Robin (00:01):

Hello. Welcome to today's installment of “Diagnostics in Practice,” where we discuss the latest diagnostic offerings from Mayo Clinic Laboratories. I'm Robin Huiras, a senior marketing specialist with Mayo Clinic Labs and firm believer in the power of advanced testing to change patients' lives. I'm really happy to be here today with two guests to discuss Mayo Clinic Laboratories’ newest neuro-oncology test. The Neuro-Oncology Genome-wide Methylation Array Analysis, which goes by the test code MTNON, is an innovative assay that provides the methylation status of central nervous system or CNS tumors. Joining me for today's conversation is Dr. Christiane Ida, molecular pathologist and neuropathologist at Mayo Clinic and consultant with Mayo Clinic's Genetics and Genomics Lab, and Dr. Kenneth Aldape, a neuropathologist and senior associate consultant with the Genetics and Genomics Lab at Mayo Clinic. Thank you both so much for being here today.

Cris Ida (00:59):

Thank you. Good to be here.

Kenneth Aldape (01:01):

Thanks for having me.

Robin (01:02):

Now, before we dig into today's discussion, I'm hoping you both can provide our listeners with just a little bit of background on your experience with molecular testing at Mayo Clinic. Dr. Ida, can you go first?

Cris Ida (01:14):

Sure. I have had the privilege of working in the Genetics and Genomics Laboratory for the past 10 years, and I continue to be inspired by the exceptional multidisciplinary teams that make our clinical test development possible. Developing a clinical assay is a complex, demanding process that requires careful design, rigorous analytical validation, and integration across laboratory systems. And the work doesn't end at launch. Ongoing quality assurance and quality control are essential to ensure continued accuracy and reliability. At Mayo Clinic, we are deeply committed to excellence in molecular testing. We invest the time, expertise, and resources necessary to deliver high-quality, cost-effective, and clinically useful tests for the patients we serve.

Kenneth Aldape (02:07):

I joined Mayo Clinic several months ago, largely on the basis of its commitment to high-level molecular diagnostics of cancer, especially for central nervous system tumors. Most recently, I was the head of the Department of Pathology at the National Cancer Institute in Maryland, where I developed a program on molecular diagnostics of central nervous system tumors, including diagnostic methylation profiling. While I was there, I created the NCI Bethesda Methylation Classifier, which is the same methylation classifier that will be used here at the Mayo Clinic. I'm excited to continue my involvement with this program here at the Mayo Clinic.

Robin (02:45):

Thanks so much, both of you, for sharing a little bit about yourselves. Now, let's get started. Methylation profiling has emerged as a powerful molecular testing tool to clarify diagnosis and classify CNS tumors, and methylation has been recognized by the World Health Organization as a key molecular biomarker to help classify these types of cancers. Dr. Ida, perhaps you can get us started by providing a little bit of background on what methylation status is in the context of tumor profiling, and why it's important or helpful to know the methylation profile of a tumor.

Cris Ida (03:18):

Yes, happy to. DNA methylation is an epigenetic modification, meaning it acts on top of the genetic code without altering the DNA sequence. It consists of addition of chemical methyl marks on the DNA across the genome. Methylation plays a key role in the regulation of gene expression. It is important for normal development through, for example, its effect on gene in printing, X-chromosome activation, and transcriptional silencing of repetitive elements. In cancer, patterns of DNA methylation are disrupted and are a result of a combination of cell of origin and somatically acquired changes that are stable over time. These abnormal patterns generate a unique DNA methylation profile functioning as a methylation fingerprint or signature of the cancer. As such, DNA methylation profile has emerged as a molecular diagnostic biomarker for CNS tumors. In fact, DNA methylation profile has been listed as either essential or desirable diagnostic criteria for over 50% of CNS tumor types included in the 2021 Fifth WHO classification.

Robin (04:36):

Thanks for that quick intro into methylation profiling, Dr. Ida. Let's talk a little bit more about the advantages of methylation testing and clarifying diagnosis or even prognosis of CNS tumors.

