A new way to classify endometrial cancer

 

We classify endometrial cancer so that we can communicate and define each patient’s disease status, the potential for harm, and the likelihood that adjuvant therapies might provide help. Traditional forms of classification have clearly fallen short in achieving this aim, as we all know of patients with apparent low-risk disease (such as stage IA grade 1 endometrioid carcinoma) who have had recurrences and died from their disease, and we know that many patients have been subjected to overtreatment for their cancer and have acquired lifelong toxicities of therapy. This column will explore the newer, more sophisticated molecular-based classifications that are being validated for endometrial cancer, and the ways in which this promises to personalize the treatment of endometrial cancer.

We historically considered endometrial cancer with respect to “types”: type 1 cancer being estrogen dependent, featuring PTEN mutations, and affecting more obese patients; type 2 cancer being associated with p53 mutations, not estrogen dependent, and affecting older, less obese individuals.¹ These categories were reasonable guides but ultimately oversimplified the disease and its affected patients. Additionally we have used histologic types, International Federation of Gynecology and Obstetrics grading, and surgical staging to categorize tumors. Unfortunately, histologic cell type and grade are limited by poor agreement among pathologists, with up to 50% discordance between readers, and surgical staging information may be limited in its completeness.² Therefore, these categorizations lack the precision and accuracy to serve as prognosticators or to direct therapy. Reliance upon these inaccurate and imprecise methods of characterization may be part of the reason why most major clinical trials have failed to identify survival benefits for experimental therapies in early-stage disease. We may have been indiscriminately applying therapies instead of targeting the patients who are the most likely to derive benefit.

Breast cancer and melanoma are examples of the inclusion of molecular data such as hormone receptor status, HER2/neu status, or BRAF positivity resulting in advancements in personalizing therapeutics. We are now moving toward this for endometrial cancer.
 

What is the Cancer Genome Atlas?

In 2006 the National Institutes of Health announced an initiative to coordinate work between the National Cancer Institute and the National Human Genome Research Institute taking information about the human genome and analyzing it for key genomic alterations found in 33 common cancers. These data were combined with clinical information (such as survival) to classify the behaviors of those cancers with respect to their individual genomic alternations, in order to look for patterns in mutations and behaviors. The goal of this analysis was to shift the paradigm of cancer classification from being centered around primary organ site toward tumors’ shared genomic patterns.

In 2013 the Cancer Genome Atlas published their results of complete gene sequencing in endometrial cancer.³ The authors identified four discrete subgroups of endometrial cancer with distinct molecular mutational profiles and distinct clinical outcomes: polymerase epsilon (POLE, pronounced “pole-ee”) ultramutated, microsatellite instability (MSI) high, copy number high, and copy number low.
 

POLE ultramutated

An important subgroup identified in the Cancer Genome Atlas was a group of patients with a POLE ultramutated state. POLE encodes for a subunit of DNA polymerase, the enzyme responsible for replicating the leading DNA strand. Nonfunctioning POLE results in proofreading errors and a subsequent ultramutated cellular state with a predominance of single nucleotide variants. POLE proofreading domain mutations in endometrial cancer and colon cancer are associated with excellent prognosis, likely secondary to the immune response that is elicited by this ultramutated state from creation of “antigenic neoepitopes” that stimulate T-cell response. Effectively, the very mutated cell is seen as “more foreign” to the body’s immune system.

Approximately 10% of patients with endometrial cancer have a POLE ultramutated state, and, as stated above, prognosis is excellent, even if coexisting with a histologic cell type (such as serous) that is normally associated with adverse outcomes. These women tend to be younger, with a lower body mass index, higher-grade endometrioid cell type, the presence of lymphovascular space invasion, and low stage.
 

MSI high

MSI (microsatellite instability) is a result of epigenetic/hypermethylations or loss of expression in mismatch repair genes (such as MLH1, MSH2, MSH6, PMS2). These genes code for proteins critical in the repair of mismatches in short repeated sequences of DNA. Loss of their function results in an accumulation of errors in these sequences: MSI. It is a feature of the Lynch syndrome inherited state, but is also found sporadically in endometrial tumors. These tumors accumulate a number of mutations during cell replication that, as in POLE hypermutated tumors, are associated with eliciting an immune response.

