A Real-World Study Reporting the Use of Foundation Medicine^® Testing in Portugal

Round 1

Reviewer 1 Report

 

The authors present the first real-world data on NGS testing of 67 cancer in Portugal using three commercial tests of Foundation Medicine. Although there are no novel findings or insights, the paper provides a realistic picture of Foundation One testing in real-world conditions, though in a limited number and variable types of tumors. I have several comment and suggestions for consideration:

 

1.     In general clinical practice, a large proportion of NGS testing refers to non-small cell lung cancer. The authors should explain why NSCLC was not part of this study. They should also briefly discuss the impact and yield of NGS testing for NSCLC (e.g. by Foundation Medicine) in the literature (e.g. PMID 35397297).

2.     The low number of only around 72 patients within a time period of 3.5 years indicates that only a small portion of NGS testing in Portugal is done by Foundation Medicine. Is there a certain selection process by which the specimens were directed towards Foundation Medicine? Apparently, all the NSCLC testing is done using different approaches.

3.     The material was restricted to tumor tissue. Why weren’t there any cytology specimens, which make a significant proportion of NGS testing in clinical practice?

4.     Turn-around time (TAT) is an issue, and a median TAT of 15 days appears high to me, considering that there are other platforms/solutions with a TAT of around 5 days. There should be more details on the TAT: Working days or calendar days? Time from sending out to receiving the result or time from receiving the material to sending out the report?

5.     Table 4: it would be informative to show whether the alterations were CNVs, SNVs  or rearrangements (three additional columns). E.g. ERBB2 can be amplified or mutated. Listing TMPRSS and ERG as a separately altered genes makes no sense, since they belong to the same unique TMPRSS2 -ERG rearrangement in prostate cancer. It is sufficient to call ERG rearrangement.

6.     The main body of data is from the 63 FoundationONE CDX. It might be ok to lump them with the few Liquid CDX and Heme data. Nevertheless, the data from the liquid CDX and Heme analyses should also be shown separately.  Otherwise, we have no idea if there were any useful data coming from them.

Author Response

1. In general clinical practice, a large proportion of NGS testing refers to non-small cell lung cancer. The authors should explain why NSCLC was not part of this study. They should also briefly discuss the impact and yield of NGS testing for NSCLC (e.g. by Foundation Medicine) in the literature (e.g. PMID 35397297). Response 1 In our laboratory, NSCLC is tested in a specific lung cancer test, Oncomine Customized Lung Cancer Test from Thermo-Fisher, not by the Foundation Medicine Test. Therefore, it was out of the scope of this study. In the study period (Jul 2017 – Dec 2020) 36% of the tests performed in our laboratory corresponded to lung cancer. The following sentence was added to the discussion, Line 243-248: “It is also important to note that samples from lung cancer, namely non-small cell lung cancer (NSCLC) were tested in a specific panel and not by Foundation Medicine. Despite that, it is important to highlight that for patients with newly diagnosed non-squamous advanced NSCLC, the use of NGS testing may help to avoid potentially missed targeted therapy options and improve uptake of testing for recently approved biomarkers and access to clinical trials of molecularly guided therapies[34]. ”

 

2. The low number of only around 72 patients within a time period of 3.5 years indicates that only a small portion of NGS testing in Portugal is done by Foundation Medicine. Is there a certain selection process by which the specimens were directed toward Foundation Medicine? Apparently, all the NSCLC testing is done using different approaches. Response 2: The period of this study corresponds to the implementation of Foundation Medicine in Portugal. Therefore, the prescription of the test was not available in all the centers and IPATIMUP was only performing the referenced ones.  The reviewer was correct. NSCLC testing was done using Oncomine Customized Lung Cancer Test from Thermo-Fisher implemented in our lab since 2014.