Cris Ida (04:48):

There are more than 80 types of CNS tumors, and many tumors that look identical by microscopic evaluation may differ in clinical behavior. Similarly, morphologically distinct tumors may share a common molecular basis. DNA methylation profiling has revealed previously unrecognized tumor subgroups within established categories, refined distinctions between overlapping entities, and even identified entirely new tumor types. Thus, it has been shown to be a powerful discovery tool that has refined CNS tumor classification. It has also been shown to be a robust diagnostic tool through AI machine learning-based classifiers leveraging the various unique DNA methylation signatures. As a result, methylation testing strengthens diagnostic accuracy, which directly impacts patient management.

Robin (05:45):

Thanks so much for that great explanation of the value of methylation testing. It sounds like there are just so many benefits to knowing the methylation status of these tumors. Dr. Aldape, do you have anything to add to what Dr. Ida has said?

Kenneth Aldape (05:58):

I really do think that methylation testing is about improving the diagnostic accuracy for brain tumors. And there is a great need for this because conventional means of how we diagnose tumors sometimes don't lead to a specific diagnosis. Additionally, we find that there are surprises in diagnostics that can only be revealed by in-depth methylation profiling. It can be hard for many pathologists to keep track of all the 80-plus types of brain tumors straight. Some are common, but some are uncommon. An individual pathologist might see only a few examples of such rare cases over the entirety of their career, so it's hard to keep them straight when you don't have the experience to see everything. And there's also inter-observer interpretation. Some folks have more experience than others in interpreting the test results, and mistakes can be made. When diagnostic errors are made, there can be clinical consequences.

The patient might be given a diagnosis of something that gets them a certain type of treatment that they may not have needed, or they may be overtreated or, conversely, they may be undertreated, depending on the diagnosis. What methylation testing does is it gives you the right diagnosis, and for pathologists who might not see that many brain tumor cases in their career, or for difficult cases, they can just get the test so there's no ambiguity.

Robin (07:18):

Thanks, Dr. Aldape, for those insights. It all makes so much sense to me. I'm still curious, though, because Mayo Clinic Labs offers a number of other molecular testing options, including DNA sequencing, chromosomal microarray, and then, a recently launched whole transcriptome RNA sequencing panel. Why is methylation testing needed? Dr. Ida, can you take this one?

Cris Ida (07:40):

Sure. Methylation is one of the four molecular diagnostic biomarkers in the integrated histomolecular diagnosis recommended by the WHO guidelines. The other biomarkers are mutations, fusions, and copy number variants. TMB and MSI are also important biomarkers in cancer, but neither diagnostic nor common in CNS tumors. Next-generation sequencing tests for mutations, chromosome microarray for copy number variants, and whole transcriptome RNA sequencing for fusions. None of these testing strategies provide tumor DNA methylation profile, so we need a different platform. Also, one may perform testing for mutations, copy number variants, and fusions, and still not get a diagnosis, or can only definitively diagnose a tumor based on the DNA methylation profile. For instance, high-grade astrocytoma with piloid features is one example of a new tumor type that can only be diagnosed by methylation profiling. So optimally, testing for all four biomarkers would allow a complete picture of the tumor and an accurate integrated histomolecular diagnosis.

Robin (08:54):

Thanks for breaking that down, Dr. Ida. I'm wondering, are there any other commercial laboratories that offer all four types of molecular tumor testing?

Cris Ida (09:03):

No. Mayo Clinic Laboratories is the only large commercial lab that we know of in the United States that offers all four types of molecular tumor testing.

Robin (09:14):

And then, what is the main benefit for patients to have access to all four types in one laboratory? Dr. Aldape, do you want to take this one?

Kenneth Aldape (09:23):

Yes. The number one challenge of molecular testing, in my opinion, is time. What people care about is wanting to know the answer as soon as possible. Having all options available means time to diagnosis likely will be shortened. And this is especially important for CNS tumors, because some of these tumors are highly aggressive and require specific treatment to be initiated as soon as possible.

Robin (09:48):

Thanks, Dr. Aldape. Now I'm wondering, would the methylation profile array be ordered before or after, or maybe at the same time as testing for the other molecular biomarkers?

