Provider Faqs

What is Whole-Genome DNA Methylation Profiling using the Epignostix Classifier?

Whole-Genome DNA Methylation Profiling is a molecular diagnostic assay that measures DNA methylation at more than 900,000 CpG sites across the genome using the Illumina Infinium MethylationEPIC v2.0 BeadChip.

Gliolab analyzes the resulting methylation profile using the Heidelberg CNS Tumor Methylation Classifier v12.8 implemented through Epignostix epxCNS. The classifier compares the tumor’s methylation signature with reference profiles of well-characterized CNS tumors to provide molecular classification.

From the same assay, Gliolab also provides MGMT promoter methylation prediction and genome-wide copy-number information, including copy-number assessment of 29 CNS tumor-relevant gene regions. The ability to obtain classification, MGMT status, and copy-number information from a single methylation assay can also help conserve limited tumor tissue.

CNS tumors with different underlying biology can have overlapping histologic features. DNA methylation profiling provides an objective molecular approach that can help confirm or refine a diagnosis, resolve diagnostically challenging cases, and provide additional molecular characterization of the tumor.

cIMPACT-NOW Update 9 describes DNA methylation profiling as a valuable tool for CNS tumor diagnosis and recommends that results be interpreted in the context of the entire case. The WHO Classification of CNS Tumors also identifies methylation profiling as an essential or desirable diagnostic criterion for many CNS tumor types.

Experience from multiple centers has demonstrated improved diagnostic accuracy using DNA methylation-based classification.

Methylation profiling should therefore be used as an adjunct to, rather than a replacement for, histopathologic and clinical evaluation.

Whole-Genome DNA Methylation Profiling may be particularly valuable for:

  • Newly diagnosed glioblastoma — to provide molecular confirmation or refinement of the diagnosis and additional tumor characterization.
  • Diagnostically challenging CNS tumors — particularly when histologic and molecular findings are ambiguous or discordant.
  • Molecular classification and subclassification — when further molecular characterization of the tumor is clinically useful.
  • MGMT promoter methylation assessment — an important biomarker for treatment considerations in glioblastoma.
  • Copy-number characterization — including genome-wide alterations and assessment of 29 CNS tumor-relevant gene regions.
  • Recurrent glioblastoma when tissue from repeat surgery is available — to characterize the recurrent tumor and, when results from the original tumor are available, compare molecular features over the course of disease in order to guide treatment.

cIMPACT-NOW specifically recommends considering DNA methylation profiling when it may help resolve a differential diagnosis, when a diagnosis is challenging or findings are discrepant, or when additional molecular subtyping is needed.

Gliolab uses Epignostix epxCNS to implement the Heidelberg CNS Tumor Methylation Classifier v12.8, an extensively developed machine-learning classifier for CNS tumors.

The v12.8 classifier was trained on 7,495 CNS tumor methylation profiles and expanded the previous classifier from 91 classes to 184 subclasses. It uses a four-tier hierarchical structure consisting of superfamily, family, class, and subclass.

The hierarchical structure allows a tumor to be classified at the level best supported by its methylation profile, increasing diagnostic accuracy and assisting in treatment selection.

The classifier provides a calibrated score ranging from 0 to 1 that reflects the confidence of the tumor’s match to a particular methylation class. Higher scores indicate greater confidence in the classification.

Gliolab reports calibrated scores at the superfamily, family, class, and subclass levels. A calibrated score of ≥0.90 is considered a confident match in Gliolab’s validated reporting criteria.

Because classification is hierarchical, a tumor may have a confident match at a broader level, such as family, while having a lower-confidence match at the more specific class or subclass level.

Scores should not be interpreted in isolation. cIMPACT-NOW emphasizes that classifier scores represent a continuum and should be evaluated together with clinical, radiologic, histologic, and other molecular findings.

The Gliolab clinical report provides:

  • Molecular classification at the superfamily, family, class, and subclass levels, supported by the classifier
  • Calibrated classifier scores
  • MGMT promoter methylation prediction, including predicted probability and confidence interval
  • Genome-wide copy-number profile
  • Copy-number assessment of 29 CNS tumor-relevant gene regions
  • Relevant chromosomal gains and losses
  • An integrated interpretation of the molecular findings

The report also includes methodology, assay limitations, and relevant references.

MGMT promoter methylation is inferred from the methylation-array data using the validated MGMT-STP27 prediction model implemented through Epignostix epxCNS.

