Skip to main content
Download PDF

Variant allelic frequencies of driver mutations can identify gliomas with potentially false-negative MGMT promoter methylation results

Acta Neuropathologica Communications volume 11, Article number: 175 (2023) Cite this article

Abstract

MGMT promoter methylation testing is required for prognosis and predicting temozolomide response in gliomas. Accurate results depend on sufficient tumor cellularity, but histologic estimates of cellularity are subjective. We sought to determine whether driver mutation variant allelic frequency (VAF) could serve as a more objective metric for cellularity and identify possible false-negative MGMT samples. Among 691 adult-type diffuse gliomas, MGMT promoter methylation was assessed by pyrosequencing (N = 445) or DNA methylation array (N = 246); VAFs of TERT and IDH driver mutations were assessed by next generation sequencing. MGMT results were analyzed in relation to VAF. By pyrosequencing, 56% of all gliomas with driver mutation VAF ≥ 0.325 had MGMT promoter methylation, versus only 37% with VAF < 0.325 (p < 0.0001). The mean MGMT promoter pyrosequencing score was 19.3% for samples with VAF VAF ≥ 0.325, versus 12.7% for samples with VAF < 0.325 (p < 0.0001). Optimal VAF cutoffs differed among glioma subtypes (IDH wildtype glioblastoma: 0.12–0.18, IDH mutant astrocytoma: ~0.33, IDH mutant and 1p/19q co-deleted oligodendroglioma: 0.3–0.4). Methylation array was more sensitive for MGMT promoter methylation at lower VAFs than pyrosequencing. Microscopic examination tended to overestimate tumor cellularity when VAF was low. Re-testing low-VAF cases with methylation array and droplet digital PCR (ddPCR) confirmed that a subset of them had originally been false-negative. We conclude that driver mutation VAF is a useful quality assurance metric when evaluating MGMT promoter methylation tests, as it can help identify possible false-negative cases.

Introduction

Temozolomide (TMZ) improves survival in high-grade gliomas [1,2,3] and is standard of care for these tumors [4]. TMZ alkylates DNA bases in tumor cells, especially guanine. O6-methylguainine is a common product of TMZ alkylation, and plays a key role in TMZ-driver tumor cell apoptosis. During DNA replication, O6-methylguanine residues erroneously pair with thymine on the new DNA strand. DNA mismatch repair enzymes excise the thymine residues, but methylguanine residues persist on the template DNA strand, triggering a series of “futile repair” cycles, leading to tumor cell apoptosis [5, 6]. The enzyme O6-methylguanine-DNA-methyltransferase (MGMT) counteracts TMZ by removing methyl groups from O6-methylguanine residues [7,8,9]. The gene encoding the enzyme, MGMT, is located at 10q26 [10], and methylation of its promoter inhibits gene expression, making tumor cells more vulnerable to TMZ [11,12,13] and improving clinical response to the drug [14]. MGMT promoter methylation is also associated with improved prognosis even in the absence of TMZ treatment [15].

Due to its importance in prognosis and in predicting TMZ response, testing for MGMT promoter methylation is required in the diagnostic workup of adult-type diffuse gliomas. Multiple approaches have been developed for testing MGMT promoter methylation, including pyrosequencing, methylation-specific PCR, methylation-specific high-resolution melting, and MethyLight™. Because pyrosequencing is simple, reproducible, and a good predictor of TMZ response, it has become the most widespread test [16,17,18,19]. More recently, genomic DNA methylation array and droplet digital PCR (ddPCR) have emerged as alternative methods [20,21,22,23,24].

In MGMT pyrosequencing, the promoter is usually considered “methylated” (positive) when ≥ 10% of measured CpG sites in the promoter sequence are methylated, and “unmethylated” (negative) when < 10% are methylated. This cutoff has been validated clinically [17] and is used for MGMT promoter testing at many institutions, although other cutoffs have been suggested [25, 26]. Accurate results depend on sufficient tumor cellularity in analyzed samples. Methylation of the promoter is abnormal, occurring only in neoplastic cells. Thus, if tumor cellularity is too low in a specimen, with too many admixed non-neoplastic cells (including astrocytes, neurons, inflammatory cells, etc.), MGMT promoter methylation signals could be diluted, leading to a false-negative result. A minimum of 70% tumor cellularity is preferred for MGMT promoter analysis, a cutoff also recommended in other diagnostic assays such as methylation profiling arrays [27, 28]. Tumor cellularity is usually estimated via subjective light microscopic evaluation of formalin fixed, paraffin embedded (FFPE) tumor tissue stained with hematoxylin and eosin (H&E). Moreover, specimens with < 70% estimated tumor cellularity are nearly always still tested if no better tissue block is available.

Adult-type diffuse gliomas usually have characteristic driver mutations whose relative amounts in an analyzed sample are quantifiable by next generation sequencing (NGS). TERT promoter mutation is found in ~ 85% of IDH-wildtype glioblastomas and > 95% of IDH-mutant and 1p/19q co-deleted oligodendrogliomas [29, 30]. Mutations in IDH1 or IDH2 are disease-defining features of IDH-mutant and 1p/19q co-deleted oligodendroglioma and IDH-mutant astrocytoma [31]. These specific driver mutations are almost always heterozygous and are not usually affected by copy number alterations [32,33,34]. Thus, tumor cellularity in most adult-type diffuse gliomas can be estimated by simply doubling the measured variant allelic frequency (VAF) of the driver mutation. Previous work, including our own, supports this approach for quantifying tumor cellularity [35, 36].

Here, we investigated driver mutation VAF as a way to evaluate MGMT promoter methylation test results by pyrosequencing, DNA methylation array, and ddPCR in a large single-institution observational study of adult-type diffuse gliomas as defined by the 5th edition of the WHO classification [31]. The central hypothesis was that glioma samples with low VAF are at risk of false negative MGMT assay results due to dilution of tumor methylation signals by non-neoplastic cells.

Materials and methods

Collection of patient samples

The cohort comprised 691 consecutive CNS WHO grade 2–4 adult-type diffuse gliomas from 658 patients resected between 2006 and 2022. Between 2019 and 2022, all tumors underwent NGS and MGMT promoter methylation testing, the latter by pyrosequencing (445 samples) or by DNA methylation array (246 samples) at Northwestern Memorial Hospital. All tumors were classified according to 2021 WHO guidelines [31]. Tumor types included glioblastoma, IDH-wildtype, CNS WHO grade 4 (IDHwt GBM); IDH-mutant astrocytoma, CNS WHO grades 2–4 (IDHmut astrocytoma); and IDH mutant and 1p/19q co-deleted oligodendroglioma CNS WHO grades 2–3 (IDHmut oligodendroglioma). Among IDHwt GBMs, only those with TERT promoter mutation were included due to the high frequency of that mutation in IDHwt GBM and its utility as a marker of tumor purity [35, 36]. MGMT and NGS data were obtained from the same tissue block in each case. Key characteristics of the patient cohort are summarized in Table 1. Characteristics of tumor samples and patients are also illustrated in Fig. 1.

Table 1 Patient cohort characteristics
Full size table
Fig. 1
figure 1

Cohort characteristics. Flowchart for key characteristics of tumor samples and patients. Illustration prepared with Biorender. GBM: IDH-wildtype glioblastoma

Full size image

MGMT promoter methylation testing

All MGMT promoter methylation assessments were carried out according to manufacturer instructions for each respective assay. For pyrosequencing, DNA was extracted from FFPE tissue using Purigen Biosystems (Pleasanton, CA, USA), then bisulfite-modified using the Qiagen EpiTect bisulfite Kit (Qiagen). DNA was amplified by PCR, and pyrosequencing was performed using primers in four CpG sites in exon 1 of the human MGMT gene (sequence on chromosome 10 from 131,265,519 to 131,265,537: CGACGCCCGCAGGTCCTCG). PyroGold Q24 SQA Reagents and the Pyro Q-CpG software on a PyroMark ID pyrosequencer (Qiagen, Crawley, UK) were used. For DNA methylation array, analysis of two CpG sites in the MGMT promoter (cg12434587, chr10:131,265,209–131,265,210 and cg12981137, chr10:131,265,575–131,265,576) was performed according to the Bady algorithm as previously described [20]. Briefly, the M-values of the methylated and unmethylated intensities at those 2 positions were used as input in a logistic regression model (MGMT-STP27). The MGMT score was obtained by logit-transformation of the probability that the MGMT promoter is methylated. The predicted probabilities, MGMT score, confidence intervals, and MGMT classification were directly obtained by the function MGMT predict from the R package mgmtstp27 (https://github.com/badozor/mgmtstp27). MGMT assessment via droplet digital PCR (ddPCR) was performed as previously described [24]. Briefly, genomic DNA samples were subjected to bisulfite conversion using the Epi-Tect Fast Bisulfite Conversion kit as per manufacturer’s instruction (Qiagen, Germantown MD, USA). MGMT promoter methylation was quantified using primers and probes targeting the converted and unconverted template (Supplementary Table S1) and ddPCR Supmermix for Probes as per manufacturer’s instructions (Bio-Rad, Hercules, CA, USA). Our ddPCR assay targeted the same 2 MGMT promoter CpG sites as the Bady algorithm, which show good correlation with TMZ response and survival [37].

