Optimal qMSP cutoff value for MGMT promoter methylation in glioblastoma and its validation for clinical significance

Abstract

Background

Glioblastoma (GBM) is the most common and aggressive primary brain tumor, with limited survival despite multimodal treatment strategies. O6-Methylguanine-DNA Methyltransferase (MGMT) promoter methylation is a well-established predictive biomarker for response to temozolomide (TMZ) therapy. However, determining an optimal quantitative methylation-specific PCR (qMSP) cut-off value remains a challenge in clinical practice.

Objective

This study aimed to establish an optimal qMSP cut-off value for MGMT promoter methylation and validate its prognostic significance in GBM patients. The impact of MGMT methylation status on survival outcomes was analyzed concerning surgical extent, tumor localization, and white matter tract involvement.

Methods

A retrospective analysis of 101 GBM patients (IDH-wildtype) diagnosed between 2008 and 2022 was performed. All patients underwent surgical resection (total/partial excision or stereotactic biopsy) followed by standard chemoradiotherapy. MGMT promoter methylation status was assessed using real-time qMSP. The optimal cut-off value was determined via receiver operating characteristic curve analysis. Kaplan-Meier survival analysis and Cox regression models evaluated the association between MGMT methylation levels, clinical characteristics, and overall survival (OS).

Results

Among 101 patients with IDH-wildtype glioblastoma, a qMSP cut-off value of 0.242% demonstrated strong diagnostic performance for MGMT methylation status (AUC = 0.875), with 78% sensitivity and 86% specificity. Patients with high methylation levels (≥ 0.242%) showed significantly longer median overall survival compared to those with low methylation (24 vs. 12 months; p = 0.006). This prognostic relevance persisted across surgical and anatomical subgroups. Multivariable Cox regression identified high qMSP methylation (HR ≈ 0.45, p < 0.001) and extent of resection ≥ 90% (HR ≈ 0.30, p = 0.002) as independent predictors of improved survival, whereas TERT promoter mutation (HR ≈ 1.9, p = 0.017) was associated with worse survival. Stratified analysis revealed that TERTp-mutant tumors with low methylation had the worst outcomes. Additionally, excisional surgery and neocortical tumor involvement were associated with significantly better survival (p = 0.0010 and p = 0.0218, respectively). These findings validate within our institutional setting the clinical utility of the 0.242% qMSP threshold for prognostic stratification in glioblastoma, although external multicenter validation is warranted before generalization to routine clinical practice.

Conclusion

The identified qMSP cut-off value (0.242) based on the procedure described in this study provides a robust prognostic stratification tool for GBM patients. High MGMT methylation correlates with improved survival, supporting its integration into clinical decision-making. Further multi-center validation studies are warranted to establish standardized MGMT assessment methodologies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-15225-2.

Introduction

Glioblastoma (GBM) is one of the most common primary tumors of central nervous system (CNS), 14.5% of all tumors [1]. Over the years, there has been an enormous development in the treatment modality of GBMs, although the treatment is mainly the extensive surgical resection followed by radiotherapy and chemotherapy. The increased use of molecular testing has become the standard for histological classification and serves as a prognostic marker. O6-Methylguanine-DNA Methyltransferase (MGMT) promoter methylation is considered as one of the most studied prognostic and predictive biomarker in GBMs. It has shown that tumor cells with low levels of MGMT protein have a higher susceptibility to the alkylating agents, resulting in substantial survival benefits [2]. Thus, MGMT promoter methylation has been demonstrated as a predictive marker for tumor response to Temozolomide (TMZ), one of the most common chemotherapeutic agent used for the treatment of GBM [3].

Different methods for determining MGMT promoter gene methylation are available in clinical practice. Those are pyrosequencing, methylation-specific PCR (MSP), Quantitative real-time PCR high-resolution melt(PCR-HRM), Methylation-specific multiplex ligation-dependent probe amplification (MS-MLPA), immunohistochemistry (IHC) and Infinium Methylation EPIC BeadChip Array (27k, 450k, or 850k) [4]. All assays rely on bisulfite conversion of unmethylated cytosines to uracils. Controversies about the best method for analyzing MGMT status and its prognostic potential are still debated. MSP is the first described technique that uses primers specific to methylated or unmethylated sequences to detect fully methylated alleles qualitatively. While it offers high specificity, it has lower sensitivity and is prone to equivocal results when signals are faint [5]. Quantitative MSP (qMSP) provides a methylation number even in faint signals by using quantitative PCR technique and recognizes fully methylated sequences [6, 7]. qMSP is supported by many studies in terms of clinical utility due to the technique is cheaper and widely available in addition to its shorter turnaround time.

