A Meta-Analysis of Randomized Clinical Trials Assessing the Efficacy of PARP Inhibitors in Metastatic Castration-Resistant Prostate Cancer

Round 1

Reviewer 1 Report

With pleasure, I read the paper titled “Meta-Analysis of Randomized Clinical Trials Assessing Efficacy of PARP Inhibitors in Metastatic Castration-Resistant Prostate Cancer”. The topic is clinically relevant to practice, and of importance to the readers of the Current Oncology journal. Overall, the manuscript reads well. However, some changes are required as indicated below:

Abstract:

·        Please improve the writing of the abstract and correct typos.

Introduction:

·        Please mention how many RCTs were included by Niazi et al. moreover, there is another meta-analysis recently published on the same topic by Iannantuono et al (Cancer Treat Rev. 2023 Sep 9;120:102623. doi: 10.1016/j.ctrv.2023.102623). This is very important to highlight the significance of your research and how it contributes to bridging the gap in the existing literature.

·        Please conclude the introduction section with some proposed hypotheses.

Methods:

·        In addition to PRISMA statement, please mention if the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions were followed during the preparation of this research.

·        Retrospective registration of the research protocol is highly advised, however, it is NOT mandatory. If registration was not done, it is perfectly fine, however, please acknowledge this clearly in the methods section and highlight it as a limitation in the limitations section.

·        Please mention if specific filters (such as year of research, country of publication, or English language) were used during literature screening.

·        Please indicate if gray literature was searched.

·        Please indicate if the references of the recent review and/or references of the meta-analyzed studies were also manually screened to identify additional potential studies that could have been mistakenly missed.

·        Section 2.2. belongs to the results section, and not the methods section.

·        There is a mismatch between the screened databases in Figure 1 (n=5 databases) and section 2.1 (n=3 databases). Please correct.

·        For inclusion criteria, please specify the “control” group.

·        For quality assessment, the authors used Cochrane risk of bias tool version 1. However, I recommend using the updated risk of bias tool (version 2) to be in line with the up-to-date Cochrane guidelines. Section 2.3 should be moved to results section.

·        For assessment of between-study heterogeneity, have you also considered the p-value of the Cochran’s Q test (i.e., p<0.1)?

·        For statistical significance, please correct the typo of p-value from <0.005 to <0.05.

·        Please mention in the methods section all the types of subgroup analyses that will be performed. 

Results:

·        Several sections in the methods section should be moved to the results section.

·        The section should start with a summary of the database screening and inclusion of studies. It also should provide a summary of the baseline characteristics and study quality.

·        The results should be reported in  better and standard way for DFS and OS. The way it is written, is NOT scientific and not up to the standards. The way authors wrote it in the abstract is pretty standard and correct. Perform the analysis according to only one method (either fixed- or random-effects model according to the heterogeneity of the data).

·        I have a concern about the selected studies as I believe some studies may have been missed. Please refer to this article and double-check: https://pubmed.ncbi.nlm.nih.gov/37716332/ 

Discussion

·        Please acknowledge additional limitations, such as: (a) lack of PROSPERO registration, hence analysis could be subjected to some bias, (b) almost all, if not all, outcomes had high heterogeneity which could be ascribed to differences in study durations, patient characteristics, and control groups, and (c) publication bias could not be assessed reliably as the number of included studies per outcome was less than the required cutoff (n=10).

Overall

·        The manuscript needs some polishing for English language and editing.

Moderate English editing is needed

Author Response

Reviewer 1

 Abstract:

• Please improve the writing of the abstract and correct typos.

Response: The typos have been removed and language has been improved.

Introduction:

• Please mention how many RCTs were included by Niazi et al. moreover, there is another meta-analysis recently published on the same topic by Iannantuono et al (Cancer Treat Rev. 2023 Sep 9;120:102623. doi: 10.1016/j.ctrv.2023.102623). This is very important to highlight the significance of your research and how it contributes to bridging the gap in the existing literature.
• Please conclude the introduction section with some proposed hypotheses.

Response:  Changes have been made to last paragraph of introduction to

• Conclude with a hypothesis
• meta-analysis recommended by reviewers has been added in the last paragraph of introduction.

Methods:

• In addition to PRISMA statement, please mention if the guidelines of the Cochrane Handbook for Systematic Reviews of Interventions were followed during the preparation of this research.

• Retrospective registration of the research protocol is highly advised, however, it is NOT mandatory. If registration was not done, it is perfectly fine, however, please acknowledge this clearly in the methods section and highlight it as a limitation in the limitations section.

