A Narrative Approach to Mismatch Repair-Deficient Endometrial Cancer

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This review focuses on the diagnosis and management of mismatched repair deficient endometrial carcinomas. The review is well written, comprehensive, and should be useful to individuals interested in familiarizing themselves with this field. Some of the comments need further clarification or discussion.

 

Specific points:

1. Lines 51 to 54 list the different molecular types of endometrial carcinomas. There should be a new line for each subtype in the text and it would be useful to provide additional information pertaining to each subtype. The POLE mutated subtype shares several features with the mismatch repair deficient subtype, which should be pointed out. For example, both types are characterized by high tumor mutation burden and both are responsive to immunotherapy. The current text suggest that these features are only associated with the mismatch repair deficient subtype. In addition, tumors with POLE mutations have been described showing microsatellite instability without silencing of any of the mismatch repair enzymes; it has been suggested that such tumors may represent a different subtype. Prognostic differences between the different molecular subtypes should also be mentioned. These points could also be discussed on page 6.
2. On page 3, both in the text and in Table 1, a common very mechanism of dMMR that is not mentioned is loss of heterozygosity. Double somatic mutation is extremely rare.
3. Table 1: The information in the first 2 lines (MLH1 and PMS2 and MSH2 and MSH6) and the last 2 lines (MLH1 and MSH2) is redundant.
4. On line 156 it is stated that there is high concordance between immunohistochemistry and PCR based approaches in evaluating MSI status. This is misleading because up to 15% discordance has been observed between the two approaches. Comments regarding concordance between NGS and these 2 approaches are also misleading because discordance has also been documented.
5. It should be stated somewhere that micro satellite instability can occur in normal tissues and that the difference between clinically significant and insignificant instability is based on frequency. Thus, several biomarkers need to be examined, as the diagnosis is based on a statistical argument. Comments should be made on the frequency needed to make a diagnosis of microsatellite instability, and on the minimum number of micro satellite bio markers that should be examined.
6. On line 170, it is said that BRAF testing has no role in triage for MMR gene mutation testing. It would be more accurate to state that a role for BRAF testing has not been validated and its significance is currently unknown for endometrial carcinoma. The presence of BRAF variants could be helpful in distinguishing Lynch syndrome from sporadic micro satellite instability in endometrial carcinoma, but all studies so far have been inconclusive due to lack of statistical power because such a variants are rare in these tumors.

 

Minor points:

1. Line 420: tobioinformatically

Author Response

Comment 1… The POLE mutated subtype shares several features with the mismatch repair deficient subtype, which should be pointed out….

Thanks to the Reviewer 1 to give us the opportunity to describe the shared biological characteristics of POLE mutated and dMMR ECs. A new paragraph has been inserted in the introduction, line 81 et seq.

"

It is important to note that POLE mutated tumors share biological characteristics with dMMR ones. The mutations localized in the proofreading domain of the gene impact on DNA replication, leading to proofreading defects and result in an ultramutated tumor with over 100 mutations per Mb and a favourable prognosis (16).The tumor neoantigens produced by POLE mutated ECs are 15-fold higher than that of dMMR/MSI-high and more than 100-fold higher than that of pMMR/MSS (17). The high TMB and the immunogenic profile are pivotal parameters for I-O benefit (18).  This ultra-high mutation rate is usually associated with MSS. However, there are rare examples of POLE-mutated ECs with concomitant dMMR, a more aggressive molecular subtype with a high TMB and frequent relapses (19)."

Comment 2. ...a common very mechanism of dMMR that is not mentioned is loss of heterozygosity…

Thanks to the Reviewer  for this observation. We inserted comment in paragraph 2, line 122:

"

"The complete inactivation of the tumour suppressor genes of MMR machinery can be reached by loss of heterozygosity (LOH), (i.e.loss of function of both the alleles). dMMR can arise when of the genes has one inherited or somatic pathogenic variant (first hit), and a second hit on the wild type allele. In Lynch syndrome–associated tumours, LOH affecting the wild type allele of a MMR gene is a mechanism that can convert a heterozygous germline mutation into full functional loss and results in MSI high/dMMR phenotype (26)."

Comment 3.....The information in the first 2 lines (MLH1 and PMS2 and MSH2 and MSH6) and the last 2 lines (MLH1 and MSH2) is redundant...

Thank you for the observation. But we prefer to maintain the text as it is; we believe that is more readable.