Next generation sequencing

Two commercially available NGS panels were used. Oncomine Comprehensive Assay Version 3 (OCAV3) was performed on 519 cases between 2019 and 2022, according to manufacturer’s instructions. PGDx Solid Tumor NGS (adopted by Northwestern Memorial Hospital in 2022) was performed on 172 cases, according to the manufacturer’s recommendations. Detailed methodological descriptions for both NGS panels can be found in the “Supplementary Methods” section.

TERT promoter droplet digital PCR

TERT promoter mutation assessment by droplet digital PCR (ddPCR) was performed using the BioRad™ QX200 ddPCR system, according to the manufacturer’s recommendations. Briefly, genomic DNA was extracted, and partitioned into droplets. The PCR reaction was performed using primers for the two most common TERT promoter mutations (c.-124 C > T and c.-146 C > T). Results were quantified and analyzed using a droplet reader, as previously described [38].

DNA methylation array-based tumor classification

For each of the 246 samples in which the MGMT promoter was assessed by genomic DNA methylation array, DNA methylation-based tumor classification was also performed using our previously validated in-house classifier [21, 39]. Briefly, extracted DNA underwent bisulfite conversion using the EZ-96 DNA methylation kit (Zymo Research Corp., Tustin, CA). After denaturation with 0.1 N NaOH and cleaning with a ZR-96 DNA concentrator-5 kit (Zymo Research Corp.), bisulfite-converted DNA was hybridized to Infinium Human Methylation EPIC BeadChips (EPIC, 850 K). Arrays were scanned and IDAT files generated. Each BeadChip had quality metrics assessed for red and green staining, hybridization, and bisulfite conversion. Methylation array data was analyzed using our classifier. Cases with methylation class score ≥ 0.9 were considered a match.

Data collection

For each tumor sample, pathology reports were reviewed. Nine cases needed their final diagnoses adjusted to fit with the 2021 WHO classification system based on molecular and morphologic criteria.

Parallel testing from different tumor regions

We performed parallel testing for TERT promoter VAF and MGMT promoter methylation on 5 IDH-wildtype GBM samples from the Northwestern University Nervous System Tumor Bank (NSTB). Each sample had previously undergone pathology review and clinical molecular testing, had an established diagnosis of IDH-wildtype GBM with TERT mutation, and had a positive test for MGMT promoter methylation by DNA methylation array. For each sample, one area of at least 0.5 cm2 with the highest relative cellularity (estimated by light microscopy of an H&E section) and a corresponding area of similar size with the lowest relative cellularity were differentially delineated via light microscopy of H&E slides. TERT VAF and MGMT promoter methylation were measured in parallel by ddPCR on tissue dissected from each delineated area.

Re-testing of tumor samples

We re-tested 12 IDHwt GBM samples (Table 2) for MGMT promoter methylation, by both DNA methylation array and ddPCR, that had previously been tested by pyrosequencing. Samples were selected based on tissue availability, TERT VAF, and availability of clinical follow-up data. In all cases, re-testing was performed on the same FFPE tissue blocks that had originally been tested with pyrosequencing. Six samples had TERT VAF < 0.1, and 6 had VAF > 0.25. Samples from the low-VAF and high-VAF cohorts were sex- and age-matched.

Table 2 Analysis of differentially dissected tumor samples
Full size table

Data processing and statistical analysis

Differences between observed versus expected frequency of positive and negative test results were compared with Fisher’s exact test. The Fisher’s exact test function of Cutoff Finder [40], available at https://molpathoheidelberg.shinyapps.io/CutoffFinder_v1/, was used to identify VAF cutoff points with maximal differences in test outcomes. TERT mutation VAF was used for analysis of IDHwt GBM, and IDH1/2 VAF for both IDHmut astrocytoma and IDHmut oligodendroglioma. Differences between means of 2 groups were compared with Student’s t-test. Non-parametric comparisons between 2 groups were done with Mann-Whitney U test (Wilcoxon rank-sum test).

To calculate cumulative mean positivity rates of MGMT promoter assays, samples were ranked by increasing VAF, and the mean positivity rate was measured for each sample plus all samples below it. This process was repeated successively (mean of the first 2 samples with the lowest VAF, then the first 3, etc.) until the entire cohort had been measured. Cumulative means and medians were calculated for MGMT promoter pyrosequencing scores in a similar fashion. Simple linear regression was performed comparing MGMT promoter pyrosequencing score to driver mutation VAF for each sample. Two-part linear regressions were performed on cumulative mean positivity rate data to analyze the variation in likelihood of positive MGMT pyrosequencing results with increasing driver mutation VAF. The data curves for cumulative mean positivity rate of each glioma subtype showed two distinct portions. Regression lines were fitted to both portions of each curve. In all cases, the slope of the line at lower VAF values was much steeper than at high VAF values. The regression lines were separated at the VAF value where the slopes had the greatest magnitude of difference (i.e., slope value M for the low-VAF regression line minus slope value M for the high-VAF regression line). All VAF values were evaluated for each pair of regression lines. All linear regressions are described in terms of slope values of best-fit lines, R2 values for goodness-of-fit, and p-values for slope deviation from zero. Tumor cellularity was estimated from TERT or IDH driver mutation VAF by simply doubling VAF values and multiplying by 100% (these mutations are usually heterozygous). In rare cases with VAF > 0.5, the cellularity estimate was capped at 100%.

For all statistical tests, p-values less than 0.05 were considered significant. For non-parametric tests, approximate p-values are reported. Data organization, processing, statistical analysis, and figure preparation were carried out with Microsoft Excel 2016, GraphPad Prism 5, Cutoff Finder, and BioRender.

Results

Cohort characteristics

The cohort was composed of 691 consecutive adult-type diffuse glioma samples, including 468 IDHwt GBMs, 144 IDHmut astrocytomas, and 80 IDHmut oligodendrogliomas. Tumor samples were from 658 patients (405 male, 253 female), 153 samples were from residual/recurrent tumors, and 36 were paired original and recurrent tumors. Among 445 samples assessed via pyrosequencing, 217 (49%) tested positive for MGMT promoter methylation (101 IDHwt GBMs, 68 IDHmut astrocytomas, 48 IDHmut oligodendrogliomas), while 228 (51%) tested negative (200 IDHwt GBMs, 25 IDHmut astrocytomas, 3 IDHmut oligodendrogliomas). Among 246 samples assessed via DNA methylation array, 140 (57%) tested positive for MGMT promoter methylation (79 IDHwt GBMs, 32 IDHmut astrocytomas, 29 IDHmut oligodendrogliomas), while 106 (43%) tested negative (88 IDHwt GBMs, 18 IDHmut astrocytomas, 0 IDHmut oligodendrogliomas). Among the 246 cases analyzed by methylation array, 194 (79%) had a methylation classifier score ≥ 0.9. One tumor matched to methylation class “control tissue: inflammatory tumor microenvironment,” but morphologic and molecular features indicated a diagnosis of IDH-wildtype GBM; NGS revealed very low TERT VAF (0.09) in that case. Despite the low VAF, the tumor tested positive for MGMT promoter methylation on the array. Among the 36 paired original-recurrent samples, recurrent tumors from two patients had discordant MGMT results with the primary tumors. In one case, a newly diagnosed grade 4 IDH-mutant astrocytoma was positive for MGMT promoter methylation on pyrosequencing (35.9% CpG site methylation), but the recurrent tumor 8 months later was negative on pyrosequencing (5.4%). In the second case, a newly diagnosed IDH-wildtype GBM was negative by pyrosequencing (3.9%), but the recurrence 4 months later was positive by methylation array. Patient characteristics are summarized in Table 1. Cohort characteristics are illustrated in Fig. 1. Additional details are available in Supplementary Table S2 (pyrosequencing data) and Supplementary Table S3 (DNA methylation array data). Details of the paired original and recurrent samples are available in Supplementary Table S4.