Response: This has been mentioned.

• Please mention if specific filters (such as year of research, country of publication, or English language) were used during literature screening.
• Please indicate if gray literature was searched.
• Please indicate if the references of the recent review and/or references of the meta-analyzed studies were also manually screened to identify additional potential studies that could have been mistakenly missed.

Response: This has been mentioned under search strategy and literature search.

Further

• Section 2.2. belongs to the results section, and not the methods section.

Response: PRISMA flowsheet and Table of Study Characteristics has been moved to “Results” section.

• There is a mismatch between the screened databases in Figure 1 (n=5 databases) and section 2.1 (n=3 databases). Please correct.

Response: Changes have been made.

• For inclusion criteria, please specify the “control” group.

Response: Accepted.

• For quality assessment, the authors used Cochrane risk of bias tool version 1. However, I recommend using the updated risk of bias tool (version 2) to be in line with the up-to-date Cochrane guidelines. Section 2.3 should be moved to results section.

Response: Will move section 2.3

• For assessment of between-study heterogeneity, have you also considered the p-value of the Cochran’s Q test (i.e., p<0.1)?

Response: It was used and has been added in Methods section

• For statistical significance, please correct the typo of p-value from <0.005 to <0.05.
• Please mention in the methods section all the types of subgroup analyses that will be performed. 

Response: Accepted

Results:

• Several sections in the methods section should be moved to the results section.
• The section should start with a summary of the database screening and inclusion of studies. It also should provide a summary of the baseline characteristics and study quality.

Response: Accepted

• The results should be reported in  better and standard way for DFS and OS. The way it is written, is NOT scientific and not up to the standards. The way authors wrote it in the abstract is pretty standard and correct. Perform the analysis according to only one method (either fixed- or random-effects model according to the heterogeneity of the data).

Response: Accepted and changes have been made. We will restrict the analysis just to fixed form as per the I2 statistics.

 I have a concern about the selected studies as I believe some studies may have been missed. Please refer to this article and double-check: https://pubmed.ncbi.nlm.nih.gov/37716332/

Response: None of the Phase-III Randomized control trials from the meta-analysis recommended by editors, has been missed.

Discussion

• Please acknowledge additional limitations, such as: (a) lack of PROSPERO registration, hence analysis could be subjected to some bias, (b) almost all, if not all, outcomes had high heterogeneity which could be ascribed to differences in study durations, patient characteristics, and control groups, and (c) publication bias could not be assessed reliably as the number of included studies per outcome was less than the required cutoff (n=10).

Response:  Accepted.

Overall

• The manuscript needs some polishing for English language and editing.

Reviewer 2 Report

In this article the authors present the results of meta-analysis of randomized clinical trials that assess efficacy of PARP inhibitors in metastatic castration-resistant prostate cancer.

General comment:

The topic of the study is relevant. The limitations of the study are fairly presented.

Specific points:

Table 1: It would be good to add the number of patients that are included in each study; just like the authors did for study 'Clarke et al. (NCT01972217)'.

Lines 59-56 nicely introduce the rationale of using PARP inhibitors in patients with breast and ovarian cancers. However, the rationale and the potential mechanisms of PARP action in prostate cancer could be briefly introduced.

Section '2. Methodology' is followed by section '3. Progression-free survival'. However, it should be clearly stated if the section 3. belongs to the Methodology or Results section.

In the file with original images, Figure 1 is of an optimal resolution; however, in the body of the manuscript the resolution is not so good; additionally, it would be good if the arrows from Figure 1 are under the right angle (not skewed).

It is not clear why is the title of section 4 ('4. OVERALL SURVIVAL') written in capital letters.

Author Response

Reviewer 2

Table 1: It would be good to add the number of patients that are included in each study; just like the authors did for study 'Clarke et al. (NCT01972217)'.

Response: Accepted. It has been mentioned for all studies except MAGNITUDE trial and PROFOUND  for the sake of simplicity since study as the arms were complicated in these two studies and this will require a lot of space in the table.

Lines 59-56 nicely introduce the rationale of using PARP inhibitors in patients with breast and ovarian cancers. However, the rationale and the potential mechanisms of PARP action in prostate cancer could be briefly introduced. 

Response: Authors agree with valuable suggestion. However, the proposed mechanism of action of PARPi in mCRPC has been mentioned twice in discussion (5^th paragraph and second-last paragraph). Authors would try to avoid mentioning it again in introduction for the sake of redundancy.