Kenneth Aldape (10:00):

As Dr. Ida said, all four types of molecular testing work together. There are certain tumors where you're not sure where the answer will lie, and you don't know ahead of time which platform's going to give them the result they need. Will it lie in the chromosomal microarray? Will it lie in next-generation sequencing? Will it lie in methylation? If it's a difficult case, you might want to hedge your bets by ordering the full portfolio with the expectation that at least one of them will give the answer. You wouldn't want to do this sequentially, because if you do, if you order one test, it might take two to three weeks. The second test will take just as long, and no one wants to wait. What people care about is wanting to know the answer as soon as possible. So when you order all four test types at the same time, the time to diagnosis will likely be shortened.

And this is especially important for CNS tumors, because diagnostic accuracy is the key to precision oncology that we want to deliver to patients affected by CNS tumors. So I think folks are going to be sending CNS tumors for simultaneous testing.

Robin (11:01):

I see. Thanks for sharing your opinions on that, Dr. Aldape. It really does make so much sense. I'd like to switch gears a little bit now and dig into the mechanics of our methylation array. How does this test, again, test code MTNON, work to identify the methylation status and better classify CNS tumors? Dr. Ida, do you want to take this one?

Cris Ida (11:22):

Yes. So the methylation molecular biomarker is very unique and distinct in how it's generated compared to how we detect other biomarkers. In terms of general lab workflow and chemistry, methylation array profiling is similar to chromosomal microarray. We start with DNA extraction from the formalin-fixed paraffin-embedded tumor tissue. Then we perform an initial step called bisulfate conversion. This step is very harsh and degrades the DNA, and as a result, we require more tissue than next-generation sequencing or copy number testing by chromosomal microarray. For example, the minimum tissue requirements for next-generation sequencing and chromosomal microarray testing is about two six-by-six millimeter tissue squares to get a minimum of 20 nanograms. Whereas for methylation array testing, we need about three to four tissue squares to reach at least 75 nanograms, which is almost four times more. We also need higher proportion of tumor, the so-called tumor purity, for methylation testing, and require at least 60% tumor.

For next-generation sequencing and chromosomal microarray, we are able to provide robust results with as little as 20% tumor. After bisulfate conversion, we use the Infinium MethylationEPPIC V2 kit to perform a highly multiplexed genotyping to evaluate methylation levels at over 900,000 sites across the genome.

The really unique and innovative aspect for the methylation array profiling is how we analyze the sample data to provide a classification based on the profile we get for the tested sample. Specifically, our test uses the NIH-developed NCI Bethesda CNS tumor classifier version two algorithm. We also developed an in-house quality control module and the NN method. The NN method is a nearest neighbors assisted unsupervised analysis that is complementary to the classifier. It allows us to objectively determine how a tested sample clusters with the classifier reference dataset and may further support classifier results. Of note, methylation array also provides MGMT promoter status, which is an established prognostic and predictive biomarker for patients with glioblastoma, the most common primary CNS malignant tumor in adults.

So in summary, our MTNON test reports that NCI/Bethesda classifier tumor classification, the NN method, and the MGMT promoter methylation status for a CNS formalin-fixed paraffin-embedded tumor tissue sample that meets our specimen requirements of at least 60% tumor and 75 nanogram DNA input.

Robin (14:19):

Thanks for sharing all those details, Dr. Ida. This test really does sound so sophisticated, and I'm curious about the classifier. If you don't mind, I'd like to get into the weeds a little bit about how that works. Dr. Aldape, since you were involved in developing the classifier, maybe you can elaborate a little bit more on the intricacies of this tool.

Kenneth Aldape (14:38):

Happy to. When I was at the NIH and we began development of this test, we started by accumulating data on the signatures of these tumors and found that particular signatures correlated quite well with a particular diagnosis. Therefore, we began to reverse engineer a classifier. In simplistic terms, if you have many samples of tumor type X and the signature of X, and many samples of tumor type Y and the signature of Y, you can then take a new sample and run it through the classifier and you can determine if it is more similar to X versus Y. The classifier is trained to recognize hundreds of unique methylation signatures. And when it recognizes that a sample has a match to a methylation family and a methylation class, it provides the classifier result, a confidence score, which could be high or low depending on how well the sample matches the classifier's training, and also the diagnostic implications of the findings.

