.e. they are not from the same group of patients. This raises the question of how representative the RNA dataset is of this cohort of patients?

            Response: We understand the Reviewer’s concerns.  Because we used the H&E stained sections from the NIH NLST data we were unable to perform RNA sequencing on the same samples.  This is a limitation of our study as we mentioned in the Discussion.  Instead we used an independent data set that further confirmed the changes in collagen we observed with the NLST data set.  Please see page 10, lines 322-323 in the revised manuscript.

 

Comment 4: Another key point refers to the description of statistical methods, which is very brief and lacks some details, including justification for test choice (why one sided and what was the directional hypothesis?), sample size, assumptions for the t-test (are the data homogenous and normally distributed?), correction methods, and effect size. This should be addressed by the authors.

Response: We used a one-tailed unpaired Student t-test as we hypothesized that the recurrent tumors would have a denser Col1 1 fiber distribution compared to the non-recurrent tumor patient groups based on our previously published data with breast cancer.  The mean Col1 fiber data, averaged over all FOVs from the same patient, were approximately normally distributed.  We have clarified this in the revised manuscript.  Please see pages 4-5, lines 169-184, and Supplementary Figure 1 and corresponding figure legend in the revised manuscript.

 

Comment 5: In addition, it is also not clear how the values were obtained for Figure 2. Does each data point correspond to the average fiber volume for each patient? Was this obtained from a single or many regions of interest?

            Response: The mean fiber volume for each patient was obtained from an average of 6-18 randomly selected FOVs per patient. Each point in Figure 2 represents the mean fiber volume for each patient in the non-recurrent and recurrent group. We have further clarified this in the Methods and in the legend for Figure 2 in the revised manuscript.  Please see page 4, lines 124-128 and revised figure legend 2 in the revised manuscript.

 

Comment 6: Additional questions/minor comments:

• Have the authors considered polarization-resolved measurements (PIPO SHG)?
• Abbreviations or acronyms must be defined on their first occurrence in the abstract or main body of the manuscript (e.g. NSCLC, CAF, etc).

            Response: The Reviewer makes an excellent suggestion for us to consider for future studies.  We have mentioned this in the discussion.  Please see page 10, line 342 in the revised manuscript.

We have carefully checked the abbreviations and acronyms in the body of the manuscript.

 

Comment 7: In summary, I think this investigation is quite interesting, but, as it stands, not significant enough. The work can be significantly improved. Considering the lack of relevance of RNA data, the manuscript is left with fiber volume data obtained from SHG images, which I believe is not sufficiently compelling for publication.

            Response: We thank the Reviewer for their insightful comments.  We have revised the manuscript to address the concerns to the best of our ability to make the manuscript suitable for publication.

Reviewer 3 Report

Comments and Suggestions for Authors

·         Please explain what is SHG microscopy and how does it work in detection of COL1 fibers?
At what time points were the tissues obtained? Were they fresh? How old were they? And how were they stored before analysis was done?

·         Please provide a detailed explanation on how the molecular analysis was done. List reagents and detailed methodology for all the genes analyzed.

·         How was Col1 fiber volume quantified? Please define all metrics in detail. Please list the excitation and emission filters used for fluorescence microscopy.

·         Figure 1 shows the difference in Col1 fibers in recurrent and non-recurrent patients, did all non-recurrent patients have a similar fiber volume? Were there any cases of ambiguity? Were there any blind measurements done? Were there cross sections were Col1 fibers were not as dense in recurrent patients? Please write a section on potential challenges and inform readers of cases where no difference were noticed in recurrent and non-recurrent patient.

·         It would be helpful to add a few tissue sample analysis from healthy/ cancer-free patients in order to validate your hypothesis. What was the control used?

·         Please include patient numbers in the study. Attach a patient number by the side of each image, so readers have a reference as to which patient you are referring to. For example, non recurrent patients can be identified as NR 1, NR 2… And recurrent patients can be identified as R1, R2.. Since the patient size is so small, add a supplementary section showing tissue sections from each patient with the patient ID.

·         Figure 2, how is the percentage value obtained. Please provide a detailed explanation of that. Please provide justification on why a specific statistical test was used to determine significance. A small section on statistical methods would be helpful. Was the t-test parametric? Non-parametric? Define the distribution and identify the correct statistical test.

·         Did you do statistical power analysis? Is your sample size enough to draw conclusions about Col1 fibers? Please write a paragraph about this.