Section '2. Methodology' is followed by section '3. Progression-free survival'. However, it should be clearly stated if the section 3. belongs to the Methodology or Results section.

Response: Authors will make changes to the manuscript.

In the file with original images, Figure 1 is of an optimal resolution; however, in the body of the manuscript the resolution is not so good; additionally, it would be good if the arrows from Figure 1 are under the right angle (not skewed).

Response: Authors will make changes to the manuscript.

It is not clear why is the title of section 4 ('4. OVERALL SURVIVAL') written in capital letters.

Response: Authors will make changes to the manuscript.

Reviewer 3 Report

The authors seek to characterize the progression-free survival  and overall survival (OS) in metastatic CRPC patients treated with PARP inhibitors.

Undoubtedly a great limitation of the studies is the great heterogeneity of the outcome.

The authors report that " PSA levels concur to determine the initial staging and prognosis of prostate cancer" and PSA increases with the progression to ADT.  I have checked the reference and one reports about the role of PSA screening and the other on the epidemiology. Psa is the reference test for risk prediction of advanced PCA , and this concept is  relevant for this paper. Thus I suggest to improve the information about the role odf PSA in the risk stratification  according to a recent pubblication Ferraro S,et al. Individual risk prediction of high grade prostate cancer based on the combination between total prostate-specific antigen (PSA) and free to total PSA ratio. Clin Chem Lab Med. 2023

Author Response

Reviewer 3

The authors report that " PSA levels concur to determine the initial staging and prognosis of prostate cancer" and PSA increases with the progression to ADT.  I have checked the reference and one reports about the role of PSA screening and the other on the epidemiology. Psa is the reference test for risk prediction of advanced PCA , and this concept is  relevant for this paper. Thus I suggest to improve the information about the role odf PSA in the risk stratification  according to a recent pubblication Ferraro S,et al. Individual risk prediction of high grade prostate cancer based on the combination between total prostate-specific antigen (PSA) and free to total PSA ratio. Clin Chem Lab Med. 2023

Response: We will add the concept of free to total PSA ratio. In the introduction without going into the detail of it’s explanation. We will add the study by Ferraro et al. as well.

Round 2

Reviewer 1 Report

With pleasure, I read the revised paper titled “Meta-Analysis of Randomized Clinical Trials Assessing Efficacy of PARP Inhibitors in Metastatic Castration-Resistant Prostate Cancer”. The topic is clinically relevant to practice, and of importance to the readers of the Current Oncology journal. Unfortunately, the manuscript is not suitable for publication in its current stage, and the authors did not adequately address the major issues. Accordingly, I still recommend MAJOR revision prior to considering acceptance of the manuscript. Please note my comments below:

Abstract:

·         Please improve the writing of the abstract and correct typos. You need to define ADT and HRR terms. Once an abbreviation is spelled out, you need to just use the abbreviation. Be consistent with terminology; sometimes PARP inhibitors and sometimes PARPi. Regarding “When calculated using the random and fixed models, results favored PARP inhibitors, hazard 23 ratio (HR) 0.855(0.752-0.974), p=0.018 I2=0%”, you need to let readers understand that (0.752-0.974) represents the 95% confidence interval (CI) and I2 represents the heterogeneity statistic, and so on. Please be consistent in reporting the findings in your abstract. Why you mention both fixed- and random-effects models were used; you just need to use one only. There is a mismatch between the screened databases in Figure section and abstract section. Pease correct. Looking at the abstract itself, it gives bad impression about your entire article, although your article is really important clinically.

Introduction:

·         Please conclude the introduction section with some proposed hypotheses.

Methods:

·         There is a mismatch between the screened databases in Figure section and abstract section. Pease correct.

·         For statistical significance, please determine the p-value that establishes statistical significance.

·         Please mention in the methods section all the types of subgroup analyses that will be performed.

·         You should use only one method of analysis; fixed- or random-effects model, but not both, and this will be determined based on the I2 value; Please correct. This is not right: “Analysis was done using random and fixed models and both were reported”.

·         You have used the p-value of the Cochrane Q statistic to determine heterogeneity in your results section, and therefore you should acknowledge that in your methods section too. 

Results:

·         Section 5.1.1 should be moved to methods section.

·         The results section should start with a summary of the database screening and inclusion of studies. It also should provide a summary of the baseline characteristics and study quality. It should be followed by summary of the quality of studies.

