Casey M. Cosgrove, MD, discussed the current state of biomarkers in the realm of endometrial cancer and how they may be optimally leveraged for prognostic and predictive purposes.
Biomarkers such as mismatch repair deficiency (dMMR), microsatellite instability (MSI), p53 positivity, and POLE mutations possess the potential to inform treatment decisions in patients with endometrial cancer, but several questions still need to be addressed, according to Casey M. Cosgrove, MD.
For example, Cosgrove added that dMMR or MSI status and increased tumor mutational burden are more clearly defined for use in clinical practice, but it is still unclear whether a patient with a tumor that harbors aPOLE mutation can safely forego treatment.
“[We] must continue to integrate the right biomarkers that we have information for into clinical practice,” Cosgrove said. “Certainly, it is easy to order a lot of testing, but if we do not have a clear game plan with what we are going to do with that testing, then we [could] harm our patients. [We need to practice some] caution when we are discussing biomarkers and new technologies with any cancer types.”
In an interview with OncLive® during the 2022 SGO Winter Meeting, Cosgrove, an assistant professor in the Department of Gynecologic Oncology at The Ohio State University College of Medicine, discussed the current state of biomarkers in the realm of endometrial cancer and how they may be optimally leveraged for prognostic and predictive purposes.
Cosgrove: We know that with many different types of cancers, we’re starting to look at a more specialized approach as to how we can better take care of our patients, and endometrial cancer certainly has many opportunities where we can identify molecular markers or things that we can’t see under the microscope to help guide the best way to treat these patients.
We can use biomarkers to treat patients, to guide which trial metric to use, [or] which type of treatment we want to use. But we can also use biomarkers to tell us which patients should be treated, or maybe even which patients should not be treated. It’s an incredibly exciting area of research for endometrial cancer because it allows us to home in on their individual types of cancers and treat them with the most individualized approach.
One of the things that needs to be highlighted about endometrial cancer is it's one of the few cancers that is getting worse. We have more patients dying of endometrial cancer now than we did last year, and that's a travesty. Endometrial cancer is aligning itself to overtake ovarian cancer to become the deadliest gynecologic cancer in the United States in the near future. We need our researchers and clinicians to be aligned with what patients that are having poor outcomes are seeing.
With biomarkers in endometrial cancer, there are 2 fundamental questions. The first question is, ‘Who should we be treating?’ Several studies have looked at the best ways that we can predict whether patients' cancers will come back or not.
Classically, we have been guided by [disease] stages or the histology, as well as several other features. However, now, we are starting to appreciate that we can use molecular signatures to tell us whether these cancers are at higher or lower risk for recurrence.
There have been reports from our group, as well as others, that have looked at different markers that have predicted higher-risk cancers; this can include p53 or genomic alterations, and several other mutations that we have noted. If we identify patients who have some of those higher-risk features, they may derive benefit from receiving extra treatment. Another marker, [referred to as] POLE, is needed to help the DNA repair itself; this allows for the immune system, without any additional treatment, to fight cancer cells better. We see that those patients [with POLE] tend to not recur.
A question that we have also been asking is, ‘Can we not treat these patients, even if they traditionally would have had higher-risk features?’ [Specifically regarding] biomarkers, [we want [to see whether] we [can] work smarter, and not necessarily harder. Can we target therapies that are based off molecular signatures?
Perhaps the one that we have the most research and the most experience with is deficiencies in MMR proteins or MSI within endometrial cancers; these markers predict response for immunotherapies. As such, when we look at an endometrial cancer that has 1 of these markers, perhaps we should jump right to immunotherapy as opposed to going through our classical chemotherapy algorithms. We also know that MMR and MSI status can also be an important prognostic biomarker because these patients seem to also have poor outcomes.
As such, taken collectively, a patient that has a dMMR tumor or MSI-high tumor may benefit from receiving more aggressive treatment and a [potentially] smarter [approach] with immunotherapy as opposed to traditional chemotherapy.
