Concordance of HER2 Expression in Paired Primary and Metastatic Sites of Endometrial Serous Carcinoma and the Effect of Intratumoral Heterogeneity

Comments 1: Small amount of cases: n= 16. Some more cases would have been great!

Response 1: Thank you for pointing this out. I/We agree with this comment. However, endometrial serous carcinoma is very rare tumor. In our institution, which specializes in gynaecology and obstetrics, this is all the cases that we could find.

Comments 2. With variations in heterogeneity patterns between techniques!

Probably you should add some information for the reader that he understands what you mean by pattern.

Response 2: Thank you for pointing this out. I/We agree with this comment. We have defined the heterogeneity patterns by clustered and mosaic as supported by Buza et al in line 68 onwards showing difference in patterns. Furthermore IHC can also be patchy as well. This correlates well with clustered and mosaic heterogeneity. This can be explained that each tumor cell has variable levels of HER2 gene transcription and copy number. If it is isolated, this would result in a mosaic pattern. If it is amplified with many gene copy numbers this would result in a clustered pattern and thus by international definition classified as HER2 amplified.

Comments 3. With a combined concordance rate of up to 68%: What does it mean concordance rate?

Probably you mean the combination of the results of your study and that of Halle et al.

So, you have to state this exactly.

Response 3: Thank you for pointing this out. I/We agree with this comment. We have revised it to state this exactly in line 243.

 Comments 4. Introduction:

Line 52: Recent research encompassing of 2,192 cases of endometrial serous carcinomas of which 52 1,686 (76.9%) were primary tumors and 506 (23.1%) were metastatic tumors highlighted 53 the incidence of HER2 status is significantly influenced by the diverse testing methods 54 and interpretation guidelines employed.

Probably I missed the reference for this citation.

Response 4: Thank you for pointing this out. I/We agree with this comment. We have revised it and added the reference.

 Comments 5. Line 65: Although a universally accepted 65 definition of HER2 amplification is lacking, Albarello et al defined HER2 heterogeneity in gastric cancers as a 2-pont difference in HER2 expression in at least 5% of the tumor. 14

You mix HER2 amplification (first sentence) with the immunohistochemically detected expression (second sentence). HER2 heterogeneity can be evaluated without definition of amplification.

Response 5: Thank you for pointing this out. I/We agree with this comment. We have revised it from amplification to mean heterogeneity.

 Comments 6. Line 73: More recently, Sherry et al demonstrated that HER2 genetic intratumoral heterogeneity defined as HER2 amplification in 5% to 50% of tumor cells examined by FISH was identified in 19/53 cases of high-grade endometrial carcinomas.15

As far as I know: Sherry is the first name and Shen the family name.

Response 6: Thank you for pointing this out. I/We agree with this comment. We have revised it to Shen in ling 74.

 Comments 7:  Material and Methods:

Line 110: 2.2. HER2 Dual in-situ hybridization (DISH)

Who evaluated DISH? Only FY (see Table 3)?

Response 7: Thank you for pointing this out. FY and YW evaluated DISH

 Comments 8:  

Line 113: 2.3. Fluorescence in-situ hybridization (FISH)

Who evaluated FISH?

Response 7: Thank you for pointing this out. FISH was evaluated by Joanne Peverall.

 Comments 9:  Line 118: 2.4. Scoring of DISH and FISH:

How did you define tumor heterogeneity? Probably I missed the definition you used for heterogeneity.

Line 120: If there was heterogenous amplification in clusters, 20 contiguous tumor cells with the highest amplification were counted to obtain the HER2/CEP17 ratio and average HER2 copy number per cell.

Do you think that it is a good idea to call a tumor amplified for HER2 if you only count 20 tumor cells with the highest amplification? Would you take Herceptin if only 10% of the entire tumor cells are amplified and the rest not? Wouldn´t it make more sense to tell the clinicians and the patients the percentage of amplified cells?

Line 123: In mosaic heterogeneity, isolated tumor cells with HER2/CEP17 ratio of more than or equals to 2.0 in a background that was unamplified, 20 contiguous cells should be counted and if the ratio was close to 2, up to 50 contiguous cells were counted in FISH and up to 40 contiguous cells were counted in DISH, recommended by the manufacturer, based on the breast protocol .18, 19

The same as above. Wouldn´t it be more appropriate to talk of subclones which may be amplified and to give the percentage of the subclone?