Gliolab reports the predicted MGMT promoter methylation status, probability of promoter methylation, confidence interval, and cutoff.

MGMT promoter methylation is an important predictive biomarker in glioblastoma and is associated with greater likelihood of benefit from alkylating chemotherapy, particularly temozolomide.

DNA methylation profiling can provide increasingly specific molecular characterization beyond the broader diagnosis of glioblastoma.

The current Heidelberg classifier recognizes molecularly distinct methylation groups within Glioblastoma, IDH-wildtype, including mesenchymal, RTK1, RTK2, etc.

The v12.8 classifier organizes results hierarchically as superfamily, family, class, and subclass. Established diagnostic entities generally reside at the family or class level, while some more recently recognized molecular distinctions occur at the subclass level.

Class and subclass information can provide additional insight into tumor biology and may be useful when considering clinical trials or to guide treatment.

Yes. If tumor tissue is available from repeat surgery, DNA methylation profiling can characterize the recurrent tumor and determine whether it remains molecularly classifiable as glioblastoma.

When methylation results from the original tumor are available, comparison of the primary and recurrent tumors may identify changes in methylation class or subclass, MGMT promoter methylation, and copy-number alterations that have occurred during tumor evolution and treatment.

The clinical significance of changes observed between primary and recurrent tumors should be interpreted in conjunction with histopathology, treatment history, and other clinical findings.

The methylation-array data are also used to generate a genome-wide copy-number profile, providing information about chromosomal gains and losses.

Gliolab reports copy-number alterations involving 29 CNS tumor-relevant gene regions, as well as characteristic chromosomal patterns relevant to glioblastoma, including chromosome 7 gain and chromosome 10 loss.

Copy-number findings are categorized as amplification, gain, neutral, loss, or homozygous deletion according to defined thresholds.

Gliolab accepts formalin-fixed, paraffin-embedded (FFPE) tumor tissue as unstained slides or scrolls, as well as extracted tumor DNA.

For FFPE tumor tissue, submit either:

15 unstained slides, 5 μm thick, plus 1 H&E slide with the tumor area circled and tumor percentage indicated;

OR

10 FFPE scrolls, 50 μm thick, plus 1 H&E slide with the tumor area circled and tumor percentage indicated.

For extracted tumor DNA, submit ≥500 ng total DNA at a concentration of ≥15 ng/μL.

The expected turnaround time is 7–14 business days after receipt of an acceptable specimen.

Whole-Genome DNA Methylation Profiling must be ordered by a physician or other authorized healthcare provider through Gliolab’s secure LabLink provider portal.

New providers first request a LabLink User ID and provide their contact information, institution or facility, and ordering physician NPI. After the account is approved, the provider receives an email with instructions for establishing a password and accessing LabLink.

Once logged in, the provider can access Resource Materials, download and complete the Test Requisition Form (TRF), and provide the required information.

Our Whole-Genome DNA Methylation test is covered by Medicare.

Gliolab is committed to providing testing to glioblastoma patients regardless of insurance status. If you have questions about insurance, billing, or financial assistance, please contact Gliolab.

At this time, Gliolab is not accepting specimens originating from California, New York, Maryland, or Pennsylvania due to state laboratory licensing requirements. Gliolab is working to expand testing availability in these states.

If you created a LabLink User ID, you can obtain test results by logging in to LabLink when you receive an email informing you that your results have been uploaded. 

If you emailed your Test Requisition Form to us, we will send you an email informing you that your results are available. 

Whole-Genome DNA Methylation Profiling must be ordered by a physician or other authorized healthcare provider through Gliolab’s secure LabLink provider portal.

New providers first request a LabLink User ID and provide their contact information, institution or facility, and ordering physician NPI. After the account is approved, the provider receives an email with instructions for establishing a password and accessing LabLink.

Once logged in, the provider can access Resource Materials, download and complete the Test Requisition Form (TRF), and provide the required information.

Gliolab currently accepts patient samples from all U.S. states except California, Maryland, New York, and Pennsylvania.

 

These four states have additional laboratory licensing requirements. Gliolab is currently pursuing licensure in these states and plans to make testing available once the required licenses are obtained.

Currently, Gliolab only accepts tumor samples a pathologist has found to be Glioblastoma, but in the future we will accept all brain cancer tumor samples.

Scroll to Top