Nomogram Based on Liver Function Test Indicators for Survival Prediction in Nasopharyngeal Carcinoma Patients Receiving PD-1 Inhibitor Therapy
Round 1
Reviewer 1 Report
hello
dear Authors thank you for an intersting paper
please structurise the Abstract as suitable for original papers
what was the inclusion criteria- only th NPC carcnioma? all types and usbtypes of NPc?
DOES each case of NPC regardless its clinical C/T- TNM was involved in the study?
what if patients had N+ neck, with nodal metastasis, was RTH/CTH-RTH also considered as treatment of choice? what about surgery in N+ nodes?
I would consider a chart-flow diagram to clearly describe the main NPC cases study group
point 2,.1, end of the chapter, please explain the last paragraph
please add a chapter, what authors do and how classify patients with cT1-T4, and N0-N1 with and without nodal involvement - is it all realted with AJCC guidelines?
figure 2-5, a very great diagram, congratulations
what if a patient's liver with NPC is not present, removed in the past or not working properly?
should patients before starting NPC treatment should undergo a strict/detailed examination of liver and co-related markers?
are patients who abuse alcohol and have liver problems be a limitation to this study?
how can we asses liver metastasis - pet. SPECT? others? mayby markers?
should patients with hepatitis b/c or other viral factors included or exluded from the study or presented the scope of treatment in NPC?
interesting paper, congratulations
please complete and add mentioned data in text, and paper will be a great addition to the literature
with regards
Author Response
Dear reviewer:
On behalf of all the contributing authors , I would like to express our sincere appreciations of your letter and reviewers ' constructive comments conceming our article entitled "Nomogram based on liver function tests indicators for survival prediction of Nasopharyngeal carcinoma patients combined with PD-1 inhibitors therapy” ( Manuscript No : 22422905). These comments were all valuable and helpful for improving our article . According to the editor and reviewers ' comments , we have made extensive modifications to our manuscript and supplemented extra data to make our results convincing . In this revised version , changes to our manuscript were all highlighted within the document by using red - colored text . Point - by - point responses to the nice associate editor and two nice reviewers are listed below this letter .
1.Please structurise the Abstract as suitable for original papers.
Reply:The abstract has been structured and words such as purpose, method, result and conclusion have been added to make it suitable for the original paper.
2.What was the inclusion criteria- only th NPC carcnioma? all types and usbtypes of NPc? DOES each case of NPC regardless its clinical C/T- TNM was involved in the study?what if patients had N+ neck, with nodal metastasis, was RTH/CTH-RTH also considered as treatment of choice? what about surgery in N+ nodes?I would consider a chart-flow diagram to clearly describe the main NPC cases study group.
Reply:The criteria for inclusion are nasopharyngeal carcinoma patients who were willing to accept PD-1 immunotherapy. Because PD-1 immunotherapy is still in the clinical testing stage, and the treatment price is relatively expensive, not all patients are willing to accept this treatment mode, so only 162 patients can meet the requirements of this study.I am sorry it is not possible to chart the NPC Cases Study Group.
3.Point 2,.1, end of the chapter, please explain the last paragraph
Reply:The last paragraph of the original 2.1 was incorrectly placed in the clinical data because it was a biochemical index with truncation value calculated by X-tile software. This paragraph has been changed in the 3 results section.
- Please add a chapter, what authors do and how classify patients with cT1-T4, and N0-N1 with and without nodal involvement - is it all realted with AJCC guidelines?
Reply:Yes, this is all related to the AJCC guidelines of NPC 8thedition, and the article has added the classification of patients with lymphatic involvement in the last paragraph of 2.1.
5.What if a patient's liver with NPC is not present, removed in the past or not working properly?
Reply:The clinical baseline data of this study were examined, and none of the 162 patients had a liver that was absent, had been removed in the past, or did not function properly, so the inclusion criteria were increased in the revised manuscript:(4)The patients had no liver resection or transplantation.
6.Should patients before starting NPC treatment should undergo a strict/detailed examination of liver and co-related markers?
Reply:Yes, our study was conducted with CT, MRI (plain scan and enhanced scan), PET-CT, liver function indicators and other markers before starting NPC treatment.
- Are patients who abuse alcohol and have liver problems be a limitation to this study?