Comments 4 and 5 . 

4)….. Comments regarding concordance between NGS and these 2 approaches are also misleading because discordance has also been documented.

5) ---- Comments should be made on the frequency needed to make a diagnosis of microsatellite instability, and on the minimum number of micro satellite bio markers that should be examined

Thank you for this important alert. We have commented your observations in a single paragraph which describes both the causes  of discordance between NGS and other techniques than the number of microsatellite loci necessary to reach an affordable result (line 184)

"

However, if the concordance among different tests is high, a clinically meaningful percentage of cases shows discordant results with obvious implications on the therapeutical options. The molecular mechanisms involved in discordance are mainly due to MSH6 mutations with functional compensation by MSH3, and MLH1 loss predominantly driven by promoter methylation (40). MSI status can be impacted by factors that fall outside the protein expression, mainly by POLE alterations and/or by the detection thresholds that may limit the identification of faint MMR alterations.

 The comprehensive profile offered by NGS represents an invaluable tool in result interpretation and consequently in diagnostic accuracy. The most advanced NGS assays eliminates the limit of PCR-based techniques: the low number of evaluated loci. NGS-based MSI can simultaneously detect hundreds to thousands of microsatellite loci, limiting some drawbacks such as ethnicity-associated variations or variations related to tumor type (41). But these high performing assays need bioinformatic algorhytms to detect the percentage of instable loci and specialized personnel. As reported by Swaerts et al (42), MSI evaluation can be performed in different ways : (A) by comparing indel distributions of microsatellites in paired tumor-normal samples, on exome data with computational tools (MSIsensor, MANTIS, MSIseq) or targeted gene panels (USCI-msi), (B) by comparing indel distributions in a tumor sample versus a fixed reference set on targeted gene panels (mSINGS, MSIFinder), or (C) by analysing the number of single nucleotide variants and indels throughout the genome to detect the hypermutated status as consequence of dMRR, (MSIseq)."

Comment 6. ....it is said that BRAF testing has no role in triage for MMR gene mutation testing...

Thank you for this comment. The phrase was changed as follows(line 216):

" 

In colorectal cancers BRAF p.V600E mutation helps to exclude LS in patients with MSI-high tumours, allowing those who are BRAF wild-type to proceed with further LS screening. In EC the rare BRAF mutations (mainly the p.V600E) are not considered efficient biomarkers to distinguish LS from sporadic MSI; currently these alterations are not validated in clinical use for predicting MMR-negative mutation status, (45, 46).

Reviewer 2 Report

Comments and Suggestions for Authors

1 The title "A Narrative Approach..." is appropriate. However, the abstract and introduction could more explicitly state the review's specific aims and scope. For instance, will it focus equally on biology, diagnostics, and therapeutics, or is there a particular emphasis? A brief sentence outlining the review's structure in the introduction would enhance readability.

2 While the TCGA/ProMisE molecular classification is well-described, its critical role in interpreting the historical clinical trial data (e.g., the failure of KEYNOTE-28 vs. the success of KEYNOTE-158) could be emphasized more strongly in sections 2.4 and 2.5. This is a key educational point for the reader.

3 This is an excellent and important section. The summary stating that current data supports dMMR as a predictive biomarker regardless of the mechanism of MMR loss is clear. However, the conflicting data from the phase II trial by Bellone et al. (reference 47, showing worse PFS/OS for hypermethylated tumors) versus the phase III trials (references 59, 60, showing no significant difference) deserves a more nuanced discussion. Could this be due to trial design (phase II vs. III), sample size, the specific immune checkpoint inhibitor used, or differences in the studied populations? A brief speculative analysis would add value.

4 This section is a strength of the manuscript. To improve flow, consider introducing a brief table summarizing the discussed resistance mechanisms (e.g., Impairment of Antigen Presentation, TME features, Gene Signatures) and their key associated biomarkers (B2M, JAK1, CD8+/PD-L1 spatial profiles, etc.). This would provide a quick reference for the reader.

5 The discussion on liquid biopsy and microbiome is relevant. However, the paragraph on microbiome (lines 447-468) ends on a somewhat negative note regarding methodological challenges. While this is accurate, the section could be balanced by also mentioning the potential of this field and what rigorous future studies should entail (as hinted at in the conclusion), rather than primarily highlighting the current pitfalls.