MGMT promoter methylation outcomes by pyrosequencing according to driver mutation VAF in adult-type diffuse gliomas

Among all adult-type diffuse gliomas combined, the optimal driver mutation VAF (either TERT for IDHwt GBM or IDH1/2 for IDHmut astrocytoma and IDHmut oligodendroglioma) in the context of MGMT promoter methylation pyrosequencing was determined by Cutoff Finder at 0.325 (Supplementary Figure S1A). At or above that cutoff, 152/270 samples (56%) tested positive for MGMT promoter methylation, compared to 65/175 (37%) below (p < 0.0001 via Fisher’s exact test, Fig. 2A). We next plotted cumulative mean MGMT promoter positivity rates as a function of driver mutation VAF. Results for all gliomas pooled together showed that maximal positivity rates were at a VAF of ~ 0.48 or higher, with intermediate positivity rates between 0.18 and 0.48. Below a VAF of 0.18, the likelihood of a given MGMT test being positive sharply dropped (Fig. 2B). We then plotted the cumulative mean MGMT promoter pyrosequencing score (percent of methylated CpG sites) as a function of driver VAF. The curve did not plateau until VAF was above 0.325 (Fig. 2C). Mean MGMT score at or above a VAF of 0.325 was 19.3% versus 12.7% below (p < 0.0001 via unpaired Student’s t-test, Fig. 2D). Mann-Whitney U test showed significant variation in median scores above (13.2%) and below (5.7%) this cutoff (p < 0.0001, Supplementary Figure S2A), keeping in mind that the standard cutoff for a “methylated” pyrosequencing result is 10%. The curve for cumulative median scores showed a sigmoid shape, with distinct plateaus above VAF = 0.06 and VAF = 0.4 (Supplementary Figure S2B).

Fig. 2
figure 2

MGMTpromoter methylation trends in pyrosequencing samples. (A) Frequency of positive results for all glioma samples above and below the cutoff value of VAF = 0.325. (B) Cumulative mean frequency of positive test results as a function of VAF. (C) Trends in cumulative mean MGMT promoter pyrosequencing score with increasing VAF. (D) Mean MGMT promoter pyrosequencing scores above and below VAF = 0.325. Similar results are shown for each tumor subtype, including IDH-wildtype glioblastoma (E-H), IDH-mutant astrocytoma (I-L) and IDH-mutant and 1p/19q co-deleted oligodendroglioma (M-P). (Horizontal dashed black lines: mean values for cohort; Horizontal solid red lines: MGMT positivity cutoff of 10.0%; Vertical dashed black lines: cutoff values identified by Cutoff Finder; Vertical dashed red lines: cutoff values identified by multi-part linear regression; Panels A, E, I, M: Fisher’s exact test; Panels D, H, L, P: unpaired Student’s T-test; pyroseq: pyrosequencing; *p < 0.05; ****p < 0.0001)

Full size image

Optimal VAF cutoffs via categorical Fisher’s exact tests in Cutoff Finder for the 3 major glioma subtypes (Supplementary Figure S1B-D) were 0.115 for IDHwt GBM (Fig. 2E), 0.325 for IDHmut astrocytoma (Fig. 2I), and 0.405 for IDHmut oligodendroglioma (Fig. 2M). While none of those exact cutoffs were statistically significant (see Supplementary Table S5 for details), there were clearly different slopes of MGMT promoter methylation positivity curves as a function of VAF within each glioma subtype (Fig. 2F J, and 2 N). For each curve, we performed multi-part linear regression analyses as an alternative approach to identifying cutoffs. We found that the slopes were maximally different at subtype-specific VAF inflection points: TERT VAF = 0.18 for IDHwt GBM, IDH VAF = 0.325 for IDHmut astrocytoma, and IDH VAF = 0.30 for IDHmut oligodendroglioma (Supplementary Figures S2C, S2F, and S2I, respectively). Cumulative mean MGMT pyrosequencing score trends are shown for IDHwt GBM (Fig. 2G), IDHmut astrocytoma (Fig. 2K), and IDHmut oligodendroglioma (Fig. 2O). Cumulative mean trends for each individual CpG pyrosequencing site are shown in Supplementary Figure S3. Mean MGMT scores for samples above and below VAF values identified by Cutoff Finder did not show statistically significant differences via unpaired student’s t-test (not shown). For IDHwt GBM, mean MGMT scores above and below the TERT VAF cutoff of 0.18 (identified by multi-part linear regression) were 14.06% (in the “methylated” range) and 9.56% (in the “unmethylated” range), respectively (p = 0.105 via unpaired student’s t-test, Fig. 2H). For IDHmut astrocytoma (Fig. 2L), Cutoff Finder and regression both identified the same VAF cutoff at which the difference in mean MGMT scores maximize (IDH VAF = 0.325, p = 0.05). For IDHmut oligodendroglioma (Fig. 2P), mean MGMT score differences above and below the regression cutoff (IDH VAF = 0.30) did reach statistical significance (p < 0.0001). Mann-Whitney U tests based on regression cutoffs for glioma subtypes showed significant variation of median MGMT promoter methylation scores only for IDHmut oligodendroglioma (18.2% vs. 27.2%, p = 0.0003, Supplementary Figure S2D, S2G, S2J). Cumulative median trends are shown in Supplementary Figures S2E, S2H, and S2K.

Together, these data suggest that, since the rate of MGMT promoter methylation drops sharply below a TERT promoter VAF of ~ 0.18 (i.e., ~ 35–40% tumor cellularity) for IDHwt GBM, cases with such low VAF may be at increased risk of false-negative MGMT promoter methylation results by pyrosequencing. Similar cutoffs are at ~ 0.33 and 0.30–0.40 for IDHmut astrocytomas and IDHmut oligodendrogliomas, respectively.

MGMT promoter methylation test results by genomic DNA methylation array according to driver mutation VAF in adult-type diffuse gliomas

When pooled together, gliomas in which MGMT promoter methylation was determined by DNA methylation array using the Bady algorithm showed a VAF cutoff of 0.245 (Fig. 3A, Supplementary Figure S4A). The maximal proportion of cases with MGMT promoter methylation was 57% compared to 49% by pyrosequencing (Fig. 3B versus Fig. 2B). The shape of the curve at lower VAFs was also different for methylation array than for pyrosequencing. The likelihood of MGMT promoter methylation dropped to 31% at a VAF of 0.17, then surprisingly rose slightly with further decrease in VAF before dropping again at VAFs below 0.05 (Fig. 3B). Based on subset analyses, this paradoxical interval between VAFs 0.10–0.17 was driven by IDHwt GBMs, as GBM samples with low TERT promoter VAFs were at least as likely to be MGMT promoter methylation-positive as those with higher VAFs (Fig. 3 C-D, Supplementary Figure S4B). By array, 79 of 167 IDHwt GBM samples (47%) tested positive, compared to only 101 of 301 (34%) by pyrosequencing (p = 0.004, Fig. 3E).