Reply:We supplemented the data on drinking history and performed a univariate Cox analysis, P>0.05, which was not statistically significant, representing no limitations to the study.There was additions in the revised manuscript with Table 1 and Figure 1.
8.How can we asses liver metastasis - pet. SPECT? others? mayby markers?
Reply:CT, MRI (plain scan and enhanced scan) and PET-CT, liver function indicators, and markers were used to evaluate liver metastasis.
9.Should patients with hepatitis b/c or other viral factors included or exluded from the study or presented the scope of treatment in NPC?
Reply:Only 48 of the 162 patients were tested for hepatitis A, C, and E, all of which were negative.Hepatitis B surface antigen (HBSAG) was detected in 93 patients and univariate Cox analysis was performed (P>0.05), with no statistical significance.There are additions in the revised manuscript.Thank you.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript by Liang et al describes a study of nasopharyngeal cancer (NPC) patient survival from which a nomogram was created.
The study appears very robust with both training and validation cohorts as well as detailed clinical data.
Unfortunately, the study becomes very narrow focused (PDL1 treated NPC with liver metastases) with some problems in the main results (and possibly study design) since it is likely that authors included primary and recurrent tumors in the same survival assessment, disregarded non-liver metastases, and used automatically selected cutoffs without making sure the resulting groups are of meaningful and comparable size. For example the authors stress the influence of CHE parameter, yet the cutoff used has only 12 &5 cases <5941 cutoff contrasted with 105&40 cases possibly enhancing the effects of outliers and chance. Possible correlation of liver metastases and LFT measurements can also further confound the results and complicate the interpretation. Some of the data is unclear and possibly misclassified (102 cases of lymph node metastasis listed as distant metastasis, only 27 cases without metastases at all). There is also no overall survival shown or TNM stage survival curves. There is no follow up time reported. This makes it challenging to put the results in context of what is expected.
The above (together with strange very narrow focus) is in stark contrast with apparently strong statistical background needed to run the statistics used in making the results (uni & multi Cox, KM, nomogram) which unfortunately makes the results appear less robust than what would be expected.
Some of the issues are listed below (unfortunately there is no line numbering to refer to)
P1 title, the abbreviation NPC should be avoided in the title. The same abbreviation is not really explained in the abstract either.
P2 the randomization is not really clear? Was randomization done on actual patients prospectively when administering PD-1 treatment (and what were the differences in subsequent processing of patients?) From the wording it is difficult to say whether randomization was done first and then subsequently some cases were excluded for not fulfilling inclusion criteria or not?
Was there some bias in patients meeting inclusion criteria? It is strange to see 117 vs 45 cases if randomization was done at treatment (one would expect even sizes of both groups). How many cases were removed for which reason? Hidden bias could hamper interpretation.
P2 why was the time for PFS calculated from the time of randomization? How does this relate to the date of diagnosis or date of treatment initiation? This also might hide some important details about the patient cohorts
The introduction suggests that PD-1 inhibitors are used for first / second line therapy for recurrent/metastatic NPC but also for primary treatment of advanced NPC. It might be important to acknowledge the status of patients at the time of randomization/inclusion in the study at this point. I guess that survival of primary or recurrent tumors would be significantly different.
Some “positive control” measure of the population should be provided to strengthen the credibility of the data shown. Ie some KM graphs of usual parameters like TNM stage or similar (at least as a supplement) would validate that the population studied is otherwise “behaving as expected”. What is the “expected” 1 or 2 OS and/or PFS of NPC patients at the institution or in the country? For example those not included for PDL1 treatment?
P2 “or dropping dead” is inappropriate phrase to use in the manuscript
P2 “LFTs were performed in the training cohort and validation cohort, separately” why is this highlighted? Does this mean that samples were collected at some point and then batches were assessed at some point in the future for all LFTs studied? What other differences in patient management were there between these what should be only administrative randomization groups?
P2 “TNM stage, distal metastasis” it might be worthwhile to clarify that distal metastasis means during followup if so? Was locoregional recurrence also recorded during followup? Apparently PFS was defined by “ patient's tumor progressing” which would/should also include recurrence and not only distant metastases?
P3 “regards to distant metastasis …there were 102 cases of lymph node metastasis”. The distinction between regional and distant lymph nodes should be highlighted. 8th edition AJCC manual suggests that “Involvement of lymph nodes below the clavicle (including mediastinum, infraclavicular region, axilla, or groin) is considered as distant metastases.” Do all 102 cases fulfill this criteria? Is this something usual for NPC?