6 The sentence "Actually, the adoption of immunotherapy (I-O) upon failure of platinum-based chemotherapy constitutes a decisive improvement..." (lines 21-22). The word "Actually" is slightly informal. Consider replacing it with "Currently," or "Notably,".

7 The description of IHC patterns is good. For even greater clarity, consider adding a simple schematic figure (if allowed by the journal) illustrating the heterodimer relationships (MLH1-PMS2, MSH2-MSH6) and the consequences of losing one protein.

8 The reference list is extensive and generally appropriate. Please double-check the consistency of journal name abbreviations and the formatting of DOIs/URLs per the journal's style guide.

 

Author Response

Comment 1. ....However, the abstract and introduction could more explicitly state the review's specific aims and scope….

I thank the Reviewer  for comments. The abstract has been modified according to the suggestions and all the described topics present in the review have been reported.

Comment 2.    While the TCGA/ProMisE molecular classification is well-described, its critical role in interpreting the historical clinical trial data (e.g., the failure of KEYNOTE-28 vs. the success of KEYNOTE-158) could be emphasized more strongly in sections 2.4 and 2.5. This is a key educational point for the reader...

The reviewer underlined a key point in the evolution of immunotherapy in dMMR tumours. We have inserted a paragraph in section 2.4 to emphasize this point ( line 271)

"

"The trial showed the critical importance of biomarker-guided therapy. The use of biomarkers increases the quality of diagnosis, indicates a therapeutic strategy and provides predictive elements of response to ICIs. The histological and molecular interpretation of biomarkers and their reporting must be matched by the understanding of biological mechanisms at the clinical level to optimize therapeutic options. Pathology Departments must in turn use standardized and controlled protocols to achieve robustness and reproducibility of the assays used to identify dMMR tumours"

Comment 3.....However, the conflicting data from the phase II trial by Bellone et al. (reference 47, showing worse PFS/OS for hypermethylated tumors) versus the phase III trials (references 59, 60, showing no significant difference) deserves a more ced discussion...

Thank you for this precious comment. A paragraph has been inserted in paragraph 2.5, line 309

"However, a lower response to ICIs in hypermethylated MLH1 tumors compared to mutated MLH1 has been reported (64, 65), but a clinical significance was not achieved in view of the relatively small proportion of MLH1 mutated tumors within the general dMMR population. It is therefore necessary to prospectively evaluate these different biological mechanisms in clinical trials through which it will be possible to define when ICIs therapy alone is usable or when combination therapy is more appropriate."

Comment 4.  To improve flow, consider introducing a brief table summarizing the discussed resistance mechanisms 

Thank you for your suggestion. A summary table of the main mechanisms of resistance and related biomarkers has been inserted at the beginning of paragrafh 2.7

Comment 5.   ...The paragraph on microbiome (lines 447-468) ends on a somewhat negative note regarding methodological challenges

Thank you, we changed our conclusion as follows:

"In conclusion these studies suggest that there are associations between gut microbiome profiling and outcomes in patients with advanced ECs treated with ICIs, but these findings are preliminary and require replication in larger cohorts. Certainly, microbiota testing is a promising predictive assay in tumours candidate to I-O, but in EC it remains a research tool waiting for clinical validation."

 

 

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

1 The abstract and introduction state that the adoption of I-O "upon failure of platinum-based chemotherapy constitutes a decisive improvement". This is outdated and understates the current standard of care. ‌The pivotal message of the last 3-4 years is that dMMR/MSI-H is a predictive biomarker for the addition of I-O to first-line platinum/taxane chemotherapy‌, significantly improving PFS and OS. This should be emphasized as the central clinical advance.

2  Reframe the narrative to clearly state that for advanced/recurrent dMMR/MSI-H EC, the standard first-line treatment is now combination chemotherapy (carboplatin/paclitaxel) PLUS an immune checkpoint inhibitor (anti-PD-1 or anti-PD-L1), as established by multiple phase III trials.

3 The section "Are all dMMR endometrial cancers created equal?" (Lines 333-378) is valuable but somewhat conflicted and requires streamlining. It presents data suggesting worse clinicopathological features and potentially different immunogenicity (lower TILs, TMB) in MLH1-methylated vs. Lynch-associated tumors, but then cites clinical trial data showing no significant difference in PFS benefit from chemo-immunotherapy based on the mechanism of MMR loss.

4 Intrinsic biological differences:‌ Acknowledge that MLH1-methylated tumors may have distinct epidemiologic (older, more obese) and immunophenotypic profiles (e.g., lower T-cell counts) compared to mutation-driven dMMR.