Fig. 3
figure 3

MGMTpromoter methylation trends in DNA methylation array samples. (A) Frequency of positive results for all glioma array samples above and below the cutoff of VAF = 0.245. (B) Cumulative mean frequency of positive results for all glioma array samples as a function of VAF. (C) Frequency of positive results for GBM array samples above and below the cutoff of TERT VAF = 0.325. (D) Cumulative mean frequency of positive results in GBM array samples as a function of TERT VAF. (E) Frequency of positive results in GBM pyrosequencing samples using MGMT cutoff of 10.0% versus GBM array samples. (F) Frequency of positive results in GBM pyrosequencing samples using MGMT cutoff of 7.28% versus GBM array samples. (G) Frequency of positive results in GBM pyrosequencing samples with TERT VAF < 0.115, using MGMT cutoff of 7.28%, versus GBM array samples with TERT VAF < 0.325. (H) Frequency of positive results in GBM pyrosequencing samples with TERT VAF ≥ 0.115, using MGMT cutoff of 7.28%, versus GBM array samples with TERT VAF ≥ 0.325. (I) Frequency of positive results for IDHmut astrocytoma above and below the cutoff of IDH VAF = 0.325 (left), by methylation class match (center), and above and below the cutoff of classifier score = 0.955 (right). (J) Cumulative mean frequency of positive results in IDHmut astrocytoma array samples as a function of IDH VAF (Fisher’s exact test for panels A, C, E, F, G, H, I; GBM: IDH-wildtype glioblastoma, IDHmut astrocytoma: IDH-mutant astrocytoma)

Full size image

Further analyses suggested that variation in MGMT assay results between pyrosequencing and methylation array for IDHwt GBM may be driven by differing MGMT pyrosequencing cutoff values and divergent results in low-VAF GBMs specifically. The original work validating the Bady algorithm compared methylation array data to pyrosequencing data with a cutoff of 7.28% CpG site methylation for a positive result [41], in contrast to the generally established pyrosequencing cutoff of 10.0%. When IDHwt GBM pyrosequencing data were adjusted downward using the 7.28% cutoff, more cases (118 of 301, 39%) tested positive, which brought pyrosequencing closer to genomic methylation array percentages (p = 0.097, Fig. 3F). Subgroup analysis by TERT VAF (also using 7.28% CpG site methylation) showed that the divergent results were most evident in low-VAF GBMs. Only 3 of 14 GBM pyrosequencing samples (21%) below the pyrosequencing-specific VAF cutoff of 0.115 tested positive, versus 35 of 65 array samples (54%) below the array-specific VAF cutoff of 0.325 (p = 0.039, Fig. 3G). For GBM samples with VAF at or above the respective cutoffs, 115 of 287 pyrosequencing samples (40%) tested positive, compared to 44 of 102 array samples (43%) (p = 0.639, Fig. 3H). These results indicate that the assays are comparable in samples with higher tumor cellularity, but in cases of low cellularity, DNA methylation array may be a more reliable measure of MGMT status.

In contrast to IDHwt GBMs, IDHmut astrocytomas tested by methylation array showed less frequent positive results in samples with IDH VAF < 0.325 (Supplementary Figure S4C), similar to pyrosequencing (Fig. 3I-J versus Fig. 2I-J). As expected, IDHmut astrocytomas with high scores by the methylation classifier (cutoff score = 0.955 by Cutoff Finder, Supplementary Figure S4D) were more likely to be MGMT promoter methylation-positive (Fig. 3I). All 29 IDHmut oligodendrogliomas analyzed by DNA methylation array tested positive for MGMT promoter methylation, even with VAFs as low as 0.18 (Supplementary Table S3). Outcomes of Cutoff Finder categorical Fisher’s exact tests for pyrosequencing and DNA methylation array data are summarized in Supplementary Table S5.

Predictive value of driver mutation VAF in MGMT promoter testing and Tumor cellularity estimation

Simple linear regression showed that MGMT promoter methylation pyrosequencing scores positively correlated with driver mutation VAF among all gliomas (slope = 24.8, R2 = 0.040, p < 0.0001, Fig. 4A). Among the three major glioma subtypes, similar positive correlations were found for MGMT promoter methylation and TERT promoter VAF in IDHwt GBM (Fig. 4B), and for IDH VAF in both IDHmut astrocytoma (non-significant) and IDHmut oligodendroglioma (Fig. 4C-D). There was also a positive correlation between microscopically estimated tumor cellularity and cellularity calculated by doubling driver mutation VAF (slope = 0.246, R2 = 0.147, p < 0.0001, Fig. 4E), although there were numerous cases in which VAF did not align with microscopic estimates of tumor cellularity. To further investigate, differences between microscopy-based cellularity and VAF-based cellularity were plotted against VAF for each case. This showed a negative correlation (slope = − 126.7, R2 = 0.579, p < 0.0001, Fig. 4F), indicating that when VAF was low, microscopy-based tumor cellularity estimates tended to be higher than what was calculated by VAF.

Fig. 4
figure 4

Driver mutation VAF, MGMTpromoter methylation scores, and tumor cellularity. (A) Linear regression of MGMT promoter pyrosequencing score versus driver mutation VAF for all glioma samples. (B) Linear regression of MGMT promoter pyrosequencing score versus TERT promoter mutation VAF for GBM. (C) Linear regression of MGMT promoter pyrosequencing score versus IDH mutation VAF for IDHmut astrocytoma. (D) Linear regression of MGMT promoter pyrosequencing score versus IDH mutation VAF for IDHmut oligodendroglioma. (E) Linear regression of microscopically estimated cellularity versus cellularity calculated from driver mutation VAF (2×VAF×100%) for all glioma samples. (F) Differences between microscopically estimated cellularity and cellularity calculated from VAF (Y-axis) plotted as a function of VAF (X-axis), for all glioma samples. (G) TERT promoter mutation VAF by ddPCR in high versus low cellularity areas of GBM tissue samples. (H) MGMT promoter methylation score by ddPCR in high versus low cellularity areas of GBM tissue samples (GBM: IDH-wildtype glioblastoma, IDHmut astrocytoma: IDH-mutant astrocytoma, IDHmut oligodendroglioma: IDH-mutant and 1p/19q co-deleted oligodendroglioma, pyroseq: pyrosequencing, ddPCR: droplet digital PCR)

Full size image

To further assess the relationship between tumor cellularity, driver VAF, and MGMT assay outcomes, we performed additional analyses on 5 IDHwt GBM samples. For each sample, areas of high and low tumor cellularity were marked via light microscopy of H&E slides. TERT VAF and MGMT promoter methylation were measured in parallel by ddPCR on tissue from each area (Table 2). In 4 out of 5 cases, both TERT VAF (Fig. 4G) and MGMT promoter methylation (Fig. 4H) were lower in less cellular portions of tumors. Although the MGMT promoter methylation scores dropped in paucicellular areas, they remained above the ddPCR cutoff for a positive result (4.0%). In one case, MGMT promoter methylation increased (from 67.6 to 75.5%) in parallel with TERT VAF (from 0.386 to 0.440) although the microscopic estimate of cellularity fell (from 60 to 40%). This further reinforces the suggestion that driver mutation VAF may be a more objective and accurate measure of tumor purity than visually estimated cellularity.

False negative MGMT promoter pyrosequencing results in cases with low driver mutation VAF

When glioma driver mutation VAF is low, MGMT promoter methylation scores by pyrosequencing also tend to be lower, and negative test results are more frequent. Based on the divergence in MGMT promoter pyrosequencing outcomes with variable VAF, we investigated whether original pyrosequencing results were false-negatives by re-testing samples with DNA methylation array and ddPCR. These alternative methods were chosen based on the sensitivity of DNA methylation array in low-VAF IDHwt GBM samples (Fig. 3C-G), and on the ability of ddPCR to identify low-abundance targets [42]. A cohort of 6 IDHwt GBM samples with TERT VAF ≤ 0.10 (lower than VAF cutoff values identified by both Cutoff Finder and regression, Fig. 2E-H) and a sex- and age-matched control cohort with TERT VAF ≥ 0.25 (higher than both VAF cutoff values) were re-tested, using the same FFPE tissue blocks as initial pyrosequencing and NGS assays. Results are illustrated in Fig. 5A (pyrosequencing vs. array) and Fig. 5B (pyrosequencing vs. ddPCR) and detailed in Table 3. In the low-VAF cohort, samples from patients #2 and #3 both had MGMT promoter methylation status re-classified on re-testing. Sample #2 was called positive by array and equivocal by ddPCR, while the opposite held true for sample #3. Both samples had pyrosequencing promoter methylation levels near the positive cutoff, low TERT VAF, and favorable TMZ responses (survival of 60.3 and 20.8 months, respectively). No cases from the high-VAF cohort had their MGMT promoter methylation status changed on re-testing. DNA was insufficient for ddPCR in 3 cases (2 from the low-VAF cohort and 1 from the high-VAF cohort). These results demonstrate that DNA methylation array analysis and ddPCR can identify false-negative MGMT pyrosequencing results in low-VAF samples.