·         The results should be reported in  better and standard way for DFS and OS. The way it is written, is NOT scientific and not up to the standards. Please see other published papers. Perform the analysis according to only one method (either fixed- or random-effects model according to the heterogeneity of the data). Always provide the I2 statistic value. For example, follow this style for PFS: five studies evaluated the PFS endpoints (intervention n=xxx, control n=xxx). The pooled results found that the intervention group had significantly better PFS (HR=0.855, 95% CI [0.752, 0.974], p=0.018). The pooled analysis was homogeneous/heterogeneous (I2= xxx% and Q-test for heterogeneity p-value=xxx) and the random/fixed effect model was used.

·         Figure 1 image has some typos for “superscript” and some numbers have underlined blue color. Please correct.

Discussion

·         Please acknowledge additional limitations, such as: publication bias could not be assessed reliably as the number of included studies per outcome was less than the required cutoff (n=10).

Overall

·         The manuscript needs extensive polishing for English language and editing.

Moderate English editing is needed

Author Response

Abstract:

• Please improve the writing of the abstract and correct typos. You need to define ADT and HRR terms. Once an abbreviation is spelled out, you need to just use the abbreviation. Be consistent with terminology; sometimes PARP inhibitors and sometimes PARPi. Regarding “When calculated using the random and fixed models, results favored PARP inhibitors, hazard 23 ratio (HR) 0.855(0.752-0.974), p=0.018 I2=0%”, you need to let readers understand that (0.752-0.974) represents the 95% confidence interval (CI) and I2 represents the heterogeneity statistic, and so on. Please be consistent in reporting the findings in your abstract. Why you mention both fixed- and random-effects models were used; you just need to use one only. There is a mismatch between the screened databases in Figure section and abstract section. Pease correct. Looking at the abstract itself, it gives bad impression about your entire article, although your article is really important clinically.

 Response: Thank you for valuable suggestions. The data reporting has been improved. We have removed the I² from the abstract. Rest of the data has been standardized. Abbreviations have been explained.

Introduction:

• Please conclude the introduction section with some proposed hypotheses.

Response: This has been added.

Methods:

• There is a mismatch between the screened databases in Figure section and abstract section. Pease correct.

Response: corrected.

 For statistical significance, please determine the p-value that establishes statistical significance.

Response: This has been mentioned under statistical analysis in Line 142.

• Please mention in the methods section all the types of subgroup analyses that will be performed.

Response: A paragraph under “Study Objectives” has been added.

• You should use only one method of analysis; fixed- or random-effects model, but not both, and this will be determined based on the I2 value; Please correct. This is not right: “Analysis was done using random and fixed models and both were reported”.

Accepted

• You have used the p-value of the Cochrane Q statistic to determine heterogeneity in your results section, and therefore you should acknowledge that in your methods section too. 

Response: We have mentioned it under heading “Q-test” in methods section. Is that the one you are talking about?

 Results:

• Section 5.1.1 should be moved to methods section.
• The results section should start with a summary of the database screening and inclusion of studies. It also should provide a summary of the baseline characteristics and study quality. It should be followed by summary of the quality of studies.

 Response: Accepted

 The results should be reported in  better and standard way for DFS and OS. The way it is written, is NOT scientific and not up to the standards. Please see other published papers. Perform the analysis according to only one method (either fixed- or random-effects model according to the heterogeneity of the data). Always provide the I2 statistic value. For example, follow this style for PFS: five studies evaluated the PFS endpoints (intervention n=xxx, control n=xxx). The pooled results found that the intervention group had significantly better PFS (HR=0.855, 95% CI [0.752, 0.974], p=0.018). The pooled analysis was homogeneous/heterogeneous (I2= xxx% and Q-test for heterogeneity p-value=xxx) and the random/fixed effect model was used.

Response: Accepted.

• Figure 1 image has some typos for “superscript” and some numbers have underlined blue color. Please correct.

Response: Corrected

Discussion

• Please acknowledge additional limitations, such as: publication bias could not be assessed reliably as the number of included studies per outcome was less than the required cutoff (n=10).

 Response: Added

Reviewer 3 Report

no further comments

Author Response

Thank you so much

Round 3

Reviewer 1 Report

For the most parts, the authors have addressed the comments adequately. The manuscript now reads well, and is much robust methodologically and valid scientifically. Nonetheless, moderate English editing is required by the journal's editorial team. Apart from that, the manuscript can be accepted in its current form.

Moderate English editing is STRONGLY required before the final version of the article is published.