 

3. The material was restricted to tumor tissue. Why weren’t there any cytology specimens, which make a significant proportion of NGS testing in clinical practice? Response 3 During the period of this study, the use of cytology specimens was not optimized. Moreover, as explained in the previous question, lung cancer testing was performed by other methods and cytology is mainly used for lung cancer.

 

4. Turn-around time (TAT) is an issue, and a median TAT of 15 days appears high to me, considering that there are other platforms/solutions with a TAT of around 5 days. There should be more details on the TAT: Working days or calendar days? Time from receiving the material to sending out the report? Response 4: At the period of this study, the platform with a TAT of around 5 days was not available. For Foundation Medicine, TAT is 14 calendar days. TAT is defined as the time from receiving the material to sending out the report. These details were included throughout the manuscript.

5. Table 4: it would be informative to show whether the alterations were CNVs, SNVs  or rearrangements (three additional columns). E.g. ERBB2 can be amplified or mutated. Listing TMPRSS and ERG as a separately altered genes makes no sense, since they belong to the same unique TMPRSS2 -ERG rearrangement in prostate cancer. It is sufficient to call ERG rearrangement. Response 5: The information about CNVs, SNVs, or rearrangements was included in Table 4. The information about TMPRSS and ERG was corrected accordingly to reviewer suggestion.

6. The main body of data is from the 63 FoundationONE CDX. It might be ok to lump them with the few Liquid CDX and Heme data. Nevertheless, the data from the liquid CDX and Heme analyses should also be shown separately.  Otherwise, we have no idea if there were any useful data coming from them. Response 6: The information about Liquid CDX and Heme data were separated in Table 4.

Reviewer 2 Report

This manuscript may be accpeted without revisions becuase it share interesting data to scientific community

Author Response

This manuscript may be accepted without revisions because it share interesting data to scientific community.

We thank the reviewer for this comment.

Reviewer 3 Report

Reports on the application of commercially-available gene sequencing panels in various populations may be of interest to clinical practitioners.

 

I have the following concerns:

 

There are some deficiencies in the English language of the report, but nothing that prevents the point from coming across (e.g. in the abstract: “A reduced number of potentially actionable variants” instead of “A limited number”).

 

Line 50: “Different reports have shown that Foundation Medicine® testing can identify additional information on genomic alterations”. What additional information?

 

Line 77: “Cohen’s Kappa coefficient (k) was calculated to determine the correlation between the number of altered genes and the TMB value.”  I don’t think Cohen’s kappa is the correct statistic to use for this.

 

Line 103: How was the percentage of tumor cells quantified?

 

Line 119: Why is TMB data only available for 62 of the samples?

 

Line 124: All patients are either MSI-stable (n=60) or could not be determined (n=7).  How were stable, unstable, and indeterminant defined in this study?

 

Line 126: “Among 67 profiled samples, 25 (37.3%) presented with potentially actionable genomic alterations by FDA and EMA-approved drugs (36.0%) or by experimental drug(s) in clinical trials (64.0%).”  This sentence is difficult to understand on first reading, perhaps say “with potentially actionable genomic alterations by either FDA and EMA-approved drugs (n=9, 36.0%) or by experimental drug(s) in clinical trials (n=16, 64.0%).”  

 

Line 192: Does this mean your definition of TMB high vs low differs from that of previous studies?  How does your definition differ and why was it chosen?

Author Response

1. There are some deficiencies in the English language of the report, but nothing that prevents the point from coming across (e.g. in the abstract: “A reduced number of potentially actionable variants” instead of “A limited number”). Response 1: The English language was polished.

2. Line 50: “Different reports have shown that Foundation Medicine® testing can identify additional information on genomic alterations”. What additional information? Response 2: Additional information refers to the genomic alterations that are included in comprehensive genomic panels (CGPs) beyond the actionable alterations. This was clarified in the manuscript, Line 50-54: “Different reports have shown that Foundation Medicine® testing, through CGP can identify additional information on genomic alterations, beyond the actionable ones and that possible treatments may apply. Moreover, these comprehensive panels also contribute to the increased availability of clinical trials [2,12-15].”