Here, we aimed to determine an optimal qMSP cut-off value for MGMT status using our institutional cohort and to validate its prognostic relevance in relation to surgical and anatomical parameters. These results provide strong internal evidence supporting the clinical utility of this threshold, although confirmation in larger multicenter cohorts will be necessary before adoption as a universal clinical standard.

Method

Retrospective study of 101 patients diagnosed with Glioblastoma, IDH- wildtype from 2008 to 2022 at Acıbadem University School of Medicine. All patients received standard treatment for GBM, including surgical resection (total resection, partial resection or biopsy alone) and radiation with concurrent Temozolomide (TMZ), followed by adjuvant TMZ therapy. Based on the evaluation of preoperative and postoperative MRI scans, patients were classified according to tumor multifocality, multicompartmental involvement, and their associations with white matter tracts (long association fibers, commissural fibers, and projection fibers) at the time of diagnosis. The extent of resection (EOR) was quantitatively assessed by comparing contrast-enhancing tumor volumes on preoperative and early postoperative T1-weighted gadolinium-enhanced MRI.Compartments were subdivided into six groups; basal ganglia, neocortex, allocortex, hypothalamus, brain stem and cerebellum. The isocortex, corresponding to the majority of the neocortex and comprises approximately 90% of the cerebral hemispheres, predominantly occupying the superficial cortical surface. The allocortex includes structures such as the hippocampal formation and olfactory cortex, which are typically situated more deeply within the cerebral architecture.

In addition, IDH1/IDH2 (target regions including IDH1 R132, IDH2 R140, IDH2 R172), and pTERT (target regions including C228T and C250T) mutations were investigated for 101 patients’ genomic DNA. These regions amplified by the nested PCR method and then mutations were detected by the minisequencing (SNaPshot) method using the SNaPshot™ Multiplex Kit produced by Applied Biosystems.Real-time PCR analysis was performed at the Acibadem Molecular Pathology Laboratory to evaluate the methylation status of 10 CpG islands (positions at 76, 77, 78, 79, 80, 81, 84, 85, 86, and 87) within the MGMT exon 1 promoter region. The genomic DNA was extracted from 5 μm tissue sections of containing well-preserved tumor cells with minimum 30% tumor purity using the QIAamp DNA FFPE TISSUE KIT(Qiagen). The quality and concentration of the DNA were assessed using the QIAxpert (Qiagen), and samples containing a minimum DNA concentration of ≥ 10ng/µl were included in the study.

Bisulfite conversion was performed using the EZ DNA Methylation-Lightning Kit (Zymo Research, USA, cat no: D5030). The methylation status of MGMT promoter was performed by the quantitative methylation-specific real-time PCR method using geneMAP™ MGMT Methylation Analysis Kit (Genmark Saglık Ürünleri, Istanbul, Turkey), which has a sensitivity of at least 0.1%. qPCR reactions were carried out on the CFX Connect Real Time System (Bio-Rad). Methylation ratio was calculated for each sample using formula specified in the kit.

Statistical analysis

The demographic, clinical, and molecular characteristics of the cohort were summarized using descriptive statistics. Continuous variables were expressed as mean ± standard deviation [8] for normally distributed data or as median (Q1–Q3) for non-normally distributed data. Categorical variables were reported as frequencies and percentages. Comparisons between groups were conducted using the independent samples t-test for normally distributed continuous variables and the Mann-Whitney U test for non-normally distributed variables. Categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate. Kaplan-Meier survival analysis was performed to evaluate overall survival (OS), with survival curves compared using the log-rank test. Cox proportional hazards regression analysis was employed to assess the association between OS and potential prognostic factors, such as qMSP cut-off values, TERTp mutations, Ki-67 index, multifocality and extent of resection (EOR). Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated for each variable. An optimal qMSP cut-off value was determined using receiver operating characteristic [3] curve analysis to maximize the difference in OS between groups. Statistical significance was set at p < 0.05. All analyses were conducted using SPSS version 28.0 (IBM Corp., Armonk, NY, USA).