Fig. 5
figure 5

False negative results in IDH-wildtype glioblastoma with lowTERTVAF. (A) MGMT promoter methylation results for 12 GBM samples (6 with TERT VAF ≤ 0.10, 6 with TERT VAF ≥ 0.25) comparing initial pyrosequencing methylation scores (left Y-axis, cutoff for positive = 10.0%, horizontal solid red line) to results on re-testing with DNA methylation array (right Y-axis). (B) MGMT promoter methylation results for the same 12 GBM samples comparing initial pyrosequencing methylation levels to results on re-testing with ddPCR. (GBM: IDH-wildtype glioblastoma, pos: positive, equiv: equivocal, neg: negative, QNS: quality/quantity of DNA not sufficient for reliable test result, ddPCR: droplet digital PCR)

Full size image
Table 3 Re-testing pyrosequencing samples with DNA methylation array and droplet digital PCR
Full size table

Discussion

MGMT promoter methylation testing provides vital information for adult-type diffuse glioma therapy planning [14, 43,44,45,46]. Several phase 3 trials utilizing TMZ for newly diagnosed IDHwt GBM showed that the median overall survival of patients with MGMT promoter-methylated IDHwt GBM was 21.2–23.2 months versus 14.0-15.3 months for patients with unmethylated IDHwt GBM [1, 14, 47, 48]. Many ongoing trials therefore use MGMT unmethylated status as an inclusion criterion for testing therapeutic regimens that do not involve TMZ [44]. TMZ might not even be prescribed for patients with low Karnofsky performance status scores and MGMT-unmethylated tumors, as the benefit-to-toxicity ratio might be too low. Conversely, MGMT promoter methylation is often used as an inclusion criterion for glioma trials involving TMZ-sensitizing strategies. Elderly patients, whose tolerance for TMZ is generally lower, may still receive TMZ if their IDHwt GBMs have MGMT promoter methylation [46, 49]. Likewise, a patient taking TMZ may continue therapy even if radiology suggests tumor progression, so long as the glioma has MGMT promoter methylation. Thus, accurate MGMT promoter methylation testing is essential for patient management.

MGMT promoter methylation does not occur in non-neoplastic cells, meaning they can upregulate MGMT gene expression and be less sensitive to TMZ than glioma cells with MGMT promoter methylation [50]. Since infiltrative glioma samples are always an admixture of tumor and non-tumor cells, and MGMT testing is done on bulk tissue samples, an excess of non-tumor cells can mask a positive signal from glioma cells (Fig. 6). Accurate results therefore depend on adequate tumor cellularity. 70% tumor cellularity is the usual benchmark for specimen acceptance, but light microscopic estimates are highly subjective. Moreover, it is unusual for specimens to be rejected for testing even if they do not reach the 70% benchmark. Thus, false-negative results are possible in some instances, although to date the frequency of such cases and objective ways of catching them have been unclear.

Fig. 6
figure 6

Central hypothesis. MGMT promoter methylation is pathologic, and occurs only in tumor cells. Cellular glioma samples are rich in DNA from tumor cells, whereas paucicellular glioma samples contain a large fraction of DNA from non-tumor cells, which can “dilute” positive methylation signals from tumor cell DNA, leading to false-negative results

Full size image

Here, we show that glioma driver mutation VAF, which is included in most NGS reports, can serve as a useful metric for judging the reliability of MGMT promoter methylation pyrosequencing results, more so than microscopic cellularity estimates. Based on these data, driver VAF should be considered when interpreting pyrosequencing results for clinical decision-making, and when considering alternative testing methods with higher sensitivity, such as ddPCR. We also show that, despite the widely used minimum cellularity of 70%, results may be considered reliable even when IDHwt GBM cellularity is much lower (35–40%). In contrast, cellularity of 60–80% appears to be needed for IDH-mutant astrocytoma and IDH-mutant oligodendroglioma. The different VAF/cellularity cutoffs for tumor types could be related to differences in intratumoral heterogeneity of MGMT promoter methylation in IDH-mutant versus IDH-wildtype gliomas [51, 52]. Tumors with greater heterogeneity would require higher purity to detect positive results on bulk analyses. The lower cutoff for IDHwt GBM may also be due to more genomic instability, which might cause small copy number alterations involving the TERT promoter locus in a subset of tumors [53]. Even so, previous work, including our own, shows that TERT VAF is a reliable tumor purity marker in the majority of IDHwt GBM cases [35, 36]. The low-VAF cases ultimately identified as false negatives had MGMT pyrosequencing levels near the positive cutoff, and patients responded well to TMZ, as would be expected. This might help explain the unexpected benefit of TMZ reported in a subset of patients whose gliomas tested negative for MGMT promoter methylation [14, 54]. Based on the fraction of IDHwt GBM samples in our cohort falling below the VAF/cellularity cutoffs identified by Cutoff Finder and regression (4.3% and 12.6%, respectively) and the known frequency of MGMT promoter methylation in IDH-wildtype GBM of ~ 35% [55], we estimate that 1.5-4.4% of IDHwt GBM MGMT pyrosequencing samples may have false-negative results, potentially impacting TMZ treatment decisions.

Array-based DNA methylation profiling is revolutionizing how brain tumors are diagnosed by recognizing distinctive DNA methylation “fingerprints” [27]. Among multiple advantages is its ability to interrogate MGMT promoter methylation status without the need for a separate test [20, 41, 56]. Our current data suggest that it is more sensitive than pyrosequencing. Our internal validation studies also support this conclusion. Validation runs of the methylation array at our institution identified 4 cases as positive which were negative by pyrosequencing. Similarly, ddPCR can detect MGMT promoter methylation in paucicellular tumor samples due to low background signals [42]. The fact that ddPCR successfully detected MGMT promoter methylation in paucicellular areas (Table 2) also demonstrates its sensitivity.

The CpG sites analyzed by each assay may be a factor in variable sensitivity. The 2 sites analyzed by the Bady algorithm from array data and by our ddPCR assay have a particularly strong association with MGMT transcriptional repression and TMZ sensitivity [41]. In contrast, the Qiagen™ pyrosequencing assay utilized at Northwestern interrogates 4 CpG sites in Exon 1 of the MGMT promoter (see methods), none of which correspond to the 2 utilized in the Bady algorithm and our ddPCR assay. CpG sites for pyrosequencing assays are often selected due to correlation with CpG methylation levels across the entire promoter region rather than their individual importance in MGMT transcriptional regulation [57]. More widespread use of genomic DNA methylation arrays and/or ddPCR in MGMT promoter methylation assessment may be beneficial in glioma diagnostic testing, due to the apparent sensitivity of both tests.

One limitation of this study is the retrospective nature of the cohort. Another is the lack of cases in which both pyrosequencing and methylation array were done. A third is the low number of IDHwt GBMs with both MGMT unmethylation and low VAF, which limited the statistical power of analyses of variation in positivity rates and variation in methylation level. Finally, this approach obviously cannot work in the minority of IDHwt GBMs lacking TERT promoter mutation.

Conclusions

In summary, our findings demonstrate the value of driver mutation VAF as a quality assurance tool in MGMT promoter methylation testing of adult-type diffuse gliomas. VAF is particularly useful for identifying cases at risk of being falsely negative. This also provides another rationale for routine NGS of glioma samples (or targeted testing for driver mutations via ddPCR) along with MGMT testing, as mutation data not only refines the histopathologic diagnosis, but also indicates how reliable the MGMT result is likely to be. Future studies could focus on the recently recognized role of MGMT promoter methylation intratumoral heterogeneity in glioma biology and response to therapy [51, 52]. The roles of individual CpG sites within the MGMT promoter, and of other transcriptional regulatory elements such as the MGMT gene enhancer [58] could also be studied further.

Data Availability

De-identified case data is available in Supplementary Tables S2 and S3. Additional materials are available from the corresponding author upon request.