 

3. Line 77: “Cohen’s Kappa coefficient (k) was calculated to determine the correlation between the number of altered genes and the TMB value.”  I don’t think Cohen’s kappa is the correct statistic to use for this. Response 3: The reviewer is correct. Spearmen´s rho correlation was used. Methods (line 95) and results (line 140) were modified accordingly: Methods: “Spearman's rho was calculated to determine the correlation between the number of altered genes and the TMB value.” Results: “A positive correlation was found between the number of altered genes and TMB for the samples tested with FoundationOne CDX (r=0.400; p=0.001)”

 

4.Line 103: How was the percentage of tumor cells quantified? Response 4: This is clarified in the methods section, line 71 “For this study the estimation of tumor cells was obtained through the ratio between the nuclei quantification of neoplastic cell vs non-neoplastic cells, as previously described by Gullo I, et al[16]”

 

  5. Line 119: Why is TMB data only available for 62 of the samples? Response 5:  At the study period, TMB information was only available for Foundation Medicine CDx (63 samples). However, one of the samples was described as poor quality that could impact the determination of TMB: can not be determined.  

 

6.Line 124: All patients are either MSI-stable (n=60) or could not be determined (n=7).  How were stable, unstable, and indeterminant defined in this study? Response 6: This is clarified in the methods section, line 77 “To determine a patient’s MSI status, Foundation Medicine® employs a fractional based (FB) MSI algorithm to categorize a tumor specimen as MSI-High (MSI-H) or microsatellite stable (MSS). The FB-MSI algorithm calculates the fraction of microsatellite loci determined to be altered or unstable (i.e., the fraction unstable loci score) based on a genome-wide analysis across >2000 microsatellite loci. For a given microsatellite locus, non-somatic alleles are discarded, and the microsatellite is categorized as unstable if remaining alleles differ from the reference genome. The final fraction unstable loci score is calculated as the number of unstable microsatellite loci divided by the number of evaluable microsatellite loci. Two FB-MSI score thresholds are applied to classify a tumor specimen as having MSI-H or MSS status. MSI-H status is reported for patients with solid tumors whose samples have FB-MSI scores ≥ 0.0124 while MSS status is reported for patients with solid tumors whose samples have FB-MSI scores ≤ 0.0041. Patients with solid tumors whose samples have FB-MSI scores >0.0041 and <0.0124, an MSI “Cannot be Determined” result is reported.”

 

7. Line 126: “Among 67 profiled samples, 25 (37.3%) presented with potentially actionable genomic alterations by FDA and EMA-approved drugs (36.0%) or by experimental drug(s) in clinical trials (64.0%).”  This sentence is difficult to understand on first reading, perhaps say “with potentially actionable genomic alterations by either FDA and EMA-approved drugs (n=9, 36.0%) or by experimental drug(s) in clinical trials (n=16, 64.0%).”  Response 7: We thank the reviewer for this comment. It was clarified in the text, line 144, according to reviewer suggestions.

 

8. Line 192: Does this mean your definition of TMB high vs low differs from that of previous studies?  How does your definition differ and why was it chosen? Response 8: As explained in the methods section and in line 197: “In our study, samples have been classified in high-TMB (≥10 mutations/Mb) and low-TMB (<10 mutations/Mb) [17,24,25]”. This is the cut-off recommended by Foundation Medicine. The study that we mention in the discussion included a different definition of TMB:  “TMB was dichotomized into two groups; low-to-intermediate (0–19 mutations/mb) vs. high (≥20 mutations/mb)”

Reviewer 4 Report

I have reviewed a manuscript titled: "Real-world study reporting the use of Foundation Medicine® testing in Portugal". There are several concerns that the authors might want to follow-up or further discuss. These concerns are as follow:

1. The major problem of this manuscript is that it seems there were no any novelty.

a. The kit is FDA approved, ready to use, readily available in the market and not developed by the authors. I do not see any novel modifications or other novelty in this kit.

b. I do not mean the kit must be developed or modified by the authors, but even with the use of commercial kit, some novel results can also be obtained.