P3 Table 1 suggests that all percentages are rounded to 1 decimal place. Spacing issues before brackets. “no .” should be replaced by N as in “N (%)”. Column heading shouldn’t begin with 162 patients, suggest renaimin to “Total cohort” or something similar. Format the variable names to be more distinct from factor levels, ie. align them to the left of the row (Gender, Age, BMI) to make the table more readable. Were there T0 cases? Usually N is grouped as N0 vs N1+? ECOG grouping is meainingless as given 0-1 vs 2 since there are only 2 cases with “=2”. “Radiotherapy& Chemo-therapy” consider replacing with chemoradiotherapy as one word.
P3 Table 1 could/should be updated with information about the median follow up time as well as final status at last contact (alive, dead, alive with progression of disease) so that readers can assess the actual number of events in later analyses
P3 some supplementary table should be made to summarize the patient cohort data for measured values (at least using the cutoffs provided). This might suggest that some cutoffs are inadequate for subsequent statistics, ie. ALT has only 2 patients at risk in the validation cohort for the <9.4 level making KM comparisons uninformative.
P4 Why are only liver metastases used in Cox regression? (there were more cases of lung and bone metastases so why exclude them from consideration? Is it accurate that there were only 27 cases without any metastases (“None”) during follow-up (or is it onset?)? While the introduction does suggest some additional importance of liver metastases it doesn’t really justify disregarding other metastases completely when making COX models.
P4 are liver metastases correlated with LFT measurements? If so, using them together in the regression analysis could be challenging and there are conflicting opinions on such cases.
P5 Why was initial M stage not considered? Having a metastasis at diagnosis ( 85 cases in total) could significantly affect the cox regression (while including ECOG makes no sense)
P5 while the Figure 1 is a very nice way to present the results please double check whether the 95% CI numbers for the multivariate analysis agree with the confidence interval on the intervening graph. For example the CHE CI should stretch to 5.009 but in the the figure the range seems to go closer to 6.0 than it should.
P6 consired relabeling the figures to replace raw variable names. Ie “Liver_m=0” could be replaced as “No” since the legend already specifies the Liver metastasis data is shown.
Typos
Each heading has “|” character in numbering ie. “2.1|. Study population”
P2 “(8thedition).”
P3 inconsistencies in spacing before brackets ie “…total bilirubin (TBIL), indirect bilirubin(IBIL)..”
P3 no space between parameters listes ie. “ALT(9.4U/L),AST(15.4U/L)”
Author Response
Dear reviewer:
On behalf of all the contributing authors , I would like to express our sincere appreciations of your letter and reviewers ' constructive comments conceming our article entitled "Nomogram based on liver function tests indicators for survival prediction of Nasopharyngeal carcinoma patients combined with PD-1 inhibitors therapy” ( Manuscript No : 22422905). These comments were all valuable and helpful for improving our article . According to the editor and reviewers ' comments , we have made extensive modifications to our manuscript and supplemented extra data to make our results convincing . In this revised version , changes to our manuscript were all highlighted within the document by using red - colored text . Point - by - point responses to the nice associate editor and two nice reviewers are listed below this letter .
- Unfortunately, the study becomes very narrow focused (PDL1 treated NPC with liver metastases) with some problems in the main results (and possibly study design) since it is likely that authors included primary and recurrent tumors in the same survival assessment, disregarded non-liver metastases, and used automatically selected cutoffs without making sure the resulting groups are of meaningful and comparable size. For example the authors stress the influence of CHE parameter, yet the cutoff used has only 12 &5 cases <5941 cutoff contrasted with 105&40 cases possibly enhancing the effects of outliers and chance. Possible correlation of liver metastases and LFT measurements can also further confound the results and complicate the interpretation. Some of the data is unclear and possibly misclassified (102 cases of lymph node metastasis listed as distant metastasis, only 27 cases without metastases at all). There is also no overall survival shown or TNM stage survival curves. There is no follow up time reported. This makes it challenging to put the results in context of what is expected.The above (together with strange very narrow focus) is in stark contrast with apparently strong statistical background needed to run the statistics used in making the results (uni & multi Cox, KM, nomogram) which unfortunately makes the results appear less robust than what would be expected.