 

These tumors tend to be associated with a higher-grade endometrioid cell type, the presence of lymphovascular space invasion, and an advanced stage. Patients with tumors that have been described as MSI high are candidates for “immune therapy” with the PDL1 inhibitor pembrolizumab because of their proinflammatory state and observed favorable responses in clinical trials.⁴
 

Copy number high/low

Copy number (CN) high and low refers to the results of microarrays in which hierarchical clustering was applied to identify reoccurring amplification or deletion regions. The CN-high group was associated with the poorest outcomes (recurrence and survival). There is significant overlap with mutations in TP53. Most serous carcinomas were CN high; however, 25% of patients with high-grade endometrioid cell type shared the CN-high classification. These tumors shared great molecular similarity to high-grade serous ovarian cancers and basal-like breast cancer.

Those patients who did not possess mutations that classified them as POLE hypermutated, MSI high, or CN high were classified as CN low. This group included predominantly grades 1 and 2 endometrioid adenocarcinomas of an early stage and had a favorable prognostic profile, though less favorable than those with a POLE ultramutated state, which appears to be somewhat protective.
 

Molecular/metabolic interactions

While molecular data are clearly important in driving a cancer cell’s behavior, other clinical and metabolic factors influence cancer behavior. For example, body mass index, adiposity, glucose, and lipid metabolism have been shown to be important drivers of cellular behavior and responsiveness to targeted therapies.^5,6 Additionally age, race, and other metabolic states contribute to oncologic behavior. Future classifications of endometrial cancer are unlikely to use molecular profiles in isolation but will need to incorporate these additional patient-specific data to better predict and prognosticate outcomes.

Clinical applications

If researchers can better define and describe a patient’s endometrial cancer from the time of their biopsy, important clinical decisions might be able to be tackled. For example, in a premenopausal patient with an endometrial cancer who is considering fertility-sparing treatments, preoperative knowledge of a POLE ultramutated state (and therefore an anticipated good prognosis) might favor fertility preservation or avoid comprehensive staging which may be of limited value. Similarly, if an MSI-high profile is identified leading to a Lynch syndrome diagnosis, she may be more inclined to undergo a hysterectomy with bilateral salpingo-oophorectomy and staging as she is at known increased risk for a more advanced endometrial cancer, as well as the potential for ovarian cancer.

Postoperative incorporation of molecular data promises to be particularly helpful in guiding adjuvant therapies and sparing some women from unnecessary treatments. For example, women with high-grade endometrioid tumors who are CN high were historically treated with radiotherapy but might do better treated with systemic adjuvant therapies traditionally reserved for nonendometrioid carcinomas. Costly therapies such as immunotherapy can be directed toward those with MSI-high tumors, and the rare patient with a POLE ultramutated state who has a recurrence or advanced disease. Clinical trials will be able to cluster enrollment of patients with CN-high, serouslike cancers with those with serous cancers, rather than combining them with patients whose cancers predictably behave much differently.

Much work is still needed to validate this molecular profiling in endometrial cancer and define the algorithms associated with treatment decisions; however, it is likely that the way we describe endometrial cancer in the near future will be quite different.
 

Dr. Rossi is an assistant professor in the division of gynecologic oncology at the University of North Carolina at Chapel Hill. She has no disclosures.

References

1. Bokhman JV. Two pathogenetic types of endometrial carcinoma. Gynecol Oncol. 1983;15(1):10-7.

2. Clarke BA et al. Endometrial carcinoma: controversies in histopathological assessment of grade and tumour cell type. J Clin Pathol. 2010;63(5):410-5.

3. Cancer Genome Atlas Research Network. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497(7447):67-73.

4. Ott PA et al. Pembrolizumab in advanced endometrial cancer: Preliminary results from the phase Ib KEYNOTE-028 study. J Clin Oncol. 2016;34(suppl):Abstract 5581.

5. Roque DR et al. Association between differential gene expression and body mass index among endometrial cancers from the Cancer Genome Atlas Project. Gynecol Oncol. 2016;142(2):317-22.

6. Talhouk A et al. New classification of endometrial cancers: The development and potential applications of genomic-based classification in research and clinical care. Gynecol Oncol Res Pract. 2016 Dec;3:14.