Abbreviations

ddPCR:

Droplet digital polymerase chain reaction

FFPE:

Formalin-fixed, paraffin-embedded

H&E:

Hematoxylin & Eosin

IDH:

Isocitrate dehydrogenase

IDHmut astrocytoma:

IDH-mutant astrocytoma

IDHmut oligodendroglioma:

Oligodendroglioma, IDH-mutant and 1p/19q co-deleted

IDHwt GBM:

IDH-wildtype glioblastoma

MGMT:

O6-methylguanine-DNA-methyltransferase

NGS:

Next generation sequencing

NSTB:

Nervous system tumor bank

OCAV3:

Oncomine comprehensive NGS assay, version 3

QNS:

Quantity/quality not sufficient

TMZ:

Temozolomide

VAF:

Variant allelic frequency

WHO:

World health organization

References

  1. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, Belanger K, Brandes AA, Marosi C, Bogdahn Uet al et al (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352:987–996. https://doi.org/10.1056/NEJMoa043330

    Article  CAS  PubMed  Google Scholar 

  2. SongTao Q, Lei Y, Si G, YanQing D, HuiXia H, XueLin Z, LanXiao W, Fei Y (2012) IDH mutations predict longer survival and response to temozolomide in secondary glioblastoma. Cancer Sci 103:269–273. https://doi.org/10.1111/j.1349-7006.2011.02134.x

    Article  CAS  PubMed  Google Scholar 

  3. van den Bent MJ, Baumert B, Erridge SC, Vogelbaum MA, Nowak AK, Sanson M, Brandes AA, Clement PM, Baurain JF, Mason WP al (2017) Interim results from the CATNON trial (EORTC study 26053 – 22054) of treatment with concurrent and adjuvant temozolomide for 1p/19q non-co-deleted anaplastic glioma: a phase 3, randomised, open-label intergroup study. Lancet 390:1645–1653. https://doi.org/10.1016/S0140-6736(17)31442-3

    Article  PubMed  PubMed Central  Google Scholar 

  4. Weller M, van den Bent M, Preusser M, Le Rhun E, Tonn JC, Minniti G, Bendszus M, Balana C, Chinot O, Dirven Let al et al (2021) EANO guidelines on the diagnosis and treatment of diffuse gliomas of adulthood. Nat Rev Clin Oncol 18:170–186. https://doi.org/10.1038/s41571-020-00447-z

    Article  PubMed  Google Scholar 

  5. Friedman HS, Kerby T, Calvert H (2000) Temozolomide and treatment of malignant glioma. Clin Cancer Res 6:2585–2597

    CAS  PubMed  Google Scholar 

  6. Lee SY (2016) Temozolomide resistance in Glioblastoma Multiforme. Genes Dis 3:198–210. https://doi.org/10.1016/j.gendis.2016.04.007

    Article  PubMed  PubMed Central  Google Scholar 

  7. Day RS 3rd, Ziolkowski CH, Scudiero DA, Meyer SA, Lubiniecki AS, Girardi AJ, Galloway SM, Bynum GD (1980) Defective repair of alkylated DNA by human tumour and SV40-transformed human cell strains. Nature 288:724–727. https://doi.org/10.1038/288724a0

    Article  CAS  PubMed  Google Scholar 

  8. Yarosh DB, Foote RS, Mitra S, Day RS 3rd (1983) Repair of O6-methylguanine in DNA by demethylation is lacking in mer- human Tumor cell strains. Carcinogenesis 4:199–205. https://doi.org/10.1093/carcin/4.2.199

    Article  CAS  PubMed  Google Scholar 

  9. Kaina B, Christmann M, Naumann S, Roos WP (2007) MGMT: key node in the battle against genotoxicity, carcinogenicity and apoptosis induced by alkylating agents. DNA Repair (Amst) 6:1079–1099. https://doi.org/10.1016/j.dnarep.2007.03.008

    Article  CAS  PubMed  Google Scholar 

  10. Gardner E, Rydberg B, Karran P, Ponder BA (1991) Localization of the human O6-methylguanine-DNA methyltransferase gene to chromosome 10q24.33-qter. Genomics 11:475–476 Doi 10.1016/0888–7543(91)90162-8

    Article  CAS  PubMed  Google Scholar 

  11. Qian XC, Brent TP (1997) Methylation hot spots in the 5’ flanking region denote silencing of the O6-methylguanine-DNA methyltransferase gene. Cancer Res 57:3672–3677

    CAS  PubMed  Google Scholar 

  12. Watts GS, Pieper RO, Costello JF, Peng YM, Dalton WS, Futscher BW (1997) Methylation of discrete regions of the O6-methylguanine DNA methyltransferase (MGMT) CpG island is associated with heterochromatinization of the MGMT transcription start site and silencing of the gene. Mol Cell Biol 17:5612–5619. https://doi.org/10.1128/MCB.17.9.5612

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  13. Esteller M, Hamilton SR, Burger PC, Baylin SB, Herman JG (1999) Inactivation of the DNA repair gene O6-methylguanine-DNA methyltransferase by promoter hypermethylation is a common event in primary human neoplasia. Cancer Res 59:793–797

    CAS  PubMed  Google Scholar 

  14. Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, Kros JM, Hainfellner JA, Mason W, Mariani Let al et al (2005) MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 352:997–1003. https://doi.org/10.1056/NEJMoa043331

    Article  CAS  PubMed  Google Scholar 

  15. Rivera AL, Pelloski CE, Gilbert MR, Colman H, De La Cruz C, Sulman EP, Bekele BN, Aldape KD (2010) MGMT promoter methylation is predictive of response to radiotherapy and prognostic in the absence of adjuvant alkylating chemotherapy for glioblastoma. Neuro Oncol 12:116–121. https://doi.org/10.1093/neuonc/nop020

    Article  CAS  PubMed  Google Scholar 

  16. Estival A, Sanz C, Ramirez JL, Velarde JM, Domenech M, Carrato C, de Las Penas R, Gil-Gil M, Sepulveda J Armengol R (2019) pyrosequencing versus methylation-specific PCR for assessment of MGMT methylation in Tumor and blood samples of glioblastoma patients. Sci Rep 9: 11125 https://doi.org/10.1038/s41598-019-47642-2

  17. Xie H, Tubbs R, Yang B (2015) Detection of MGMT promoter methylation in glioblastoma using pyrosequencing. Int J Clin Exp Pathol 8:1790–1796

    PubMed  PubMed Central  Google Scholar 

  18. Quillien V, Lavenu A, Karayan-Tapon L, Carpentier C, Labussiere M, Lesimple T, Chinot O, Wager M, Honnorat J, Saikali S et al (2012) Comparative assessment of 5 methods (methylation-specific polymerase chain reaction, MethyLight, pyrosequencing, methylation-sensitive high-resolution melting, and immunohistochemistry) to analyze O6-methylguanine-DNA-methyltranferase in a series of 100 glioblastoma patients. Cancer 118: 4201–4211 https://doi.org/10.1002/cncr.27392

  19. Quillien V, Lavenu A, Ducray F, Joly MO, Chinot O, Fina F, Sanson M, Carpentier C, Karayan-Tapon L, Rivet P et al (2016) Validation of the high-performance of pyrosequencing for clinical MGMT testing on a cohort of glioblastoma patients from a prospective dedicated multicentric trial. Oncotarget 7: 61916–61929

  20. Bady P, Delorenzi M, Hegi ME (2016) Sensitivity analysis of the MGMT-STP27 model and impact of genetic and epigenetic context to predict the MGMT methylation status in Gliomas and other tumors. J Mol Diagn 18:350–361. https://doi.org/10.1016/j.jmoldx.2015.11.009

    Article  CAS  PubMed  Google Scholar 

  21. Santana-Santos L, Kam KL, Dittmann D, De Vito S, McCord M, Jamshidi P, Fowler H, Wang X, Aalsburg AM, Brat DJ et al (2022) Validation of whole genome methylation profiling classifier for Central Nervous System tumors. J Mol Diagn. https://doi.org/10.1016/j.jmoldx.2022.04.009

    Article  PubMed  Google Scholar 

  22. Barault L, Amatu A, Bleeker FE, Moutinho C, Falcomata C, Fiano V, Cassingena A, Siravegna G, Milione M, Cassoni P al (2015) Digital PCR quantification of MGMT methylation refines prediction of clinical benefit from alkylating agents in glioblastoma and metastatic Colorectal cancer. Ann Oncol 26:1994–1999. https://doi.org/10.1093/annonc/mdv272

    Article  CAS  PubMed  Google Scholar 

  23. Olmedillas-Lopez S, Olivera-Salazar R, Garcia-Arranz M, Garcia-Olmo D (2022) Current and emerging applications of Droplet Digital PCR in Oncology: an updated review. Mol Diagn Ther 26:61–87. https://doi.org/10.1007/s40291-021-00562-2

    Article  CAS  PubMed  Google Scholar 

  24. Patel SBN, Panchal K, Trivedi P, Trivedi T (2021) MGMT promoter methylation status using droplet digital PCR in glioma patients: association with treatment response and survival. IJARESM 9:324–333

    Google Scholar 

  25. Hosoya T, Takahashi M, Honda-Kitahara M, Miyakita Y, Ohno M, Yanagisawa S, Omura T, Kawauchi D, Tamura Y Kikuchi M (2022) MGMT gene promoter methylation by pyrosequencing method correlates volumetric response and neurological status in IDH wild-type glioblastomas. J Neurooncol 157: 561–571 https://doi.org/10.1007/s11060-022-03999-5