Response 9 : Thank you for pointing this out.

Although a universally accepted definition of HER2 heterogeneity is lacking, Albarello et al defined HER2 heterogeneity in gastric cancers as a 2-pont difference in HER2 expression in at least 5% of the tumor. ¹ Buza et al further categorized heterogeneity, defining clustered heterogeneity as involving groups of tumor cells (20 cells or more) and mosaic heterogeneity as isolated amplified tumor cells within a background of unamplified tumor cells. We have followed Buza’s recommendation on tumour heterogeneity.

Thank you for your comments. We/I agree that statement. However, this is the current international guidelines to reviewing 20 cells as the standard but also for practical reasons. As for FISH, they can only analyse a small circle of cells, as they have to manually count the number of copies per cell which is labour intensive. Perhaps in the future, with AI, we could expand the counting to more cells. I also agree that if there was only a small proportion of cells that are highlighted we do not yet know the significance of this. In majority of the cases with clustered heterogeneity they were HER2 amplified and this would also be seen in other cells that are beyond the 20 cells that were counted. But again as per international guidelines we do not report more than 20 cells. While in mosaic heterogeneity, it is only isolated cells that have increased HER2 gene copy numbers and do not yet meet the criteria for amplification.

This would be similar to PDL1 status in lung as not all the tumor cells would highlight with PDL1.

 Comments 10:  Results:

Line 135: 8/10 cases scored HER2 IHC 0 or 1+, and 2/10 cases scored equivocal, HER2 IHC 2+, case 10 and 14, Table 3. Amongst the 6 HER2 amplified cases, 4/6 cases were HER2 IHC 3+, 2/6 cases were equivocal by HER2 IHC 2+, case 7 and case 11, Table 3.

Table 3: Case 14: YW and FY differed in their evalution of HER2: Why did you take YW for the final evaluation? The same happens in cases 7 and 9 e.g.

Response 10: Thank you for pointing this out.  YW is more senior and thus has more experience in evaluating HER2 IHC, thus her results were taken instead of FY.

 Comments 11:  Table 3: Was there no polysomy of chromosome 17? Don´t you think that a pure amplification is quite different to an amplification with a polysomy of chromosome 17?

Table 3: IHC-Score: Please provide not only numbers, but add +: 1+, 2+, 3+.

.

Response 11: Thank you for pointing this out.  No, there was no polysome of chromosome 17 in our cases. Yes, it would be very different if there was polysomy of chromosome 17 as highlighted in the breast cancer guidelines for HER2.

Thank you for pointing this out. We have adding  +: 1+, 2+, 3+.

 Comments 12:  Line 150: 3.1. Intratumoral heterogeneity

Didn´t you also find some heterogeneity in immunohistochemistry?

You only describe the ISH testing.

Once again: How do you define heterogeneity?

.

Response 12: Thank you for pointing this out.  We have defined heterogeneity as above. Based on the HER2 breast guidelines, they assess HER2 based on the cut off of 10%. The criteria that Fader and Buza et al, and most recently CAP have recommended is 30%. I agree that there is some heterogeneity in immunohistochemistry, but the significance of which is difficult to determine. But in general, most of our cases with HER2 IHC 3+ these cases showed diffuse with much more than 30% IHC staining. Those with 2+ were more challenging, depending on whether the scorer believes it occupies more or less than 30% of the entire tumor cells. But in general, this has not affected the results as all those with HER2 3+ were amplified, and None of the HER2 1+ or 0 were amplified.

 Comments 13:  Discussion:

Line 177: analytical factors including, cold ischemic time, of less than 1 hour and fixation in 10% neutral buffered formalin for 6 to 72 hours. 20

What do you mean by “of less than 1 hour”?

Line 195: If only the hysterectomies were tested for HER2 amplification, two out of six 195 cases would have been falsely positive, none would have been falsely negative.

For a better understanding you may add: falsely positive concerning the metastasis.

Line 196: Similarly, if only the metastasis were tested, two of the ten cases would be falsely negative compared to hysterectomies, none would have been falsely positive.

Why not two of twelve cases instead of ten cases?

.

Response 13: Thank you for pointing this out. I/We meant cold ischemic time of less than 1 hour, have removed the comma.