 

We classify endometrial cancer so that we can communicate and define each patient’s disease status, the potential for harm, and the likelihood that adjuvant therapies might provide help. Traditional forms of classification have clearly fallen short in achieving this aim, as we all know of patients with apparent low-risk disease (such as stage IA grade 1 endometrioid carcinoma) who have had recurrences and died from their disease, and we know that many patients have been subjected to overtreatment for their cancer and have acquired lifelong toxicities of therapy. This column will explore the newer, more sophisticated molecular-based classifications that are being validated for endometrial cancer, and the ways in which this promises to personalize the treatment of endometrial cancer.

We historically considered endometrial cancer with respect to “types”: type 1 cancer being estrogen dependent, featuring PTEN mutations, and affecting more obese patients; type 2 cancer being associated with p53 mutations, not estrogen dependent, and affecting older, less obese individuals.¹ These categories were reasonable guides but ultimately oversimplified the disease and its affected patients. Additionally we have used histologic types, International Federation of Gynecology and Obstetrics grading, and surgical staging to categorize tumors. Unfortunately, histologic cell type and grade are limited by poor agreement among pathologists, with up to 50% discordance between readers, and surgical staging information may be limited in its completeness.² Therefore, these categorizations lack the precision and accuracy to serve as prognosticators or to direct therapy. Reliance upon these inaccurate and imprecise methods of characterization may be part of the reason why most major clinical trials have failed to identify survival benefits for experimental therapies in early-stage disease. We may have been indiscriminately applying therapies instead of targeting the patients who are the most likely to derive benefit.

Breast cancer and melanoma are examples of the inclusion of molecular data such as hormone receptor status, HER2/neu status, or BRAF positivity resulting in advancements in personalizing therapeutics. We are now moving toward this for endometrial cancer.
 

What is the Cancer Genome Atlas?

In 2006 the National Institutes of Health announced an initiative to coordinate work between the National Cancer Institute and the National Human Genome Research Institute taking information about the human genome and analyzing it for key genomic alterations found in 33 common cancers. These data were combined with clinical information (such as survival) to classify the behaviors of those cancers with respect to their individual genomic alternations, in order to look for patterns in mutations and behaviors. The goal of this analysis was to shift the paradigm of cancer classification from being centered around primary organ site toward tumors’ shared genomic patterns.

In 2013 the Cancer Genome Atlas published their results of complete gene sequencing in endometrial cancer.³ The authors identified four discrete subgroups of endometrial cancer with distinct molecular mutational profiles and distinct clinical outcomes: polymerase epsilon (POLE, pronounced “pole-ee”) ultramutated, microsatellite instability (MSI) high, copy number high, and copy number low.
 

POLE ultramutated

An important subgroup identified in the Cancer Genome Atlas was a group of patients with a POLE ultramutated state. POLE encodes for a subunit of DNA polymerase, the enzyme responsible for replicating the leading DNA strand. Nonfunctioning POLE results in proofreading errors and a subsequent ultramutated cellular state with a predominance of single nucleotide variants. POLE proofreading domain mutations in endometrial cancer and colon cancer are associated with excellent prognosis, likely secondary to the immune response that is elicited by this ultramutated state from creation of “antigenic neoepitopes” that stimulate T-cell response. Effectively, the very mutated cell is seen as “more foreign” to the body’s immune system.

Approximately 10% of patients with endometrial cancer have a POLE ultramutated state, and, as stated above, prognosis is excellent, even if coexisting with a histologic cell type (such as serous) that is normally associated with adverse outcomes. These women tend to be younger, with a lower body mass index, higher-grade endometrioid cell type, the presence of lymphovascular space invasion, and low stage.
 

MSI high

MSI (microsatellite instability) is a result of epigenetic/hypermethylations or loss of expression in mismatch repair genes (such as MLH1, MSH2, MSH6, PMS2). These genes code for proteins critical in the repair of mismatches in short repeated sequences of DNA. Loss of their function results in an accumulation of errors in these sequences: MSI. It is a feature of the Lynch syndrome inherited state, but is also found sporadically in endometrial tumors. These tumors accumulate a number of mutations during cell replication that, as in POLE hypermutated tumors, are associated with eliciting an immune response.