Reply: I apologize for the misrepresentation in the abstract of the original article, but the full article was indeed a survival prediction for all NPC patients with remote metastases, not just NPC patients with liver metastases.Thank you for your comments.I have made some changes in the article.Cutoff values are numerical data used using X-tile software.The criteria for inclusion are nasopharyngeal carcinoma patients who were willing to accept PD-1 immunotherapy. Because PD-1 immunotherapy is still in the clinical testing stage and the treatment price is relatively expensive, not all patients are willing to accept this treatment mode. Due to the number of cases and the existing region, this paper has limitations, which were described in the penultimate paragraph of Discuss.Thank you for finding the problem of lymphatic metastasis in the review. After follow-up data verification, 63 cases of distant lymph node metastasis were found. The data were modified in Article 3.1|Patientcharacteristics and Table 1.The survival curve of TNM stage was supplemented. In multivariate COX analysis, P was >0.05, which had no statistical significance.
2.P1 title, the abbreviation NPC should be avoided in the title. The same abbreviation is not really explained in the abstract either.
Reply:The abbreviations of the title and abstract have been changed
3.P2 the randomization is not really clear? Was randomization done on actual patients prospectively when administering PD-1 treatment (and what were the differences in subsequent processing of patients?) From the wording it is difficult to say whether randomization was done first and then subsequently some cases were excluded for not fulfilling inclusion criteria or not?
Was there some bias in patients meeting inclusion criteria? It is strange to see 117 vs 45 cases if randomization was done at treatment (one would expect even sizes of both groups). How many cases were removed for which reason? Hidden bias could hamper interpretation.
Reply:The criteria for inclusion in this study were nasopharyngeal carcinoma patients who were willing to receive PD-1 immunotherapy. Because PD-1 immunotherapy is still in the clinical testing stage, and the treatment price is relatively expensive, not all patients are willing to accept this treatment mode, so only 162 patients can meet the requirements of this study, and only 162 patients could be randomized analysis.For traditional data sets, the ratio of the training set to the test set is usually 6:4,7:3 or 8:2, and rarely 1:1.
- P2 why was the time for PFS calculated from the time of randomization? How does this relate to the date of diagnosis or date of treatment initiation? This also might hide some important details about the patient cohorts
Reply:The time of PFS was not calculated by randomization time.The time was from the first receipt of PD1 immunotherapy, and all biochemical indicators, including liver function indicators, renal function indicators, and cardiac enzyme indicators, were monitored before treatment and three weeks after treatment until the last follow-up.
- The introduction suggests that PD-1 inhibitors are used for first / second line therapy for recurrent/metastatic NPC but also for primary treatment of advanced NPC. It might be important to acknowledge the status of patients at the time of randomization/inclusion in the study at this point. I guess that survival of primary or recurrent tumors would be significantly different.
Some “positive control” measure of the population should be provided to strengthen the credibility of the data shown. Ie some KM graphs of usual parameters like TNM stage or similar (at least as a supplement) would validate that the population studied is otherwise “behaving as expected”. What is the “expected” 1 or 2 OS and/or PFS of NPC patients at the institution or in the country? For example those not included for PDL1 treatment?
P2 “or dropping dead” is inappropriate phrase to use in the manuscript
Reply:Table 1 supplemented the survival status of 8-12 weeks of treatment and the survival status data of PFS, supplemented the KM graphs of TNM.Due to the limitations of receiving PD1 therapy, this column chart is only applicable to patients receiving PD1 therapy.P2 "ordroppingdead" is changed to "ordead".
6.P2 “LFTs were performed in the training cohort and validation cohort, separately” why is this highlighted? Does this mean that samples were collected at some point and then batches were assessed at some point in the future for all LFTs studied? What other differences in patient management were there between these what should be only administrative randomization groups?
Reply:I am sorry that P2 "LFTs is carried out in the training group and the verification group respectively" was the wrong expression and has been deleted.In terms of patient management, this is again related to the limitations of the patient inclusion criteria in this study.
7.P2 “TNM stage, distal metastasis” it might be worthwhile to clarify that distal metastasis means during followup if so? Was locoregional recurrence also recorded during followup? Apparently PFS was defined by “ patient's tumor progressing” which would/should also include recurrence and not only distant metastases?
Reply:Table 1 has been modified and supplemented with the location of remote metastases to include data on recurrence and non-recurrence.