Several ongoing clinical trials are trying to refine the way that we take care of a patient with endometrial cancer. What I hope that we can see through all these trials is that we have some easily applied clinical markers that are going to directly impact our decision making. The ProMisE [Proactive Molecular Risk Classifier for Endometrial Cancer system] from the University of British Columbia, which establishes endometrial cancer into 4 different molecular groups, has the most traction right now throughout the gynecologic oncology community.1 This is testing for dMMR, POLE [mutations], as well as p53 [positivity], and grouping endometrial cancers into 1 of 4 groups.
Depending on which group [patients] fall into, then the next question will be, ‘What is the best treatment for those individuals?’ This has been evaluated through several trials. There is a lot of excitement [in this area] and we hope to get some answers over the next couple of years. The endometrial cancer [setting] has a lot of opportunity for growth in patient care, and we, as clinicians, are going to have to start integrating a lot of this information quickly as it becomes available.
With the improvements in technology, we are going to be getting more information for each cancer type in the future than we have ever had before. Assimilating that information into the clinical setting [may represent] a major barrier for patients and clinicians alike. [That being said], hopefully, we will have some standardized protocols that [can help] us to [select] the [best] way to care for each individual patient.
The RAINBO [trials] are very exciting because not only are [the investigators] using the molecular classifications to determine the risk for recurrence and the aggressiveness of treatment, but they are also integrating novel treatment strategies for the different molecular groups. As suh, I might identify that a patient is at higher risk, and I know what [treatments are] available, but I can [consider adding] more targeted therapies because of that molecular class, and I can treat them more aggressively and with a smarter therapy.
One of the things that is most notable is the vast amount of ongoing research that we have in the area right now; this also highlights the vast number of [unanswered] questions right now. One potential concern is that physicians may start integrating little pieces of this information into their clinical practice, without having a full tool belt available for what they need to be doing [with respect to] what are the right tests [to use] and the right treatments that might be available.
Within the past 1 or 2 years, the NCCN guidelines have started incorporating the algorithms for molecular classification with endometrial cancer.2 Traditionally, we have broken endometrial cancers into type 1 and type 2, with type 1 being a less aggressive type of [disease], and type 2 being more aggressive. In 2013, the Cancer Genome Atlas [TCGA] broke endometrial cancer [down further,] into 4 important groups with differing risk of recurrence and differing molecular features.3 Recent research has tried to integrate those in the clinical setting because TCGA uses technologies that we cannot apply to the bedside. By utilizing newer technologies, we are able to start integrating some of their findings into clinical care.
The NCCN has started to appreciate that, and [currently uses] the algorithm for breaking endometrial cancers into 1 of 4 molecular groups, as opposed to the 2 groups that we had traditionally [used]. We also see a greater emphasis on what we can do with that information in the NCCN guidelines. For instance, 1 of the most important molecular groups is the MSI or dMMR group. Not only is this an important prognostic biomarker, but it is also a predictive biomarker for the use of immunotherapy—that has been highlighted in the more recent editions of the NCCN guidelines, where they say if [a patient] recurs [and has] MSI-high disease or dMMR, then immunotherapy is very appropriate.
The other thing that the NCCN guidelines highlight is that even in the absence of having a biomarker, patients can receive immunotherapy with the combination of lenvatinib [Lenvima] and pembrolizumab [Keytruda]. As such, sometimes having a negative biomarker can be just as informative for guiding treatment.
One of the most important concepts to understand is that we do not have all the answers quite yet; we are still working on it. As we get more information, [we may identify other biomarkers] to look for. It is hard to cherry pick 1 or 2 things that [we] can utilize [for all patients]. The question is whether we need to have a whole grouping of testing to make the right decision for our patients.
Right now, we are [in] a situation where there are clearly defined biomarkers that can be utilized in a clinical setting, like dMMR or MSI status and increased tumor mutational burden, but we still have unclear knowledge whether a patient with a tumor with, [for instance], a POLE mutation, can safely forego treatment. If you hear that these tumors have no recurrences, [we want to know] why we are treating them with chemotherapy and radiation, which have high toxicities. However, at the same time, we have not clearly proven that withholding treatment [in these patients] is not compromising outcomes.
[It is also important to note] our reliance on classic pathologic findings, histology, and stage, and [reflect on] how we integrate that into our patient care algorithms. Once we start getting more molecular data, how do we integrate the old and new data? What do we do if we have conflicting information? This could cause problems in the future if we do not adequately study these questions.