However, from the results, I do not see any novelty neither. It seems that there is merely a collection of turn-around time, summary of the main tumor types, summary of the frequent altered genes.  These results are too superficial and provide little scientific significance. I do not see any mutation profiles of these genes, nor any detailed description of these mutations in SNV, MSI, TMB, but just a number and percentage presented.

 

c. What so special about these 72 subjects? Can these 72 samples represent the whole population of Portugal or represent any special group of persons? Or any scientific significance from this group of 72 people so that the summary of the main tumor types, summary of the frequent altered genes mean something to the scientific community?

At least I didn’t see any description within the manuscript.  Therefore, based on the above concerns, I do not find any novel and significant scientific content in this manuscript. I almost think this is the advertisement from the company, but advertisements will still provide more details like mutations profiling showing off the advantages of the kit over the competitor, but these information are missing in this mansucript. 

 

 

2. From table 2, the summary of tumor type seems add-up not equal to 63 or 72. Either 9 or 18 samples are missing. 9 samples more than the number of the second largest tumor type, while 18 is even greater than the main tumor type in this study.

 

3. The population size is too small, which is only 72 subjects.

 

4. Objective of this manuscript is the vague and not clear. Essentially everything presented can be considered as “explored” and so it means nothing.

Author Response

1.The major problem of this manuscript is that it seems there were no any novelty.

1. The kit is FDA approved, ready to use, readily available in the market and not developed by the authors. I do not see any novel modifications or other novelty in this kit.
2. I do not mean the kit must be developed or modified by the authors, but even with the use of commercial kit, some novel results can also be obtained. However, from the results, I do not see any novelty neither. It seems that there is merely a collection of turn-around time, summary of the main tumor types, summary of the frequent altered genes.  These results are too superficial and provide little scientific significance. I do not see any mutation profiles of these genes, nor any detailed description of these mutations in SNV, MSI, TMB, but just a number and percentage presented.
3. What so special about these 72 subjects? Can these 72 samples represent the whole population of Portugal or represent any special group of persons? Or any scientific significance from this group of 72 people so that the summary of the main tumor types, summary of the frequent altered genes mean something to the scientific community?

Response 1: This study was performed during the implementation of the Foundation One test in Portugal. Therefore, it is only a descriptive study reporting the main findings during this period and no novelty is expected. In the period of this study, there was an average of 24 tests Foundation Medicine per year. In the last year, our laboratory performed 154 tests, representing an increase of 529%.  

Information regarding genomic alterations and rearrangements is now included in Table 4.

At least I didn’t see any description within the manuscript.  Therefore, based on the above concerns, I do not find any novel and significant scientific content in this manuscript. I almost think this is the advertisement from the company, but advertisements will still provide more details like mutations profiling showing off the advantages of the kit over the competitor, but these informations are missing in this manuscript. 

Reponse 2: A more detailed description of mutations profile was included in Table 4. Regarding the company, as previously included in the Funding and Conflict of Interest Section: “This research was funded by Roche Farmacêutica Quimica Lda.” and “The APC was funded by Roche Farmacêutica Quimica Lda”. “The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.”

Round 2

Reviewer 3 Report

All my concerns have been addressed. 

Author Response

The authors want to thank to the reviewer. 

Reviewer 4 Report

Thank you for your response.

1. I have seen the update of the Table 4. This is for sure a better version than the previous one. However, i have some hersitation on the scientific significance of merely summarize the number of these mutations. maybe some further detailed discussion or interpretation would increase the scientific values so that it better fit the scope of this journal which read as " pre-analytical and analytical issues and the interpretation of biomarker testing in the field of diagnostic and predictive molecular pathology."