  26. Gurrieri L, De Carlo E, Gerratana L, De Maglio G, Macerelli M, Pisa FE, Masiero E, Aprile G, Follador A, Puglisi F al (2018) MGMT pyrosequencing-based cut-off methylation level and clinical outcome in patients with Glioblastoma Multiforme. Future Oncol 14:699–707. https://doi.org/10.2217/fon-2017-0437

    Article  CAS  PubMed  Google Scholar 

  27. Capper D, Jones DTW, Sill M, Hovestadt V, Schrimpf D, Sturm D, Koelsche C, Sahm F, Chavez L, Reuss DE et al (2018) DNA methylation-based classification of central nervous system tumours. Nature 555: 469–474 https://doi.org/10.1038/nature26000

  28. Capper D, Stichel D, Sahm F, Jones DTW, Schrimpf D, Sill M, Schmid S, Hovestadt V, Reuss DE, Koelsche C et al (2018) Practical implementation of DNA methylation and copy-number-based CNS tumor diagnostics: the Heidelberg experience. Acta Neuropathol 136: 181–210 https://doi.org/10.1007/s00401-018-1879-y

  29. Stichel D, Ebrahimi A, Reuss D, Schrimpf D, Ono T, Shirahata M, Reifenberger G, Weller M, Hanggi D, Wick W et al (2018) Distribution of EGFR amplification, combined chromosome 7 gain and chromosome 10 loss, and TERT promoter mutation in brain tumors and their potential for the reclassification of IDHwt astrocytoma to glioblastoma. Acta Neuropathol 136: 793–803 https://doi.org/10.1007/s00401-018-1905-0

  30. Killela PJ, Reitman ZJ, Jiao Y, Bettegowda C, Agrawal N, Diaz LA Jr., Friedman AH, Friedman H, Gallia GL, Giovanella BC al (2013) TERT promoter mutations occur frequently in gliomas and a subset of tumors derived from cells with low rates of self-renewal. Proc Natl Acad Sci U S A 110:6021–6026 Doi 1303607110 [pii]. https://doi.org/10.1073/pnas.1303607110

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  31. Louis DN, Perry A, Wesseling P, Brat DJ, Cree IA, Figarella-Branger D, Hawkins C, Ng HK, Pfister SM, Reifenberger Get al et al (2021) The 2021 WHO classification of tumors of the Central Nervous System: a summary. Neuro Oncol 23:1231–1251. https://doi.org/10.1093/neuonc/noab106

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  32. Mosrati MA, Malmstrom A, Lysiak M, Krysztofiak A, Hallbeck M, Milos P, Hallbeck AL, Bratthall C, Strandeus M Stenmark-Askmalm M (2015) TERT promoter mutations and polymorphisms as prognostic factors in primary glioblastoma. Oncotarget 6: 16663–16673 https://doi.org/10.18632/oncotarget.4389

  33. Pietrak B, Zhao H, Qi H, Quinn C, Gao E, Boyer JG, Concha N, Brown K, Duraiswami C Wooster R (2011) a tale of two subunits: how the neomorphic R132H IDH1 mutation enhances production of alphaHG. Biochemistry 50: 4804–4812 https://doi.org/10.1021/bi200499m

  34. Duncan CG, Barwick BG, Jin G, Rago C, Kapoor-Vazirani P, Powell DR, Chi JT, Bigner DD, Vertino PM, Yan H (2012) A heterozygous IDH1R132H/WT mutation induces genome-wide alterations in DNA methylation. Genome Res 22:2339–2355. https://doi.org/10.1101/gr.132738.111

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Wu Z, Abdullaev Z, Pratt D, Chung HJ, Skarshaug S, Zgonc V, Perry C, Pack S, Saidkhodjaeva L, Nagaraj Set al et al (2022) Impact of the methylation classifier and ancillary methods on CNS Tumor diagnostics. Neuro Oncol 24:571–581. https://doi.org/10.1093/neuonc/noab227

    Article  CAS  PubMed  Google Scholar 

  36. Jamshidi P, McCord M, Galbraith K, Santana-Santos L, Jennings LJ, Snuderl M, Horbinski C (2023) Variant allelic frequency of driver mutations predicts success of genomic DNA methylation classification in central nervous system tumors. Acta Neuropathol 145:365–367. https://doi.org/10.1007/s00401-023-02542-8

    Article  CAS  PubMed  Google Scholar 

  37. Rosas-Alonso R, Colmenarejo-Fernandez J, Pernia O, Rodriguez-Antolin C, Esteban I, Ghanem I, Sanchez-Cabrero D, Losantos-Garcia I, Palacios-Zambrano S, Moreno-Bueno Get al et al (2021) Clinical validation of a novel quantitative assay for the detection of MGMT methylation in glioblastoma patients. Clin Epigenetics 13:52. https://doi.org/10.1186/s13148-021-01044-2

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  38. Adachi JI, Shirahata M, Suzuki T, Mishima K, Uchida E, Sasaki A, Nishikawa R (2021) Droplet digital PCR assay for detecting TERT promoter mutations in patients with glioma. Brain Tumor Pathol 38:201–209. https://doi.org/10.1007/s10014-021-00403-4

    Article  CAS  PubMed  Google Scholar 

  39. Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, Irizarry RA (2014) Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics 30:1363–1369. https://doi.org/10.1093/bioinformatics/btu049

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Budczies J, Klauschen F, Sinn BV, Gyorffy B, Schmitt WD, Darb-Esfahani S, Denkert C (2012) Cutoff finder: a comprehensive and straightforward web application enabling rapid biomarker cutoff optimization. PLoS ONE 7:e51862. https://doi.org/10.1371/journal.pone.0051862

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Bady P, Sciuscio D, Diserens AC, Bloch J, van den Bent MJ, Marosi C, Dietrich PY, Weller M, Mariani L, Heppner FL al (2012) MGMT methylation analysis of glioblastoma on the Infinium methylation BeadChip identifies two distinct CpG regions associated with gene silencing and outcome, yielding a prediction model for comparisons across datasets, Tumor grades, and CIMP-status. Acta Neuropathol 124:547–560. https://doi.org/10.1007/s00401-012-1016-2

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Jennings LJ, George D, Czech J, Yu M, Joseph L (2014) Detection and quantification of BCR-ABL1 fusion transcripts by droplet digital PCR. J Mol Diagn 16:174–179. https://doi.org/10.1016/j.jmoldx.2013.10.007

    Article  CAS  PubMed  Google Scholar 

  43. Wick W, Platten M, Meisner C, Felsberg J, Tabatabai G, Simon M, Nikkhah G, Papsdorf K, Steinbach JP, Sabel Met al et al (2012) Temozolomide chemotherapy alone versus radiotherapy alone for malignant astrocytoma in the elderly: the NOA-08 randomised, phase 3 trial. Lancet Oncol 13:707–715. https://doi.org/10.1016/S1470-2045(12)70164-X

    Article  CAS  PubMed  Google Scholar 

  44. Hegi ME, Stupp R (2015) Withholding temozolomide in glioblastoma patients with unmethylated MGMT promoter–still a dilemma? Neuro Oncol 17:1425–1427. https://doi.org/10.1093/neuonc/nov198

    Article  PubMed  PubMed Central  Google Scholar 

  45. Bady P, Kurscheid S, Delorenzi M, Gorlia T, van den Bent MJ, Hoang-Xuan K, Vauleon E, Gijtenbeek A, Enting R, Thiessen B et al (2018) The DNA methylome of DDR genes and benefit from RT or TMZ in IDH mutant low-grade glioma treated in EORTC 22033. Acta Neuropathol 135: 601–615

  46. Malmstrom A, Gronberg BH, Marosi C, Stupp R, Frappaz D, Schultz H, Abacioglu U, Tavelin B, Lhermitte B Hegi ME (2012) Temozolomide versus standard 6-week radiotherapy versus hypofractionated radiotherapy in patients older than 60 years with glioblastoma: the nordic randomised, phase 3 trial. Lancet Oncol 13: 916–926

  47. Gilbert MR, Wang M, Aldape KD, Stupp R, Hegi ME, Jaeckle KA, Armstrong TS, Wefel JS, Won M, Blumenthal DT al (2013) Dose-dense temozolomide for newly diagnosed glioblastoma: a Randomized Phase III Clinical Trial. J Clin Oncol 31:4085–4091. https://doi.org/10.1200/jco.2013.49.6968

