Use of Propensity Score Matching to Compare Short Outcomes from Transoral and External Surgical Approaches in Patients with Deep-Lobe Parotid Pleomorphic Adenomas

Round 1

Reviewer 1 Report

The abstract of the review report form is not updated

If Histologic subtype (cellular, myxoid, classic) and the presence of pseudopodia cannot not considered, it should be clearly stated in the limitations of the study

Not all the modification are highlighted in the new version of the manuscript.

Author Response

The abstract of the review report form is not updated

Thank you for the comment. We will update the abstract of the reiew report.

If Histologic subtype (cellular, myxoid, classic) and the presence of pseudopodia cannot not considered, it should be clearly stated in the limitations of the study

Thank you for the constructive suggestion. We have discussed the histologic subtype in the limitation paragraph in the discussion section as suggested.

Not all the modification are highlighted in the new version of the manuscript.

Thank you for this comment. We have highlighted the other modifications in the revised manuscript.

Reviewer 2 Report

I thank the authors for the effort they put together to address my previous comments, including consulting a biostatistician.

However, there are still few issues that need to be addressed before accepting the paper:

• In the statistical analysis, the authors need to add the tests used in the new table 3. As both ANOVA and chi-square test should have not been used in table 3. As a matter of fact, I don’t see anywhere in the results where one-way ANOVA can be approproiately used. Please clarify.
• Line 103: “respectively” should be omitted or if referring to the previous sentence, then align with continuous and categorical variables, Also please see the previous comment.
• Line 116-117: please consult with the biostatistician on the appropriate phrase to use. Patients who did not die are the ones who get censored on their last contact date if they are lost to follow up.
• Line 120: Can you please give a link or reference to the “Free statistic software versions 1.2”

Author Response

In the statistical analysis, the authors need to add the tests used in the new table 3. As both ANOVA and chi-square test should have not been used in table 3. As a matter of fact, I don’t see anywhere in the results where one-way ANOVA can be approproiately used. Please clarify.

Sorry for the mistake. We have rewritten the statistical analysis part in the revised manuscript.

Line 103: “respectively” should be omitted or if referring to the previous sentence, then align with continuous and categorical variables, Also please see the previous comment.

Thank you for this suggestion. We have made the revisions as you suggested.

Line 116-117: please consult with the biostatistician on the appropriate phrase to use. Patients who did not die are the ones who get censored on their last contact date if they are lost to follow up.

Thank you for this comment. We have rewritten the phrases about censoring time in the revised manuscript.

Line 120: Can you please give a link or reference to the “Free statistic software versions 1.2”

Thank you for this comment. The reference about the software is

Yang Q, Zheng J, Wen D, et al. Association between metformin use on admission and outcomes in intensive care unit patients with acute kidney injury and type 2 diabetes: A retrospective cohort study. J Crit Care 2021;62:206-11. doi: 10.1016/j.jcrc.2020.12.007 [published Online First: 2021/01/11]

Round 2

Reviewer 2 Report

I thank the authors again for the effort addressing my comments.

There are still few issues especially in the statistics part that needs to be answered.

1) In the methods statistical analysis, the authors removed the tests used for comparison but didn’t replace it with anything else. Please consult the biostatistician about this. This should be corrected before publishing the paper:

• Please verify what test you used for total hospital stay. Since there are two groups, if parametric you need to use and report in the methods the t-test to compare both, and for the non-parametric you use mann-whitney U test.
• Was the Chi-square test with Yates' continuity correction used to generate the P value? Please work with the biostatistician to check if Fisher exact test is more appropriate here for some comparisons.

2)  Again, for the comment about “Free statistic software versions 1.2” in the methods sections. What is the “Free statistic software versions 1.2”. The reference provided did not have anything related to this. Also the authors need to add the appropriate reference or link to the text, to allow all reader to know what software was used other than R.

Author Response

Response to Editor and Reviewers

Reviewer 2

I thank the authors again for the effort addressing my comments.

There are still few issues especially in the statistics part that needs to be answered.

1) In the methods statistical analysis, the authors removed the tests used for comparison but didn’t replace it with anything else. Please consult the biostatistician about this. This should be corrected before publishing the paper:

Please verify what test you used for total hospital stay. Since there are two groups, if parametric you need to use and report in the methods the t-test to compare both, and for the non-parametric you use mann-whitney U test.

Was the Chi-square test with Yates' continuity correction used to generate the P value? Please work with the biostatistician to check if Fisher exact test is more appropriate here for some comparisons.