·         For the increased expression of Col1, fibronectin, laminin, nidogen-2, and aggrecan, it would be useful to provide more context about their roles in extracellular matrix (ECM) remodeling and how this correlates with cancer invasiveness.

·         The decrease in hyaluronan binding protein-2 expression is noted, but its potential implications should be discussed. How might this decrease affect tumor progression or patient prognosis

·         What are the different subtypes of CAFs? Were different subtypes of CAFS analyzed?

·         Are there any other phenotypes of the COL1 fibers that can be studied to further interpret the recurrence of disease? Such as Size, brightness, length etc

·         Please write a section on the mechanism and fuctional role of Col1 fibers in the progression/regression of the disease.

·         Have there been cases of recurrence where these Col1 fibers were not elevated?

·         Are figure3 and Figure1 images from same patients? Please have image identifiers for all patients as described in one of the comments above.

·         Figure 3, I see some sections in non-recurrent patients with long thicker patterns of Col1 fibers. For example, Figure 3A, non recurrent patient, at the 4’o clock region a zoom in might show long fibers. Figure 3, needs to show a bigger data set and different cross sections of COL1 fibers. Showing one very evident cross section may not be sufficient to identify recurrence in patients.

Overall this is a great study as it provides valuable insights into the differential gene expression between noninvasive and invasive lung adenocarcinoma, particularly focusing on ECM-related proteins. This is highly relevant given the importance of understanding tumor microenvironment interactions in cancer progression and potential therapeutic interventions. However, with addition of control group and addressal of the above mentioned comments, the study can be considered for publication.

Author Response

Reviewer 3

 

Comment 1: Please explain what is SHG microscopy and how does it work in detection of COL1 fibers?

At what time points were the tissues obtained? Were they fresh? How old were they? And how were they stored before analysis was done?

            Response: We have provided details regarding SHG microscopy and have provided a reference that details SHG microscopy for journal readers.  The tissue sections were H&E stained sections from the NIH NLST data base obtained during surgery.  These sections are formalin fixed and mounted on glass slides and therefore do not require any special storage conditions.  This provides the advantage of the ease of including SHG microscopy analysis of H&E stained sections as a companion diagnostic.  We have clarified these details in the revised manuscript.  Please see pages 3-4 lines, 120-124, and page 3, lines 106-110, in the revised manuscript.

 

Comment 2: Please provide a detailed explanation on how the molecular analysis was done. List reagents and detailed methodology for all the genes analyzed.

            Response: The molecular analysis was performed using publicly available expression data from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo).  The original transcriptomic study performed by RNA sequencing of 53 patient-derived RNA stage 1A lung adenocarcinoma samples (32 noninvasive and 21 invasive samples) was conducted by Yoo et.al., [3] (reference # 32 in the revised manuscript), and submitted to the GEO database for public use. We retrieved the submitted dataset (GSE166720) and analyzed differences in genes associated with ECM proteins, CAFs, immune checkpoints and T-cell markers in the noninvasive and invasive samples to further understand the potential causes and consequences of the Col1 changes observed with SHG microscopy in our study.  The dataset was analyzed using in-built software GEO2R (https://www.ncbi.nlm.nih.gov/geo/geo2r/)[4] (reference # 44 in the revised manuscript).  The interactive web tool GEO2R is based on R programing language with an in-built statistical program and graphic tools that allow identification of differentially expressed genes.  Genes expressed with at least > 0.5 or < 0.5 log2 fold change (~1.4-fold change) and a p adjusted value (padj) of <0.05 were considered significantly altered.  We have provided this information in the revised manuscript.  Please see page 4, lines 154-164 in the revised manuscript.

 

Comment 3: How was Col1 fiber volume quantified? Please define all metrics in detail. Please list the excitation and emission filters used for fluorescence microscopy.

            Response: We have provided the details in the revised manuscript. Briefly, our software quantifies the total Col1 fiber volume by first preprocessing the raw image to exclude noise and nonfibrillar shapes by using a shape filter as previous described[5-8] (references # 18-20, 27 in the revised manuscript).  The Col1 fiber structures extracted from the raw images were analyzed as the percent Col1 fibers per field of view. An excitation filter of 880 nm and an emission filter of 440 nm was used for the SHG microscopy.  We have provided these details in the revised manuscript.  Please see page 4, lines 146-150 in the revised manuscript.