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  48. Gilbert MR, Dignam JJ, Armstrong TS, Wefel JS, Blumenthal DT, Vogelbaum MA, Colman H, Chakravarti A, Pugh S, Won Met al et al (2014) A Randomized Trial of Bevacizumab for newly diagnosed Glioblastoma. N Engl J Med 370:699–708. https://doi.org/10.1056/NEJMoa1308573

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  49. Perry JR, Laperriere N, O’Callaghan CJ, Brandes AA, Menten J, Phillips C, Fay M, Nishikawa R, Cairncross JG, Roa Wet al et al (2017) Short-course Radiation plus Temozolomide in Elderly patients with Glioblastoma. N Engl J Med 376:1027–1037

    Article  CAS  PubMed  Google Scholar 

  50. Kitange GJ, Carlson BL, Schroeder MA, Grogan PT, Lamont JD, Decker PA, Wu W, James CD, Sarkaria JN (2009) Induction of MGMT expression is associated with temozolomide resistance in glioblastoma xenografts. Neuro Oncol 11:281–291. https://doi.org/10.1215/15228517-2008-090

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  51. Gempt J, Withake F, Aftahy AK, Meyer HS, Barz M, Delbridge C, Liesche-Starnecker F, Prokop G, Pfarr N, Schlegel J al (2022) Methylation subgroup and molecular heterogeneity is a hallmark of glioblastoma: implications for biopsy targeting, classification and therapy. ESMO Open 7:100566. https://doi.org/10.1016/j.esmoop.2022.100566

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Wenger A, Ferreyra Vega S, Kling T, Bontell TO, Jakola AS, Caren H (2019) Intratumor DNA methylation heterogeneity in glioblastoma: implications for DNA methylation-based classification. Neuro Oncol 21:616–627. https://doi.org/10.1093/neuonc/noz011

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  53. Richardson TE, Sathe AA, Xing C, Mirchia K, Viapiano MS, Snuderl M, Abdullah KG, Hatanpaa KJ, Walker JM (2021) Molecular signatures of chromosomal instability correlate with Copy number variation patterns and patient outcome in IDH-Mutant and IDH-Wildtype Astrocytomas. J Neuropathol Exp Neurol 80:354–365. https://doi.org/10.1093/jnen/nlab008

    Article  CAS  PubMed  Google Scholar 

  54. Stupp R, Hegi ME, Mason WP, van den Bent MJ, Taphoorn MJ, Janzer RC, Ludwin SK, Allgeier A, Fisher B, Belanger K al (2009) Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial. Lancet Oncol 10:459–466. https://doi.org/10.1016/S1470-2045(09)70025-7

    Article  CAS  PubMed  Google Scholar 

  55. Horbinski C, McCortney K, Stupp R (2021) MGMT promoter methylation is associated with patient age and 1p/19q status in IDH-mutant gliomas. Neuro Oncol 23:858–860. https://doi.org/10.1093/neuonc/noab039

    Article  PubMed  PubMed Central  Google Scholar 

  56. Bell EH, Zhang P, Fisher BJ, Macdonald DR, McElroy JP, Lesser GJ, Fleming J, Chakraborty AR, Liu Z, Becker AP et al (2018) Association of MGMT Promoter Methylation Status With Survival Outcomes in Patients With High-Risk Glioma Treated With Radiotherapy and Temozolomide: An Analysis From the NRG Oncology/RTOG 0424 Trial. JAMA Oncol 4: 1405–1409

  57. Mikeska T, Bock C, El-Maarri O, Hubner A, Ehrentraut D, Schramm J, Felsberg J, Kahl P, Buttner R, Pietsch T et al (2007) Optimization of quantitative MGMT promoter methylation analysis using pyrosequencing and combined bisulfite restriction analysis. J Mol Diagn 9: 368–381 https://doi.org/10.2353/jmoldx.2007.060167

  58. Leske H, Camenisch Gross U, Hofer S, Neidert MC, Leske S, Weller M, Lehnick D, Rushing EJ (2023) MGMT methylation pattern of long-term and short-term survivors of Glioblastoma reveals CpGs of the enhancer region to be of high prognostic value. Acta Neuropathol Commun 11:139. https://doi.org/10.1186/s40478-023-01622-w

    Article  CAS  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

The authors thank the faculty members of the Neuropathology Division and the faculty and staff of the Molecular Pathology Division at Northwestern Memorial Hospital for their assistance in identifying cases and facilitating data collection. We also thank Kirsten Bell-Burdett for consultations on statistical analyses.

Funding

This work was supported by NIH grants F32CA264883 (MM), R01NS117104, R01NS118039 (CH), and the P50CA221747 Northwestern University SPORE in brain cancer.

Author information

Authors and Affiliations

  1. Department of Pathology, Northwestern University Feinberg School of Medicine, Chicago, USA

    Matthew McCord, Pouya Jamshidi, Lucas Santana-Santos, Erica Vormittag-Nocito, David Dittman, Stephanie Parker, Joseph Baczkowski, Lawrence Jennings & Craig Horbinski

  2. Department of Neurological Surgery, Northwestern University Feinberg School of Medicine, Chicago, USA

    Vineeth Thirunavu, Jordain Walshon, Kathleen McCortney, Roger Stupp, Amy B. Heimberger & Craig Horbinski

  3. Department of Pathology, New York University Langone Health, New York, USA

    Kristyn Galbraith & Matija Snuderl

  4. Department of Preventive Medicine, Northwestern University Feinberg School of Medicine, Chicago, USA

    Hui Zhang

  5. Department of Neurology, Northwestern University Feinberg School of Medicine, Chicago, USA

    Rimas V. Lukas, Roger Stupp, Karan Dixit & Priya Kumthekar

  6. Lou and Jean Malnati Brain Tumor Institute of the Robert H. Lurie Comprehensive Cancer Center, Northwestern University, Chicago, USA

    Rimas V. Lukas, Roger Stupp, Karan Dixit, Priya Kumthekar, Amy B. Heimberger & Craig Horbinski

  7. Feinberg School of Medicine, Northwestern University, 303 E Superior Street, 6-518, Chicago, IL, 60611, USA

    Craig Horbinski

Authors
  1. Matthew McCord
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Pouya Jamshidi
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Vineeth Thirunavu
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Lucas Santana-Santos
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Erica Vormittag-Nocito
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. David Dittman
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Stephanie Parker
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Joseph Baczkowski
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Lawrence Jennings
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Jordain Walshon
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Kathleen McCortney
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Kristyn Galbraith
    View author publications

    You can also search for this author in PubMed Google Scholar

  13. Hui Zhang
    View author publications

    You can also search for this author in PubMed Google Scholar

  14. Rimas V. Lukas
    View author publications

    You can also search for this author in PubMed Google Scholar

  15. Roger Stupp
    View author publications

    You can also search for this author in PubMed Google Scholar

  16. Karan Dixit
    View author publications

    You can also search for this author in PubMed Google Scholar

  17. Priya Kumthekar
    View author publications

    You can also search for this author in PubMed Google Scholar

  18. Amy B. Heimberger
    View author publications

    You can also search for this author in PubMed Google Scholar

  19. Matija Snuderl
    View author publications

    You can also search for this author in PubMed Google Scholar

  20. Craig Horbinski
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

CH conceptualized the project. MM, PJ, JW, and KM, assembled the patient cohort. MM, PJ, VT, JW, and KM reviewed cases and recorded data. LS, EV, DD, SP, JB, and LJ supervised diagnostic testing and interpreted results. MM analyzed data and generated tables and figures. MM drafted the manuscript. PJ, VT, LS, EV, RL, HZ, RS, PK, ABH, and CH assisted with data interpretation. All authors reviewed and edited the final manuscript.

Corresponding author

Correspondence to Craig Horbinski.

Ethics declarations

Ethics approval and consent to participate

This research was carried out as a quality improvement study, in accordance with the Declaration of Helsinki. All patient data was de-identified prior to analysis. For samples from the Northwestern University Nervous System Tumor Bank, all patients provided written informed consent for research, under a Northwestern University Institutional Review Board-approved protocol (STU00095863).

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

McCord, M., Jamshidi, P., Thirunavu, V. et al. Variant allelic frequencies of driver mutations can identify gliomas with potentially false-negative MGMT promoter methylation results. acta neuropathol commun 11, 175 (2023). https://doi.org/10.1186/s40478-023-01680-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s40478-023-01680-0

Keywords

Acta Neuropathologica Communications

ISSN: 2051-5960

Contact us