Thank you for the comment. I have consulted a biostatistician as you suggested, and added the following sentence in the revised manuscript “Variables were compared using t-test for continuous variables and chi-square test for categorical variables”.

2)  Again, for the comment about “Free statistic software versions 1.2” in the methods sections. What is the “Free statistic software versions 1.2”. The reference provided did not have anything related to this. Also the authors need to add the appropriate reference or link to the text, to allow all reader to know what software was used other than R.

Thank you for this concern. Free statistic software is a menu software based on R and python, its methodology literature and web page are being created and is temporarily inaccessible. The analyses in this study were performed with the statistical software packages R (matching, survminer, and other R packages), and the results were organized with the help of Free statistic software. We decided to put the software description in the acknowledgements. The reference I provided before was a study also use the “Free statistic software” with the statistic analysis.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The Authors present a study on the efficacy of transoral approach to remove deep lobe parotid pleomorphic adenoma of the parapharyngeal space. I think that the manuscript is well written, but the follow-up time is too short, and it should be clearly stated in the discussion as a limitation of the study. I think also, that PSM should include integrity of the tumor during removal and some histologic aspects like the presence of a complete capsule or pseudocapsule, the histologic subtype (cellular, myxoid, classic), and the presence of pseudopodia.

Some questions:

Pure parapharyngeal pleomorphic adenoma arising from minor salivary glands are excluded from the analysis?

No other complications except facial paralysis?

What is the timing of oral re-alimentation?

How is the follow-up? MR every year?

Recently robotic transoral approach has been proposed as alternative to standard endoscopic transoral approach: The Authors should report and discuss this option.

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

I had a great pleasure to review the paper entitled: “Use of propensity score matching to compare outcomes from transoral and external surgical approaches in patients with deep-lobe parotid pleomorphic adenomas”

The authors retrospectively evaluated the outcome of patients with deep-lobe parotid gland pleomorphic adenomas, comparing transoral versus an external approach using propensity score matching. This is considered a big cohort of patients with deep-lobe parotid gland pleomorphic adenomas who presented to a single institution over the span of 18 years.

This is an interesting study with the following issues being identified:

• In the multivariable analysis, some parameters have three categories. What does the HR reflect in these? Which one is the reference? please clarify.
• No clarification of when the patients were censored. Please add this to the methods section and explain the reason for censoring the patients (eg. Death, lost to follow-up …. Etc.)
• In the methods section, it is stated that an unadjusted multivariable Cox proportional hazards model was used, but still, you are using other covariates in the model. Verify why this is unadjusted.
• Using both multivariable and multivariate wording is a bit confusing. Please be consistent throughout the paper and use multivariable since you are referring to only one outcome.
• It is worth summarizing the complications and duration of hospitalization in a table.
• A clear limitation section can’t be identified. Even though the authors covered few limitations in the discussion, a specific section for the limitations can be helpful. If added, consider adding the extended study period which may reflect different treatment modalities and techniques (in consistent with the transoral approach being started in 2010 in the authors institution).
• In the third section of the discussion, the authors discussed the transoral approach in terms of safety and surgeon’s acceptance. Then stated, “because of this controversy we implemented PSM …”. Since the matching factors were age, gender, and tumor size, this could be an overstatement and it would be helpful to rephrase.
• Few parts of the manuscript may need some improvement in finer points of grammar and punctuation.

Author Response

Please see attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

You present an impressively large series of pure deep-lobe PA of the parotid glands, considering that most often they are only a dumbbell extension from the superficial lobe. Your selection criteria are critical to support your results given that only these particular lesions are suitable to a pure transoral approach. I have some questions for you:

• how was the preoperative work up conducted? MRI or CT (it is not clear in the MM section). What about FNAC? How was it performed (US/ CT guided)?
• how was the follow up conducted? yearly MRI?
• please specify in detail the used approaches: eg. differences between  transcervical and transparotid?
• Although matching the two groups is a convenient technique, there is an inherent selection bias given the retrospective design: why was the group treated by non-transoral approaches? 
• MM: "an unadjusted multivariate Cox model"? please explain given how variables in the  multivariate model were included. Then, most importantly, was this model applied only on the matched cohort or on the whole group? Last, I don't understand the insertion of the variables "age" and "gender", how are they related to the outcomes? Often transoral approach is not feasible because of other covariates such as mouth opening/modified Mallampati Score etc.
• bibliography should be updated, starting from 10.1002/lary.29458

Then there are very little mispells and mistakes (eg., line 214 is cm) but overall the manuscript is well-written.

Author Response

Please see attachment.

Author Response File: Author Response.pdf