8.P3 “regards to distant metastasis …there were 102 cases of lymph node metastasis”. The distinction between regional and distant lymph nodes should be highlighted. 8th edition AJCC manual suggests that “Involvement of lymph nodes below the clavicle (including mediastinum, infraclavicular region, axilla, or groin) is considered as distant metastases.” Do all 102 cases fulfill this criteria? Is this something usual for NPC?
Reply:Thank you for finding the problem of lymphatic metastasis. After our subsequent verification of the data, it was found that 63 cases belonged to distant lymph node metastasis instead of 102 cases. The data were modified in Article 3.1|Patientcharacteristics and Table 1.
9.P3 Table 1 suggests that all percentages are rounded to 1 decimal place. Spacing issues before brackets. “no .” should be replaced by N as in “N (%)”. Column heading shouldn’t begin with 162 patients, suggest renaimin to “Total cohort” or something similar. Format the variable names to be more distinct from factor levels, ie. align them to the left of the row (Gender, Age, BMI) to make the table more readable. Were there T0 cases? Usually N is grouped as N0 vs N1+? ECOG grouping is meainingless as given 0-1 vs 2 since there are only 2 cases with “=2”. “Radiotherapy& Chemo-therapy” consider replacing with chemoradiotherapy as one word.
Reply:Table 1 has all been changed as suggested above.
10.P3 Table 1 could/should be updated with information about the median follow up time as well as final status at last contact (alive, dead, alive with progression of disease) so that readers can assess the actual number of events in later analyses
Reply:Table 1 We added fustat for 8-12 weeks and final state data at last contact.
11.P3 some supplementary table should be made to summarize the patient cohort data for measured values (at least using the cutoffs provided). This might suggest that some cutoffs are inadequate for subsequent statistics, ie. ALT has only 2 patients at risk in the validation cohort for the <9.4 level making KM comparisons uninformative.
Reply:Measurements of the patient cohort data (at least using the cut off values provided) were shown in Figure 1.Please forgive us for the limitation that only 162 patients received PD1 treatment, so the data in the verification group could not be changed. We tried hardly to make changes, but all the results in the whole paper were changed, losing the purpose of the original study.
12.P4 Why are only liver metastases used in Cox regression? (there were more cases of lung and bone metastases so why exclude them from consideration? Is it accurate that there were only 27 cases without any metastases (“None”) during follow-up (or is it onset?)? While the introduction does suggest some additional importance of liver metastases it doesn’t really justify disregarding other metastases completely when making COX models.
P4 are liver metastases correlated with LFT measurements? If so, using them together in the regression analysis could be challenging and there are conflicting opinions on such cases.
P5 Why was initial M stage not considered? Having a metastasis at diagnosis ( 85 cases in total) could significantly affect the cox regression (while including ECOG makes no sense)
Reply:In our study, the data of distant metastasis and M stage of all sites were counted, but they were not included in the original table because they were not statistically significant (P>0.05). Now the data of distant metastasis cases and M stage have been added to Figure 1.For all clinical and biochemical indicators, we first performed a single factor analysis, and when the factor P<0.05, we were included in the multivariate COX analysis.Multivariate COX analysis found that liver metastasis, ALT, CHE, TBA and TP were independent risk factors for survival and prognosis prior to first PD-1 inhibitors therapy.
13.P5 while the Figure 1 is a very nice way to present the results please double check whether the 95% CI numbers for the multivariate analysis agree with the confidence interval on the intervening graph. For example the CHE CI should stretch to 5.009 but in the the figure the range seems to go closer to 6.0 than it should.
Reply:With the addition of smoking and drinking history, M-stage and distant metastasis data, the 95%CI of multivariate analysis was consistent with the confidence interval on the intervention chart(Figure 1)
14.P6 consired relabeling the figures to replace raw variable names. Ie “Liver_m=0” could be replaced as “No” since the legend already specifies the Liver metastasis data is shown.
Reply:I am sorry that the drawing of R word software was displayed in this way, and the original variable name cannot be replaced. After trying to modify the instruction, an error is reported and the drawing cannot be made.
15.Typos
Each heading has “|” character in numbering ie. “2.1|. Study population”
P2 “(8thedition).”
P3 inconsistencies in spacing before brackets ie “…total bilirubin (TBIL), indirect bilirubin(IBIL)..”