Thank you for pointing this out. We/I have revised to say falsely positive concerning the metastasis. In line 221.

Agreed. Should be two of the twelve cases. Have revised this in line 222.

1. Summary

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

Comments 1: 1) In line 27 it was mentioned "meta-analysis" while actual meta-analysis has not been performed, because descriptive statistics performed on combined samples from just two studies cannot be considered a meta-analysis!

Response 1: Thank you for pointing this out. I/We agree with this comment. Therefore, I/we have changed it to “a review”

Comments 2: Methods' descriptions are too scarce for someone not familiar with those techniques, and stating that something has been done with autostainer is not sufficient!

All methods must be properly described (what they do and how), with enough details so anyone could repeat them. E.g., what is unclear what CEP17 is, how it is measured, and at the end what it is actually used for.

Methods' acronyms should be uniform throughout the text, since you mixed DISH, SISH and FISH, and now it is unclear what is what!

Response 2: Agree. I/We have, accordingly, revised the paper and changed all the SISH to DISH to reflect the same technique. With regards to autoimmunostainers, this is a commercial machine that is used commonly in most laboratories under Ventana and readily available. The exact workings of the machine would be proprietary information. We have provided the model and the methodology, which is through indirect polymer method at the appropriate dilutaion factor for repeatability and reproducibility. Similarly, the DISH and FISH platforms, we are using commercially available machines and probes which can be looked up via the vendors website.

Comments 3: There are lots of unreferenced text in both 'Introduction' and 'Discussion', e.g., text in lines 52-57, 57-61, 68-70, 70-72, 72-73, 199-205, 222-225, etc.

It is unclear if paragraphs in lines 206-212 and 213-219 correspond to unreferenced study mentioned in lines 199-205.

References seem improperly numbered, e.g., in lines 49-50 it was written Fader et al.¹¹, while the 11th reference in the list is Buza et al.

Response 3: Agree. I/We have, accordingly, revised and added all the references as stated in

lines 52-57, 57-61, 68-70, 70-72, 72-73, 199-205, 222-225.

Yes lines 206-212 and 213-219 correspond to referenced study and has been referenced in lines 199-205.

I have changed the reference to the one by Fader et al.

Comments 4: In line 94 and Table 3 it was mentioned "32 cases"; however, those are samples, not cases, since you had only 16 cases/patients.

Response 4: Agree. I/We have, accordingly, revised it to samples.

Comments 5) Table 3, per sample results, should be presented first, since the remaining two tables are summary results.

Response 5: Agree. I/We have, accordingly, have revised this.

Comments 6) All abbreviations and acronyms presented in tables must be explained in their footnotes.

Response 6: Agree. I/We have, accordingly, have added the footnotes.

Comments 7) Authors should present representative figures of different levels of HER2 IHC staining.

Response 7: It is not feasible to do this as not all the authors are pathologists and not all are subspecializing in gynaecology.

Comments 8) Authors should uniformly and appropriately write only approved gene and protein symbols, so it would be clearly evident when the text is about gene amplification and when about protein over-expression. Thus, the approved gene symbol is ERBB2, while protein erbB-2, HER2 or Neu.

Response 8: Agreed. The gene symbol is ERBB2 and the protein should be annotated as HER2. We did annotate this in the original manuscript. However, this became very confusing to some readers and the reviewers recommended that we use HER2 only.

Comments 9) Authors should put more effort to explain discrepancies between gene/protein expression and techniques, e.g., ERBB2/HER2 transcription/translation regulation, respectively.

Response 8: Unsure. I/We are not sure exactly which discrepancies you mean. All our cases with HER2 IHC 3+ showed concordant amplification with DISH and FISH.

With regards to HER2 2+, It is well known that HER2 IHC2 requires further testing as recommended by international guidelines as there could be some transcription of the HER2 gene copy but insufficient to meet the cut off requirements or criteria for amplification. Finally, HER2 +1/0 is graded as negative by international guidelines.  

Comments 1) Full name of genes, proteins, techniques... should not be written with the first capital letter.

Response 1: Agreed. I/We have revised it. 

Comments 2) Line 155: Word "correlation" should be used only in the context of calculated correlation coefficient, otherwise "association" should be used.

Response 2: Agreed. I/We have revised to association.