5 Clinical/Therapeutic implications:‌ Clearly state that despite these biological nuances, ‌current evidence from randomized trials does not support using the mechanism of MMR loss (methylation vs. mutation) to select or withhold first-line chemo-immunotherapy.‌ The biomarker for this treatment is the dMMR/MSI-H status itself.

6 The section titled "The problem of resistance..." (Lines 390-415) correctly identifies a critical knowledge gap but lacks specificity. The statement that resistance "cannot be exactly quantifiable at the moment" is true for the distinction between primary and acquired resistance in the first-line chemo-IO setting, but the problem itself is quantifiable (e.g., a subset of patients in the dMMR arms of the cited trials still progress).

7 Provide context: While the objective response rates (ORR) for dMMR EC with first-line chemo-IO are high (~60-70%), a significant minority (30-40%) do not achieve a durable response.

Acknowledge that most available resistance data (e.g., on B2M, JAK1 mutations) come from studies of I-O monotherapy in later-line settings, and their applicability to the first-line chemo-IO combination is an active area of research.

Emphasize that identifying predictive biomarkers for this resistant subset is the "unmet need" that drives the subsequent sections of the review.

8 Table 3 (Mechanisms of resistance and associated biomarkers) is a useful summary. However, the accompanying text (Lines 420-550) can be more logically organized to mirror the table.
a) Impairment of antigen presentation:‌ The discussion on B2M and JAK1 mutations is good. Clarify that the study by Gulhan et al. (ref 71) analyzed patients treated with I-O monotherapy (avelumab). The note about B2M alterations not precluding response in colorectal cancer is interesting but may need a transitional phrase to explain its relevance to EC.
b) Tumor Mutational Burden and Deficiency in effector T-cells:‌ The sentence "TMB probably provides limited predictive value within a dMMR/MSI-H EC population..." (Lines 474-476) is a key insight and should be more prominently featured. The reference to "clonal neoantigen presentation" (ref 83, Westcott et al.) is excellent and could be briefly explained.
c) Immune cells activity in TME & d) Gene signatures:‌ These subsections contain valuable information but would benefit from a clearer introductory sentence stating that spatial analysis of the TME (immune contexture) and multi-parameter gene signatures are being explored to improve prediction over single biomarkers like TMB or PD-L1.

Author Response

Comments 1 and 2 The abstract and introduction state that the adoption of I-O "upon failure of platinum-based chemotherapy constitutes a decisive improvement". This is outdated and understates the current standard of care. The pivotal message of the last 3-4 years is that dMMR/MSI-H is a predictive biomarker for the addition of I-O to first-line platinum/taxane chemotherapy, significantly improving PFS and OS. This should be emphasized as the central clinical advance.

We thank the reviewer for this important comment. At line 23 of the Abstract and at line 100 of the Introduction we inserted the phrase: “Currently the combination of Immunotherapy with standard therapy has become the new standard first-line treatment for dMMR advanced or recurrent ECs”. In the introduction we clearly stated that the combined therapy is the first -line treatment in for these patients.

3 The section "Are all dMMR endometrial cancers created equal?" (Lines 333-378) is valuable but somewhat conflicted and requires streamlining. It presents data suggesting worse clinicopathological features and potentially different immunogenicity (lower TILs, TMB) in MLH1-methylated vs. Lynch-associated tumors, but then cites clinical trial data showing no significant difference in PFS benefit from chemo-immunotherapy based on the mechanism of MMR loss.

4 Intrinsic biological differences:‌ Acknowledge that MLH1-methylated tumors may have distinct epidemiologic (older, more obese) and immunophenotypic profiles (e.g., lower T-cell counts) compared to mutation-driven dMMR.

5 Clinical/Therapeutic implications:‌ Clearly state that despite these biological nuances, ‌current evidence from randomized trials does not support using the mechanism of MMR loss (methylation vs. mutation) to select or withhold first-line chemo-immunotherapy.‌ The biomarker for this treatment is the dMMR/MSI-H status itself.

We thank the reviewer for these insightful comments. The section has been revised to improve clarity. We now clearly distinguish between biological differences (epidemiologic and immunophenotypic features between MLH1-methylated and mutation-driven dMMR tumors) and their clinical implications. The revised text explicitly states that, despite these biological differences, current evidence from randomized trials does not support the use of the mechanism of MMR loss (methylation vs. mutation) to guide first-line chemo-immunotherapy decisions, and that dMMR/MSI-high status remains the relevant biomarker for treatment decision.