P3 no space between parameters listes ie. “ALT(9.4U/L),AST(15.4U/L)”
Reply:All the above had been modified, thank you.
Author Response File: Author Response.docx
Reviewer 3 Report
As a surgeon I am not familiar neither with some of the statistics they utilize nor with the relevant biological responses to cancer immunotherapy. However, their reasoning (on why to base a nomogram on liver function tests) seems logical to me as well as their methodology
Nevertheless I have several “concerns” :
It is confusing for me from the very beginning. They report in the abstract that “
TNM classification is not the only risk factor for NPC with liver metastasis”, as well as in their conclusion that “Our nomogram can assist doctors in predicting the prognosis and survival of patients with hepatic metastasis NPC “ However in their analysis they include all types of NPC not only those with liver metastasis.
They give me the impression that their conclusion is that liver function tests rather than TNM stage is more important factor of survival. Further explanations are required for this unexpected conclusion
This is an important group of NPC patients however this is probably “endemic” NPC and this nomogram may require validation to other regions. Even EBV status is not shown (at least as a percentage in their known cases in order give an idea of comparisons )
Finally, it is not clearly explained how a nomogram like this can be useful e.g on earlier initiation or avoidance of a PD-1 inhibitors?
And some minor comments :
depending on how much familiar are the readers of current oncology with the various abbreviations (NPC, PFS) these should be explained or not in the abstract
several grammatical corrections are required
several grammatical corrections are required
Author Response
Dear reviewer:
On behalf of all the contributing authors , I would like to express our sincere appreciations of your letter and reviewers ' constructive comments conceming our article entitled "Nomogram based on liver function tests indicators for survival prediction of Nasopharyngeal carcinoma patients combined with PD-1 inhibitors therapy” ( Manuscript No : 22422905). These comments were all valuable and helpful for improving our article . According to the editor and reviewers ' comments , we have made extensive modifications to our manuscript and supplemented extra data to make our results convincing . In this revised version , changes to our manuscript were all highlighted within the document by using red - colored text . Point - by - point responses to the nice associate editor and two nice reviewers are listed below this letter .
1.It is confusing for me from the very beginning. They report in the abstract that “ TNM classification is not the only risk factor for NPC with liver metastasis”, as well as in their conclusion that “Our nomogram can assist doctors in predicting the prognosis and survival of patients with hepatic metastasis NPC “ However in their analysis they include all types of NPC not only those with liver metastasis.
Reply:I am sorry for the misrepresentation of the original article, but the full text was indeed a survival prediction for all NPC patients with remote metastasis, not just NPC patients with liver metastasis.Thank you for your comments.I have made some changes in the article.
2.They give me the impression that their conclusion is that liver function tests rather than TNM stage is more important factor of survival. Further explanations are required for this unexpected conclusion
Reply:In this study, when liver function indexes and TNM stage were put into multivariate COX analysis, the TNM stage P values were all >0.05, which had no statistical significance compared with independent factors in Normagram.The survival curve and Normagram independent factor of TNM stage were added in the original text, and DCA curve and NRI value were better than TNM stage.
3.This is an important group of NPC patients however this is probably “endemic” NPC and this nomogram may require validation to other regions. Even EBV status is not shown (at least as a percentage in their known cases in order give an idea of comparisons )
Reply:Guangdong Province is a region with a high incidence of nasopharyngeal cancer in China, although there were "endemic" nasopharyngeal cancer, the number of cases that can implement PD-1 drug treatment were still limited, and PD-1 immunotherapy is still in the clinical testing stage, and the treatment price is relatively expensive, not all patients are willing to accept this treatment.In the follow-up data supplement, only 12 people were tested for EB-DNA, and only 4 were positive, with a positive rate of 33.3%. The relevant data has been added in the revised paper.Moreover, limitations were also raised at the end of the revised article.
4.Finally, it is not clearly explained how a nomogram like this can be useful e.g on earlier initiation or avoidance of a PD-1 inhibitors?
Reply:In the middle part of 3.4|Constructingnomogram, the use of a nomogram was introduced, which includes the liver function indicators before PD-1 inhibitor treatment. Only by calculating the high and low risk for survival prediction can PD-1 inhibitor use be avoided by determining the predicted risk values.The timing of early initiation cannot be predicted, only the timing of the first treatment with PD-1.