 

These tumors tend to be associated with a higher-grade endometrioid cell type, the presence of lymphovascular space invasion, and an advanced stage. Patients with tumors that have been described as MSI high are candidates for “immune therapy” with the PDL1 inhibitor pembrolizumab because of their proinflammatory state and observed favorable responses in clinical trials.⁴
 

Copy number high/low

Copy number (CN) high and low refers to the results of microarrays in which hierarchical clustering was applied to identify reoccurring amplification or deletion regions. The CN-high group was associated with the poorest outcomes (recurrence and survival). There is significant overlap with mutations in TP53. Most serous carcinomas were CN high; however, 25% of patients with high-grade endometrioid cell type shared the CN-high classification. These tumors shared great molecular similarity to high-grade serous ovarian cancers and basal-like breast cancer.

Those patients who did not possess mutations that classified them as POLE hypermutated, MSI high, or CN high were classified as CN low. This group included predominantly grades 1 and 2 endometrioid adenocarcinomas of an early stage and had a favorable prognostic profile, though less favorable than those with a POLE ultramutated state, which appears to be somewhat protective.
 

Molecular/metabolic interactions

While molecular data are clearly important in driving a cancer cell’s behavior, other clinical and metabolic factors influence cancer behavior. For example, body mass index, adiposity, glucose, and lipid metabolism have been shown to be important drivers of cellular behavior and responsiveness to targeted therapies.^5,6 Additionally age, race, and other metabolic states contribute to oncologic behavior. Future classifications of endometrial cancer are unlikely to use molecular profiles in isolation but will need to incorporate these additional patient-specific data to better predict and prognosticate outcomes.

Clinical applications

If researchers can better define and describe a patient’s endometrial cancer from the time of their biopsy, important clinical decisions might be able to be tackled. For example, in a premenopausal patient with an endometrial cancer who is considering fertility-sparing treatments, preoperative knowledge of a POLE ultramutated state (and therefore an anticipated good prognosis) might favor fertility preservation or avoid comprehensive staging which may be of limited value. Similarly, if an MSI-high profile is identified leading to a Lynch syndrome diagnosis, she may be more inclined to undergo a hysterectomy with bilateral salpingo-oophorectomy and staging as she is at known increased risk for a more advanced endometrial cancer, as well as the potential for ovarian cancer.

Postoperative incorporation of molecular data promises to be particularly helpful in guiding adjuvant therapies and sparing some women from unnecessary treatments. For example, women with high-grade endometrioid tumors who are CN high were historically treated with radiotherapy but might do better treated with systemic adjuvant therapies traditionally reserved for nonendometrioid carcinomas. Costly therapies such as immunotherapy can be directed toward those with MSI-high tumors, and the rare patient with a POLE ultramutated state who has a recurrence or advanced disease. Clinical trials will be able to cluster enrollment of patients with CN-high, serouslike cancers with those with serous cancers, rather than combining them with patients whose cancers predictably behave much differently.

Much work is still needed to validate this molecular profiling in endometrial cancer and define the algorithms associated with treatment decisions; however, it is likely that the way we describe endometrial cancer in the near future will be quite different.
 

Dr. Rossi is an assistant professor in the division of gynecologic oncology at the University of North Carolina at Chapel Hill. She has no disclosures.

References

1. Bokhman JV. Two pathogenetic types of endometrial carcinoma. Gynecol Oncol. 1983;15(1):10-7.

2. Clarke BA et al. Endometrial carcinoma: controversies in histopathological assessment of grade and tumour cell type. J Clin Pathol. 2010;63(5):410-5.

3. Cancer Genome Atlas Research Network. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497(7447):67-73.

4. Ott PA et al. Pembrolizumab in advanced endometrial cancer: Preliminary results from the phase Ib KEYNOTE-028 study. J Clin Oncol. 2016;34(suppl):Abstract 5581.

5. Roque DR et al. Association between differential gene expression and body mass index among endometrial cancers from the Cancer Genome Atlas Project. Gynecol Oncol. 2016;142(2):317-22.

6. Talhouk A et al. New classification of endometrial cancers: The development and potential applications of genomic-based classification in research and clinical care. Gynecol Oncol Res Pract. 2016 Dec;3:14.