Comments 6 and 7

The section titled "The problem of resistance..." (Lines 390-415) correctly identifies a critical knowledge gap but lacks specificity. The statement that resistance "cannot be exactly quantifiable at the moment" is true for the distinction between primary and acquired resistance in the first-line chemo-IO setting, but the problem itself is quantifiable (e.g., a subset of patients in the dMMR arms of the cited trials still progress).

7 Provide context: While the objective response rates (ORR) for dMMR EC with first-line chemo-IO are high (~60-70%), a significant minority (30-40%) do not achieve a durable response.

Acknowledge that most available resistance data (e.g., on B2M, JAK1 mutations) come from studies of I-O monotherapy in later-line settings, and their applicability to the first-line chemo-IO combination is an active area of research.

Emphasize that identifying predictive biomarkers for this resistant subset is the "unmet need" that drives the subsequent sections of the review.

We thank the reviewer for these insightful suggestions. The section has been revised to improve clarity. We now state that the ORR in the dMMR subgroup treated with first-line chemo-immunotherapy is approximately 60-70%, indicating that about 30-40% of patients do not achieve a durable response. We also clarified that currently available trial reports provide only aggregate ORR and PFS/OS outcomes and do not allow a clear distinction between primary and acquired resistance, which would require IPD-level (Individual Participant Data) analyses. Finally, we added a sentence acknowledging that most currently described molecular mechanisms of resistance (e.g., alterations in immune-related genes) derive from ICI monotherapy studies in later-line settings, and that their relevance in the first-line chemo-immunotherapy context remains under investigation. The text has been modified accordingly (Lines 348-373).

Comments 8

Table 3 (Mechanisms of resistance and associated biomarkers) is a useful summary. However, the accompanying text (Lines 420-550) can be more logically organized to mirror the table.
a) Impairment of antigen presentation:‌ The discussion on B2M and JAK1 mutations is good. Clarify that the study by Gulhan et al. (ref 71) analyzed patients treated with I-O monotherapy (avelumab). The note about B2M alterations not precluding response in colorectal cancer is interesting but may need a transitional phrase to explain its relevance to EC.
b) Tumor Mutational Burden and Deficiency in effector T-cells:‌ The sentence "TMB probably provides limited predictive value within a dMMR/MSI-H EC population..." (Lines 474-476) is a key insight and should be more prominently featured. The reference to "clonal neoantigen presentation" (ref 83, Westcott et al.) is excellent and could be briefly explained.
c) Immune cells activity in TME & d) Gene signatures:‌ These subsections contain valuable information but would benefit from a clearer introductory sentence stating that spatial analysis of the TME (immune contexture) and multi-parameter gene signatures are being explored to improve prediction over single biomarkers like TMB or PD-L1.

Thank you to the Reviewer for his comments

To mirror the table 3 the subtitles of the paragraph 2.7 were modified

Moreover at line 388 we specified that the study analyzed patients treated with I-O monotherapy (avelumab)

Line 396 (B2M in colorectal cancer).

 In EC B2M mutations are a biologically important mechanism of resistance to checkpoint inhibitors, but B2M loss explains not all the non-responders and future studies need to clarify the role of gdTcells in EC.

Tumour Mutational Burden: Line 417 and following:

 However, only half of these patients take a significant clinical benefit from I-O and TMB doesn’t stratify responders. TMB does not reflect whether the mutations generate neoantigens or their immunogenicity, and the quality and clonality of neoepitopes might be the real important determinant of ICI response. Westcott et al. (83), showed that clonal neoantigen burden is predictive of ICI response in human MMRd cancers whereas overall tumor neoantigen burden is imperfectly correlated with clonal burden and ICI response. The intratumour heterogeneity of mutations or simply defined, the fraction of subclonal mutations could dilute the power of TMB as a biomarker of ICI response. Furthermore, several limitations prevent widespread implementation of TMB as a predictive biomarker, including high costs and lack of a defined cut-off.

c)Immune cells activity in TME & d) Gene signatures

We inserted an introductory sentence as required at line 467:

Spatial multiomics combines protein cell phenotyping with gene expression profiling, providing precise informations on cellular functions and disease mechanisms within the tumour and its TME. This combined approach makes it possible to identificy biomarkers in specific tissue compartments.