Robin (15:33):

Thanks for explaining that, Dr. Aldape. I'm curious, what happens when the classifier cannot identify a match beyond obviously reporting that there's not a match? I mean, can the machine use or learn from the unmatched data? Dr. Ida, do you want to take this question?

Cris Ida (15:49):

That's a good question. The classifier is only as good as what it was trained for, and for some unusual tumors, it will not be able to provide a high-confidence classification. Methylation classifiers are iterative and improved performance with expanded reference datasets. Periodically, we can and plan to review the data of unmatched cases and may be able to recognize extended spectrum of existing classes or new clusters that may correspond to new tumor groups or subgroups.

Robin (16:22):

I see. That is just so interesting. Are there other classifiers out there, or does Mayo Clinic Labs have the only one?

Cris Ida (16:29):

Yes. There are other methylation CNS tumor classifiers in use. They differ as a result of the unique data used to train their algorithms and the type of the algorithm. However, to a certain extent, they're trained using overlapping publicly available data and provide concordant results. Our test uses one of the two classifiers being used nationally in a clinical laboratory environment.

Robin (16:56):

Thanks, Dr. Ida. It's so great to hear that Mayo Clinic Labs is on the leading-edge of this testing technology. And then this leads me to my next question. Could the classifier be used to help clarify diagnosis of other cancer types? Dr. Aldape, maybe you can take this one.

Kenneth Aldape (17:13):

Yes, I'm hoping that within one to two years, we might be able to validate the methylation array using classifiers for sarcomas, kidney tumors, and maybe even pan cancer. I may be biased, but it feels like the sky is the limit, and you'd probably want a classifier for every organ system.

Robin (17:30):

That is just such great news for patients with cancer. Let's talk a little bit more about the advantages of this test. Dr. Ida, what are your thoughts on how this test or results from this test will benefit patients?

Cris Ida (17:41):

From a patient perspective, the main benefit of this test is greater diagnostic precision. Everything starts with the diagnosis. When we are more precise, we better understand the disease process, supporting more accurate and cost-effective treatment decisions. From a clinical laboratory perspective, this test completes our neuro-oncology testing portfolio. Mayo Clinic has long offered testing for other CNS molecular biomarkers, but the lack of methylation profiling represented an important gap. MTNON fills that gap, allowing us to cover all molecular biomarkers included in the current WHO classification and position us at the forefront of comprehensive CNS tumor testing. By bringing this test online using a different platform and an AI-based approach to data analysis, we are stepping into a new era of clinical diagnostics. AI is increasingly part of medicine and our daily practice, and this is one of the first clinical assays in our laboratory to meaningfully leverage the technology for patient care.

Robin (18:51):

Thanks, Dr. Ida. AI is indeed a really powerful tool, and it's just such great news that we are able to harness the benefit of it for our patients. Dr. Aldape, I'm wondering if you have any final thoughts on the advantages of this test before we wrap up our conversation.

Kenneth Aldape (19:08):

Yes. To be honest, I really think what our clients want is an answer. What is the diagnosis? Has every reasonable step been taken to make sure the diagnosis is accurate? This test does that, but it also provides data that can be mined in the future to maybe understand new tumor types. It might allow us to identify new markers, new ideas, and new tests from that data that we're accumulating. So I think it's a win-win, where the deliverable is the bottom-line diagnosis and the downstream benefits that help you understand tumor biology and new treatments to help patients now and in the future.

Robin (19:43):

Thanks, Dr. Aldape, and you too, Dr. Ida, for providing both of your perspectives on how this test will help to improve care for patients with CNS tumors. It really is such an innovative assay with the power to provide patients with better outcomes, which is always the goal for our testing here at Mayo Clinic Labs. I have so appreciated our conversation today. I learned just a ton and hope our listeners did as well. And to our listeners, thanks for joining us. We hope you enjoyed today's discussion, and will join again for our next installment of “Diagnostics in Practice.”