 

Comment 4: Figure 1 shows the difference in Col1 fibers in recurrent and non-recurrent patients, did all non-recurrent patients have a similar fiber volume? Were there any cases of ambiguity? Were there any blind measurements done? Were there cross sections where Col1 fibers were not as dense in recurrent patients? Please write a section on potential challenges and inform readers of cases where no difference were noticed in recurrent and non-recurrent patient.

            Response: Yes the SHG microscopy studies were performed with blinded samples.  The mean fiber volume for each patient was obtained from 6-18 randomly selected FOVs per patient.  Each point in Figure 2 represents the mean fiber volume for each patient in the non-recurrent and recurrent group.  As evident in this figure, there is some overlap between the mean fiber volume detected in the recurrent and non-recurrent group.  We have also provided Supplementary Figure 1 with individual FOVs displayed for the two groups together with the corresponding statistical analysis.  Expanded studies with a large data base would be required to identify a Col1 fiber threshold value for risk for recurrence for stage 1 NSCLC.  These points have been discussed in the revised submission.  Please refer to page 4, lines 124-129, Supplementary Figure 1 and corresponding figure legend, page 10, lines 339-341, and revised Figure legend 2 in the revised manuscript.

 

Comment 5: It would be helpful to add a few tissue sample analysis from healthy/ cancer-free patients in order to validate your hypothesis. What was the control used?

            Response: Our purpose here was to evaluate differences between recurrent and non-recurrent stage 1 NSCLC using the H&E sections of the NIH NLST data set.  We have clarified this in the Discussion and have included a statement to say that future studies should include analysis from cancer-free patients.  Please see page 10, lines 341-343 in the revised manuscript.

 

Comment 6: Please include patient numbers in the study. Attach a patient number by the side of each image, so readers have a reference as to which patient you are referring to. For example, non recurrent patients can be identified as NR 1, NR 2… And recurrent patients can be identified as R1, R2.. Since the patient size is so small, add a supplementary section showing tissue sections from each patient with the patient ID.

Response: We have modified the figures accordingly.   We have provided the de-identified file names associated with each patient in Figures 1 and 3.  Please see revised Figures 1 and 3 and corresponding figure legends in the revised manuscript.

 

Comment 7: Figure 2, how is the percentage value obtained. Please provide a detailed explanation of that. Please provide justification on why a specific statistical test was used to determine significance. A small section on statistical methods would be helpful. Was the t-test parametric? Non-parametric? Define the distribution and identify the correct statistical test.

            Response: Briefly, our software quantifies the total Col1 fiber volume by first preprocessing the raw image to exclude noise and nonfibrillar shapes by using a shape filter as previously described [8] (reference #27 in the revised manuscript).  The Col1 fiber structures extracted from the raw images were analyzed as the percent Col1 fibers per FOV.  A parametric 1-tailed two-sample unpaired Student’s t-test was used for statistical analysis.  We have provided these details in the revised manuscript.  We have also included Supplementary Figure 1 where the percent fiber volume from randomly FOVs from 6-18 FOVs from each tissue block, one block per patient were analyzed. To compare the percent fiber volume between patients with and without recurrence, we employed a random effects model. In this model, the tissue block served as the random effect (random intercept), while the cancer recurrence status was the fixed effect. Additionally, we log-transformed the percent fiber volume to reduce skewness before fitting the model. The model output indicates that recurrent patients exhibited a higher percent fiber volume than non-recurrent patients (two-sided test p-value = 0.078, one-sided test p-value = 0.039).  Please see page 4, lines 146-450, and pages 4-5, lines 169-184, and Supplementary Figure 1 and corresponding figure legend in the revised manuscript.

 

Comment 8: Did you do statistical power analysis? Is your sample size enough to draw conclusions about Col1 fibers? Please write a paragraph about this.

            Response: We did not perform a power analysis.  This was an exploratory study to identify trends in the data for future expanded studies.

 

Comment 9: For the increased expression of Col1, fibronectin, laminin, nidogen-2, and aggrecan, it would be useful to provide more context about their roles in extracellular matrix (ECM) remodeling and how this correlates with cancer invasiveness.

            Response: Yes we agree.  We have provided the details in the revised manuscript.  Please see pages 9-10, lines 301-312, in the revised manuscript.

 

Comment 10: The decrease in hyaluronan binding protein-2 expression is noted, but its potential implications should be discussed. How might this decrease affect tumor progression or patient prognosis.

            Response: We have discussed the potential implications in the revised manuscript.  Please see page 10, lines 314-316 and lines 318-319, in the revised manuscript.