5.depending on how much familiar are the readers of current oncology with the various abbreviations (NPC, PFS) these should be explained or not in the abstractseveral grammatical corrections are required
Reply:The NPC and PFS in the title and abstract have been explained before using abbreviations in the revised paper.Thank you.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
thank you, looks perfect
congratulations
Author Response
Dear reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers' constructive comments concerning our article entitled “Nomogram based on liver function tests indicators for survival prediction of Nasopharyngeal carcinoma patients combined with PD-1 inhibitors therapy” (Manuscript No: 22422905). These comments were all valuable and helpful for improving our manuscript.
Thanks for the reviewer's affirmation of this paper, but one reviewer was not satisfied with the original revised paper, so I tried to make a major revision of my paper again, and this change was quite large. As the cutoff value and the proportion of the study cohort changed, the independent risk factors of the original paper changed from “liver metastasis, levels of ALT, CHE, TBA and TP” to “metastasis stage, ALT, AST/ALT ratio, and LDH” were independent risk factors for prognosis of PD-1 inhibitors therapy, the manuscript of all the data, including all the tables and figures, has made major revision.
Due to the short revision time, some parts including the Quality of English Language may not be perfect. I hope you can forgive us and look forward to your comments and suggestions again. Thank you.
Sincerely yours,
Hao Chen
2023-07-09
Author Response File: Author Response.docx
Reviewer 2 Report
The revised manuscript by Liang et al. addresses some of the problems with the original. The most important correction being the distant metastasis issue.
However, some issues remain suggesting that the short time allowed for revisions was inadequate to contact all authors and have each address their parts comprehensively.
For example, the corresponding author did include follow up measures “Fustat 8-12 weeks” and “PFS fustat” which again are “raw” table names instead of proper labels suitable for publication ready material. The added variables are somewhat confusing since for progression free survival (new table 1) a person is not necessarily “dead”.
Similar previous issue 14 "not publication quality labels in figures" remains unaddressed. The authors state “After trying to modify the instruction, an error is reported and the drawing cannot be made.” suggesting that the contacted authors do not know how to use the R package which is really inappropriate to say while revising a statistically heavy manuscript.
The reply to reviewer's letter does address the issue of validation and training groups being unequal size. But again, the revised manuscript itself doesn’t include this important bit of information. The authors could state at P2 that patients were divided into two groups at random in a 2:1 ratio. The reply specifies “6:4,7:3 or 8:2, and rarely 1:1.” So maybe another target ratio is more suitable. What software was used to make randomization at the chosen ratio?
The authors state “ Reply:The time of PFS was not calculated by randomization time.” Yet the manuscript still states “PFS was outlined as the period of time from randomization to a patient's tumor progressing or dead.” Meaning that something is wrong, either the answer or the manuscript text. The sentence is now even more grammatically incorrect.
In the new Figure 2 with newly added KM graphs, there is a discrepancy in panel B (training cohort N stage) with the Table 1. Table states there should be 50 N0-N1 cases while figure has 51 patients at risk. Unfortunately I do not personally know enough about NPC in general to know whether lack of N stage differences is expected, but knowing about other cancer types I would feel the difference would be more pronounced (significant?) if groups were N0 vs N1+.
Table 1 has been revised to be more readable, but there is no space between numbers and brackets containing the percentage.
The same issue with spacing around brackets persists throughout the revised manuscript.
The authors did not provide further details about the measured parameters. And repeat that the cutoffs for numerical measurements were made by X-tile software. These automatic (?) cutoffs persist in the manuscript even if they are likely meaningless clinically. Ie both CHE and ALT parameters still have a cutoff that makes very problematic groups (as seen in the at risk tables below KM plots with only 2 patients at risk for ALT<9.4). The erroneous cutoff for ECOG is still present in Figure 1 (0-1 vs 2) even though it is now deleted/hidden from table 1. It is also still impossible to say whether other cutoffs are acceptable or not since at no place in the manuscript the authors provide information about how many patients are in each group that underlies Figure 1.
To this issue the authors reply “Please forgive us for the limitation that only 162 patients received PD1 treatment, so the data in the verification group could not be changed.”. I would think this might be a language/communication issue since the problem is not with the size of the patient cohort.