 

We classify endometrial cancer so that we can communicate and define each patient’s disease status, the potential for harm, and the likelihood that adjuvant therapies might provide help. Traditional forms of classification have clearly fallen short in achieving this aim, as we all know of patients with apparent low-risk disease (such as stage IA grade 1 endometrioid carcinoma) who have had recurrences and died from their disease, and we know that many patients have been subjected to overtreatment for their cancer and have acquired lifelong toxicities of therapy. This column will explore the newer, more sophisticated molecular-based classifications that are being validated for endometrial cancer, and the ways in which this promises to personalize the treatment of endometrial cancer.

We historically considered endometrial cancer with respect to “types”: type 1 cancer being estrogen dependent, featuring PTEN mutations, and affecting more obese patients; type 2 cancer being associated with p53 mutations, not estrogen dependent, and affecting older, less obese individuals.¹ These categories were reasonable guides but ultimately oversimplified the disease and its affected patients. Additionally we have used histologic types, International Federation of Gynecology and Obstetrics grading, and surgical staging to categorize tumors. Unfortunately, histologic cell type and grade are limited by poor agreement among pathologists, with up to 50% discordance between readers, and surgical staging information may be limited in its completeness.² Therefore, these categorizations lack the precision and accuracy to serve as prognosticators or to direct therapy. Reliance upon these inaccurate and imprecise methods of characterization may be part of the reason why most major clinical trials have failed to identify survival benefits for experimental therapies in early-stage disease. We may have been indiscriminately applying therapies instead of targeting the patients who are the most likely to derive benefit.

Breast cancer and melanoma are examples of the inclusion of molecular data such as hormone receptor status, HER2/neu status, or BRAF positivity resulting in advancements in personalizing therapeutics. We are now moving toward this for endometrial cancer.
 

What is the Cancer Genome Atlas?

In 2006 the National Institutes of Health announced an initiative to coordinate work between the National Cancer Institute and the National Human Genome Research Institute taking information about the human genome and analyzing it for key genomic alterations found in 33 common cancers. These data were combined with clinical information (such as survival) to classify the behaviors of those cancers with respect to their individual genomic alternations, in order to look for patterns in mutations and behaviors. The goal of this analysis was to shift the paradigm of cancer classification from being centered around primary organ site toward tumors’ shared genomic patterns.

In 2013 the Cancer Genome Atlas published their results of complete gene sequencing in endometrial cancer.³ The authors identified four discrete subgroups of endometrial cancer with distinct molecular mutational profiles and distinct clinical outcomes: polymerase epsilon (POLE, pronounced “pole-ee”) ultramutated, microsatellite instability (MSI) high, copy number high, and copy number low.
 

POLE ultramutated

An important subgroup identified in the Cancer Genome Atlas was a group of patients with a POLE ultramutated state. POLE encodes for a subunit of DNA polymerase, the enzyme responsible for replicating the leading DNA strand. Nonfunctioning POLE results in proofreading errors and a subsequent ultramutated cellular state with a predominance of single nucleotide variants. POLE proofreading domain mutations in endometrial cancer and colon cancer are associated with excellent prognosis, likely secondary to the immune response that is elicited by this ultramutated state from creation of “antigenic neoepitopes” that stimulate T-cell response. Effectively, the very mutated cell is seen as “more foreign” to the body’s immune system.

Approximately 10% of patients with endometrial cancer have a POLE ultramutated state, and, as stated above, prognosis is excellent, even if coexisting with a histologic cell type (such as serous) that is normally associated with adverse outcomes. These women tend to be younger, with a lower body mass index, higher-grade endometrioid cell type, the presence of lymphovascular space invasion, and low stage.
 

MSI high

MSI (microsatellite instability) is a result of epigenetic/hypermethylations or loss of expression in mismatch repair genes (such as MLH1, MSH2, MSH6, PMS2). These genes code for proteins critical in the repair of mismatches in short repeated sequences of DNA. Loss of their function results in an accumulation of errors in these sequences: MSI. It is a feature of the Lynch syndrome inherited state, but is also found sporadically in endometrial tumors. These tumors accumulate a number of mutations during cell replication that, as in POLE hypermutated tumors, are associated with eliciting an immune response.