 

Comment 11: What are the different subtypes of CAFs? Were different subtypes of CAFS analyzed?

            Response: We used known molecular markers associated with CAFs to screen for differences between the invasive and noninvasive groups.  The molecular markers that showed a significant difference between the two groups were associated with antigen-presenting CAFs (apCAFs), inflammatory CAFs (iCAFs), myofibroblast CAFs (myCAFs), and FAP-a CAFs.  This is further clarified in the revised manuscript together with a brief description of their characteristics.  Please see page 10, lines 326-327, and lines 328-331, in the revised manuscript. 

 

Comment 12: Are there any other phenotypes of the COL1 fibers that can be studied to further interpret the recurrence of disease? Such as Size, brightness, length etc

            Response: The reviewer makes an excellent point.  Yes we can examine texture, morphology, length in the future.  We have included this point in the Discussion.  Please see page 10, lines 343-345, in the revised manuscript.

 

Comment 13: Please write a section on the mechanism and functional role of Col1 fibers in the progression/regression of the disease.

            Response: This has been provided in the revised manuscript.  Please see page 9, lines 301-307 in the revised manuscript.

 

Comment 14: Have there been cases of recurrence where these Col1 fibers were not elevated?

            Response: As mentioned earlier, the mean fiber volume for each patient was obtained from an average of 12 randomly selected FOVs per patient.  Each point in Figure 2 represents the mean fiber volume for each patient in the non-recurrent and recurrent group. As evident in this figure, there is some overlap between the mean fiber volume detected in the recurrent and non-recurrent group.  We have also provided Supplementary Figure 1 with individual FOVs displayed together with the corresponding statistical analysis.  Expanded studies with a large data base would be required to identify a Col1 fiber threshold value for risk for recurrence for stage 1 NSCLC.  These points have been discussed in the revised submission.  Please see page 4, lines 124-128, Figure 2 and corresponding figure legend, Supplementary Figure 1 and corresponding figure legend, and page 10, lines 341-434, in the revised manuscript.

 

Comment 15: Are Figure 3 and Figure 1 images from same patients? Please have image identifiers for all patients as described in one of the comments above.

            Response: We have provided the de-identified file names associated with each patient in Figures 1 and 3.  Please see revised Figures 1 and 3 and correspondingly revised figure legends.

 

Comment 16: Figure 3, I see some sections in non-recurrent patients with long thicker patterns of Col1 fibers. For example, Figure 3A, non recurrent patient, at the 4’o clock region a zoom in might show long fibers. Figure 3, needs to show a bigger data set and different cross sections of COL1 fibers. Showing one very evident cross section may not be sufficient to identify recurrence in patients.

            Response: The reviewer makes an excellent point.  This exact concerns made us display the entire section as shown in Figure 3, as opposed to the randomly selected FOVs presented in Figure 1, to illustrate the differences in Col1 fibers between recurrent and nonrecurrent stage 1 NSCLC.  As mentioned earlier, the mean fiber volume for each patient was obtained from 6-18 randomly selected FOVs per patient.  As mentioned earlier, each point in Figure 2 represents the mean fiber volume for each patient in the non-recurrent and recurrent group. 

 

Comment 17: Overall this is a great study as it provides valuable insights into the differential gene expression between noninvasive and invasive lung adenocarcinoma, particularly focusing on ECM-related proteins. This is highly relevant given the importance of understanding tumor microenvironment interactions in cancer progression and potential therapeutic interventions. However, with addition of control group and addressal of the above mentioned comments, the study can be considered for publication.

            Response: We thank the reviewer for these comments and have addressed the concerns to the best of our ability.

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

I thank the reviewers for addressing the issues previously raised. However, my concerns regarding this research persist. The RNA data has low significance as it cannot be correlated to the collagen findings. On the other hand, the authors do not explore collagen SHG data beyond fiber volume, which is quite a limited analysis. For this reason, it is my opinion that this manuscript is not robust enough for publication.

Comments on the Quality of English Language

Minor typos and grammatical errors

Author Response

We thank the Reviewer for their insightful comments.  We have revised the manuscript to address the concerns to the best of our ability to make the manuscript suitable for publication.

Reviewer 3 Report

Comments and Suggestions for Authors

I appreciate the author's response to my comments. The extensive revision of the manuscript makes it more impactful.

Author Response

We appreciate your valuable comments to help us improve the quality of our manuscript.