Maybe the authors might want to submit to other journal so to better fit the scope of their nice work?

2. I am very sorry, because i may have overlooked, as i did not see any response to the following concerns posted previously, i have copied below for your reference:

"2. From table 2, the summary of tumor type seems add-up not equal to 63 or 72. Either 9 or 18 samples are missing. 9 samples more than the number of the second largest tumor type, while 18 is even greater than the main tumor type in this study.

3. The population size is too small, which is only 72 subjects.

4. Objective of this manuscript is the vague and not clear. Essentially everything presented can be considered as “explored” and so it means nothing."

Author Response

Thank you for your response.

1. I have seen the update of the Table 4. This is for sure a better version than the previous one. However, i have some hersitation on the scientific significance of merely summarize the number of these mutations. maybe some further detailed discussion or interpretation would increase the scientific values so that it better fit the scope of this journal which read as " pre-analytical and analytical issues and the interpretation of biomarker testing in the field of diagnostic and predictive molecular pathology." Maybe the authors might want to submit to other journal so to better fit the scope of their nice work?

Response 1: Thank you for sharing your thoughts on the scientific significance of summarizing the number of mutations. I understand your hesitation and agree that a more detailed discussion or interpretation could align it better with the scope of the journal, which focuses on pre-analytical and analytical issues, as well as the interpretation of biomarker testing in diagnostic and predictive molecular pathology. However, as you can see in the Table the alterations found are well-known targets or others that are common but not targets (p53 for example). All these alterations are well described in terms of their impact on the pathogenesis, diagnosis, and prognostication of the tumors. The goal of the article is to demonstrate the real-world application of this specific molecular test and in our opinion, this is of interest to the pathologists that read the journal because we must be prepared to understand these tests and be the frontline in the molecular diagnosis.

 

2. I am very sorry, because i may have overlooked, as i did not see any response to the following concerns posted previously, i have copied below for your reference:

 

"2. From table 2, the summary of tumor type seems add-up not equal to 63 or 72. Either 9 or 18 samples are missing. 9 samples more than the number of the second largest tumor type, while 18 is even greater than the main tumor type in this study.

Response 2: Table 2 only includes a summary of the most representative tumor types, with tumor frequency ≥ 3 identified in our study. This is the reason why the sum is not equal to 63 or 72.

3. The population size is too small, which is only 72 subjects.

Reponse 3: As already explained this study was performed during the implementation of the Foundation One test in Portugal. Therefore, the prescription of the test was not available in all the centers and IPATIMUP was only performing the referenced ones.  Therefore, it is only a descriptive study reporting the main findings during this period and no novelty is expected. In the period of this study, there was an average of 24 tests for Foundation Medicine per year. In the last year, our laboratory performed 154 tests, representing an increase of 529%. 

 

4. Objective of this manuscript is the vague and not clear. Essentially everything presented can be considered as “explored” and so it means nothing."

Response 4: This is descriptive real-world data reporting the implementation of Foundation One tests in Portugal in a centralized Reference Laboratory. The goal of the article is to demonstrate the real-world application of this specific molecular test and in our opinion, this is of interest to the pathologists that read the journal because we must be prepared to understand these tests and be the frontline in the molecular diagnosis

Round 3

Reviewer 4 Report

Thank you so much for your clear reply. If the authors has incorporate the comment on table 2 into the manuscript, I have no more further question. Or concern on the manuscript.  

 

In my opinion, the novelty and scientific values of this manuscript is not high (if not low), It is because as the authors have mentioned, this manuscript is merely a descriptive of the real-world application of this specific molecular test, which is not very much aligned with the scope of the journal. The manuscript is fine, seriously, but I would recommend to submit to other journal so to better fit the scope of their nice work.

 

As I have no more concerns about the manuscript and it is not a scientific concerns, in this case, I would leave it to the editor to decide whether this manuscript is fit for publish in this journal or not.

Author Response

Thank you for your comments.