Clinical utility of whole-genome DNA methylation profiling as a primary molecular diagnostic assay for central nervous system tumors-A prospective study and guidelines for clinical testing

Abstract

Background: Central nervous system (CNS) cancer is the 10th leading cause of cancer-associated deaths for adults, but the leading cause in pediatric patients and young adults. The variety and complexity of histologic subtypes can lead to diagnostic errors. DNA methylation is an epigenetic modification that provides a tumor type-specific signature that can be used for diagnosis.

Methods: We performed a prospective study using DNA methylation analysis as a primary diagnostic method for 1921 brain tumors. All tumors received a pathology diagnosis and profiling by whole genome DNA methylation, followed by next-generation DNA and RNA sequencing. Results were stratified by concordance between DNA methylation and histopathology, establishing diagnostic utility.

Results: Of the 1602 cases with a World Health Organization histologic diagnosis, DNA methylation identified a diagnostic mismatch in 225 cases (14%), 78 cases (5%) did not classify with any class, and in an additional 110 (7%) cases DNA methylation confirmed the diagnosis and provided prognostic information. Of 319 cases carrying 195 different descriptive histologic diagnoses, DNA methylation provided a definitive diagnosis in 273 (86%) cases, separated them into 55 methylation classes, and changed the grading in 58 (18%) cases.

Conclusions: DNA methylation analysis is a robust method to diagnose primary CNS tumors, improving diagnostic accuracy, decreasing diagnostic errors and inconclusive diagnoses, and providing prognostic subclassification. This study provides a framework for inclusion of DNA methylation profiling as a primary molecular diagnostic test into professional guidelines for CNS tumors. The benefits include increased diagnostic accuracy, improved patient management, and refinements in clinical trial design.

Keywords: DNA methylation; central nervous system tumors; guidelines; molecular; tumor classification.

Figure 1.

(A) This prospective study started with surgical resection of the brain tumor and tissue processing for a pathologist. All tumors received the standard of care pathology diagnosis as judged appropriate at the time of initial review, and simultaneous whole genome DNA methylation profiling. The histologic diagnosis and the DNA methylation diagnosis were compared and additional molecular studies including DNA and RNA NGS studies were performed as required to resolve discrepant cases. (B) Our cohort included 1921 primary central nervous system tumors, of which 1602 (83%) had World Health Organization (WHO) recognized diagnoses and 319 (17%) had descriptive diagnoses. (C) Of the 1602 WHO diagnoses, 1189 (74%) tumors showed concordance between histopathology and DNA methylation and were considered a complete diagnostic match, 225 (14%) tumors were a diagnostic mismatch with discrepant tumor type and/or grade, 110 (7%) tumors DNA methylation was able to add additional prognostic information, and 78 (5%) tumors did not classify by DNA methylation (referred to as “no match”). (D) Of the 319 tumors carrying descriptive diagnoses, DNA methylation provided a conclusive diagnosis in 273 (86%), 46 (14%) tumors did not classify and were therefore considered “no match.”

Figure 2.

Diagnostic utility for accurate diagnosis and prognostic stratification. Six tumor groups with the highest yield of DNA methylation included GBM, ependymoma, glioneuronal tumors, oligodendroglioma, astrocytoma IDH mutant, and medulloblastoma. (A) GBM (N = 390) were a complete match in 82% of cases, a diagnostic mismatch in 13% of cases, and did not classify with any entity by DNA methylation in 5% of cases (no match). Most misdiagnosed GBMs were reclassified as diffuse midline glioma K27 altered (31%), anaplastic pilocytic astrocytoma (10%), and pleomorphic xanthoastrocytoma (7%). (B) Ependymoma (N = 109) had a complete match rate of 69%, a diagnostic mismatch rate of 23%, and a no-match rate of 8%. Ependymomas were most commonly reclassified as myxopapillary ependymoma (24%) and subependymoma (28%), (C) Glio-neuronal tumors (N = 160) had a diagnostic complete match rate of 61%, a diagnostic mismatch rate of 27%, and a no match rate of 12%. Pilocytic astrocytoma (N = 80) were reclassified by DNA methylation in 20 cases (28%) and DNA methylation upgraded the diagnosis in 11% of these cases. Ganglioglioma (N = 32) had a diagnostic mismatch rate of 31% and DNA methylation upgraded the diagnosis in 15% of cases. (D) Oligodendroglioma (N = 66) had a complete match rate of 83%, a diagnostic mismatch rate of 15%, and a no-match rate of 2%. Tumors diagnosed histologically as oligodendroglioma are most often reclassified as astrocytoma (10%), glioblastoma (2%), and DNET (3%). (E) Astrocytoma IDH mutant (N = 96) were a complete match in 65% of cases, a diagnostic mismatch in 31% of cases, and a no match in 4% of cases. Astrocytoma IDH mutant World Health Organization (WHO) grade 2 was most reclassified as a higher-grade IDH mutant astrocytoma (11%), astrocytoma IDH mutant WHO grade 3 was most commonly reclassified as a lower-grade IDH mutant glioma in 48% of cases, and astrocytoma IDH mutant WHO grade 4 most commonly reclassified as a lower grade IDH mutant astrocytoma in 28% of cases. (F) While medulloblastoma is rarely misdiagnosed (3% of cases) DNA methylation provides prognostic information by stratifying tumors into established molecular subgroups including Shh, Wnt, group 3, and group 4.

Figure 3.

Of the entire 1921 cohort, 319 (17%) brain tumors were diagnosed descriptively and carried 195 different descriptive diagnoses. For the analysis, tumors were stratified into adult high-grade, adult low-grade, pediatric high-grade, and pediatric low-grade. In the adult high-grade group, there were 67 tumors and 49 unique descriptive diagnoses for which DNA methylation was able to provide a diagnosis in 94% of cases resulting in 20 different methylation classes (A). In the adult low-grade group, there were 99 tumors and 62 unique descriptive diagnoses for which DNA methylation was able to provide a diagnosis in 86% of cases resulting in 29 different methylation classes (B). In the pediatric high-grade group, there were 65 tumors and 24 unique descriptive diagnoses for which DNA methylation was able to provide a diagnosis in 92% of cases resulting in 17 different methylation classes (C). In the pediatric low-grade group, there were 69 tumors and unique descriptive diagnoses for which DNA methylation was able to provide a diagnosis in 91% of cases resulting in 17 different methylation classes (D). For the full list of descriptive diagnoses and DNA methylation classes see Supplementary Table 1. For the list of abbreviations see Supplementary Table 3.

Figure 4.

DNA methylation and clinical re-stratification of descriptive cases: In addition to providing accurate diagnosis, DNA methylation changed grading of the tumors. In total, 40 (26%) out of 155 descriptive low-grade tumors were upgraded to a higher-grade tumor type by DNA methylation, and 26 (22%) out of 118 were downgraded to a lower-grade tumor by DNA methylation.