 

These tumors tend to be associated with a higher-grade endometrioid cell type, the presence of lymphovascular space invasion, and an advanced stage. Patients with tumors that have been described as MSI high are candidates for “immune therapy” with the PDL1 inhibitor pembrolizumab because of their proinflammatory state and observed favorable responses in clinical trials.⁴
 

Copy number high/low

Copy number (CN) high and low refers to the results of microarrays in which hierarchical clustering was applied to identify reoccurring amplification or deletion regions. The CN-high group was associated with the poorest outcomes (recurrence and survival). There is significant overlap with mutations in TP53. Most serous carcinomas were CN high; however, 25% of patients with high-grade endometrioid cell type shared the CN-high classification. These tumors shared great molecular similarity to high-grade serous ovarian cancers and basal-like breast cancer.

Those patients who did not possess mutations that classified them as POLE hypermutated, MSI high, or CN high were classified as CN low. This group included predominantly grades 1 and 2 endometrioid adenocarcinomas of an early stage and had a favorable prognostic profile, though less favorable than those with a POLE ultramutated state, which appears to be somewhat protective.
 

Molecular/metabolic interactions

While molecular data are clearly important in driving a cancer cell’s behavior, other clinical and metabolic factors influence cancer behavior. For example, body mass index, adiposity, glucose, and lipid metabolism have been shown to be important drivers of cellular behavior and responsiveness to targeted therapies.^5,6 Additionally age, race, and other metabolic states contribute to oncologic behavior. Future classifications of endometrial cancer are unlikely to use molecular profiles in isolation but will need to incorporate these additional patient-specific data to better predict and prognosticate outcomes.

Clinical applications

If researchers can better define and describe a patient’s endometrial cancer from the time of their biopsy, important clinical decisions might be able to be tackled. For example, in a premenopausal patient with an endometrial cancer who is considering fertility-sparing treatments, preoperative knowledge of a POLE ultramutated state (and therefore an anticipated good prognosis) might favor fertility preservation or avoid comprehensive staging which may be of limited value. Similarly, if an MSI-high profile is identified leading to a Lynch syndrome diagnosis, she may be more inclined to undergo a hysterectomy with bilateral salpingo-oophorectomy and staging as she is at known increased risk for a more advanced endometrial cancer, as well as the potential for ovarian cancer.

Postoperative incorporation of molecular data promises to be particularly helpful in guiding adjuvant therapies and sparing some women from unnecessary treatments. For example, women with high-grade endometrioid tumors who are CN high were historically treated with radiotherapy but might do better treated with systemic adjuvant therapies traditionally reserved for nonendometrioid carcinomas. Costly therapies such as immunotherapy can be directed toward those with MSI-high tumors, and the rare patient with a POLE ultramutated state who has a recurrence or advanced disease. Clinical trials will be able to cluster enrollment of patients with CN-high, serouslike cancers with those with serous cancers, rather than combining them with patients whose cancers predictably behave much differently.

Much work is still needed to validate this molecular profiling in endometrial cancer and define the algorithms associated with treatment decisions; however, it is likely that the way we describe endometrial cancer in the near future will be quite different.
 

Dr. Rossi is an assistant professor in the division of gynecologic oncology at the University of North Carolina at Chapel Hill. She has no disclosures.

References

1. Bokhman JV. Two pathogenetic types of endometrial carcinoma. Gynecol Oncol. 1983;15(1):10-7.

2. Clarke BA et al. Endometrial carcinoma: controversies in histopathological assessment of grade and tumour cell type. J Clin Pathol. 2010;63(5):410-5.

3. Cancer Genome Atlas Research Network. Integrated genomic characterization of endometrial carcinoma. Nature. 2013;497(7447):67-73.

4. Ott PA et al. Pembrolizumab in advanced endometrial cancer: Preliminary results from the phase Ib KEYNOTE-028 study. J Clin Oncol. 2016;34(suppl):Abstract 5581.

5. Roque DR et al. Association between differential gene expression and body mass index among endometrial cancers from the Cancer Genome Atlas Project. Gynecol Oncol. 2016;142(2):317-22.

6. Talhouk A et al. New classification of endometrial cancers: The development and potential applications of genomic-based classification in research and clinical care. Gynecol Oncol Res Pract. 2016 Dec;3:14.