What is not acceptable is the blind adherence to X-tile software results. Someone knowledgeable in statistics must look at the output of that software and determine whether it is suitable (clinically and statistically) to use or not. And apparently this was not done either in the original manuscript or during revision. The authors should re-evaluate the cutoffs and use the X-tile only as a guideline. Apparently in X-tile software, the cutoffs can be manually adjusted and evaluated instead of simply taking the cutoff with highest chi square value. The authors already did not take the grouping into 3 subgroups which is apparently the X-tile default as described in the original publication.
The authors further state “We tried hardly to make changes, but all the results in the whole paper were changed, losing the purpose of the original study.”. which is somewhat concerning. It is impossible to know what changes the authors made, but if some minor changes led to completely different conclusions, it might mean that the results shown in the current manuscript are not as robust as needed for publication and the same could likely not be reproduced by other subsequent studies. Publishing uncertain results would only add to the noise within literature instead of new reliable knowledge. And publishing an uncertain nomogram might cause actual harm to the patients so extra care is warranted
The manuscript still requires thorough editing
Author Response
Dear reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers' constructive comments concerning our article entitled “Nomogram based on liver function tests indicators for survival prediction of Nasopharyngeal carcinoma patients combined with PD-1 inhibitors therapy” (Manuscript No: 22422905). These comments were all valuable and helpful for improving our article. According to the editor and reviewers' comments, we have made extensive modifications to our manuscript and supplemented extra data to make our results convincing. In this revised version, changes to our manuscript were all highlighted within the document by using red-colored text. Point-by-point responses to the nice associate editor and the nice reviewer are listed below this letter.
- Our previous study of the training cohort and the validation cohort were changed from the original 117:45 to 108:54 (2:1) proportional random. Thank you for your correction, which let us know that the best cutoff value using X-tile software was wrong. It was because of your timely correction that we could timely correct the error in the use of this statistical method. So we used the 'surv_cutpoint' function of the 'survminer' R package to calculate cutoff values for numerical variables in the training cohort, and the cutoff value was verified by the “MedCalc” software.
- As the cutoff value and the proportion of the study cohort changed, the independent risk factors of the original paper changed from “liver metastasis, levels of ALT, CHE, TBA and TP” to “metastasis stage, ALT, AST/ALT ratio, and LDH” were independent risk factors for prognosis of PD-1 inhibitors therapy, the manuscript of all the data, including all the tables and figures, has made major revision.
- The outcomes data has been added in the new Table 1, and the Kaplan-Meier curves labels had all been changed. With the change of at random with 2:1 ratio between the training cohort and the validation cohort, some situations in the validation cohort where the number of individually representative cases were too low were settled.
- The spacing around brackets and numbers has also been modified. Due to the short revision time, some parts may not be perfect, including the Quality of English Language. I hope you can forgive us and look forward to your comments and suggestions again. Thank you.
Sincerely yours,
Hao Chen
2023-07-09
Author Response File: Author Response.docx
Reviewer 3 Report
I believe that the authors have provided sufficient answers to my questions/comments
Author Response
Dear reviewer:
On behalf of all the contributing authors, I would like to express our sincere appreciations of your letter and reviewers' constructive comments concerning our article entitled “Nomogram based on liver function tests indicators for survival prediction of Nasopharyngeal carcinoma patients combined with PD-1 inhibitors therapy” (Manuscript No: 22422905). These comments were all valuable and helpful for improving our manuscript.
Thanks for the reviewer's affirmation of this paper, but one reviewer was not satisfied with the original revised paper, so I tried to make a major revision of my paper again, and this change was quite large. As the cutoff value and the proportion of the study cohort changed, the independent risk factors of the original paper changed from “liver metastasis, levels of ALT, CHE, TBA and TP” to “metastasis stage, ALT, AST/ALT ratio, and LDH” were independent risk factors for prognosis of PD-1 inhibitors therapy, the manuscript of all the data, including all the tables and figures, has made major revision.
Due to the short revision time, some parts including the Quality of English Language may not be perfect. I hope you can forgive us and look forward to your comments and suggestions again. Thank you.
Sincerely yours,
Hao Chen
2023-07-09
Author Response File: Author Response.docx
Round 3
Reviewer 2 Report
The revised version appears to address all major concerns and the results more robust.
some proofreading is still required, especially around brackets and other interpunction signs
ie "Creatine Kinase(CK), and"
"(TG),cholesterol (CHO),"