Determination of the optimal qMSP threshold for predicting MSP status

To determine the optimal qMSP (quantitative methylation-specific PCR) threshold corresponding to the binary classification obtained from conventional MSP (methylation-specific PCR), a receiver operating characteristic [3] analysis was performed. The analysis was restricted to cases with both valid MSP results and measurable qMSP values (n = 56). MSP status (“Methylated” = 1, “Unmethylated” = 0) was used as the binary target variable, while qMSP values were treated as the predictor. The ROC curve was generated by plotting sensitivity (true positive rate) against 1-specificity (false positive rate) across all possible qMSP thresholds. The optimal cut-off was defined as the point that maximized the Youden index (J = Sensitivity + Specificity – 1), representing the best balance between sensitivity and specificity. This approach ensured that the derived threshold reflected the added sensitivity of qMSP while remaining anchored to the established MSP-based framework. Patient-level clinical and molecular data used for statistical analyses are provided in Supplementary File 2. The distribution of qMSP values across the cohort is provided in Supplementary File 1.

Results

Patient and tumor characteristics

The median age of the participants was 59 years (Q1-Q3: 51.00–66.00), with a mean age of 58.30 ± 10.84 years. The Ki-67 proliferation index showed a median value of 30% (Q1-Q3: 20%−40%) and a mean of 30 ± 15%. The overall survival duration was 18 months (Q1-Q3: 10.00–31.00) with a mean of 23.01 ± 17.93 months. Regarding the number of compartments involved, the median was 2.00 (Q1-Q3: 1.00–3.00), and the mean was 2.19 ± 1.43. The qMSP percentage had a wide range, with a median of 0.29% (Q1-Q3: 0.02%−3.93%) and a mean of 3.21 ± 5.94%. The cohort consisted of 57 males (56.4%) and 44 females (43.6%). Most patients underwent excision surgery (74.3%), while 25.7% underwent stereotactic biopsy. MSP unmethylation was observed in 59 patients (60.8%), whereas MSP methylation was identified in 38 patients (39.2%). 78.2% had TERT promoter mutations were observed in 79 patients (78.2%), whereas TERT wild-type was identified in 22 patients (21.8%).

The majority of patients did not present with multifocality (81.2%), while 18.8% did. Similarly, 50.5% of patients had multi-compartmental involvement, while 49.5% did not. A high mortality rate was observed, with 91 patients (90.1%) classified as deceased, while only 8.9% (9 patients) alive (Table 1).

Table 1.

Demographic, Clinical, and pathological characteristics of the study cohort

Variable	Median [Q1-Q3]/Mean ± SD - n (%)
Age (years)	59.00 [51.00–66.00]/58.30 ± 10.84
Ki-67 (%)	30% [20%- 40%]/30% ± 15%
Overall Survival (months)	18.00 [10.00–31.00]/23.01 ± 17.93
Number of Compartments	2.00 [1.00–3.00]/2.19 ± 1.43
qMSP (%)	0.29 [0.02–3.93]/3.21 ± 5.94
Sex
Male	57 (56.4%)
Female	44 (43.6%)
Surgery Type
Excision	75 (74.3%)
Stereotactic biopsy	26 (25.7%)
MSP
Unmethylated	59 (60.8%)
Methylated	38 (39.2%)
Deceased	91 (90.1%)
Alive	9 (8.9%)
Multifocal
Present	19 (18.8%)
Absent	82 (81.2%)
Multi-Compartmental
Present	51 (50.5%)
Absent	50 (49.5%)

Survival analysis

The analysis revealed several factors significantly associated with overall survival (OS). Among numerical parameters, the number of compartments and Ki-67 index exhibited a negative correlation with OS, though these correlations were not statistically significant (p = 0.58 and p = 0.74, respectively). Gender showed a statistically significant difference, with female patients demonstrating a longer survival (27.5 ± 20.1 months) compared to male patients (19.4 ± 15.2 months, p = 0.0408). Surgical type had a significant impact on survival, where patients who underwent excision surgery had a mean OS of 26.0 ± 18.2 months, significantly higher than those who underwent stereotactic biopsy (14.4 ± 14.3 months, p = 0.0010). MSP methylation status also correlated with improved survival outcomes, with methylated cases showing an OS of 32.4 ± 19.4 months compared to 17.0 ± 14.4 months in unmethylated cases (p < 0.0001). Multifocality was associated with worse survival outcomes; patients with multifocal disease had a mean OS of 14.6 ± 9.0 months compared to 25.1 ± 19.0 months in non-multifocal cases (p = 0.0318).

Regarding regional involvement, isocortical involvement was the only statistically significant factor. Patients with isocortex involvement had a mean OS of 23.8 ± 18.0 months, which was significantly longer than the 8.6 ± 6.4 months observed in patients without isocortical involvement (p = 0.0218) (Fig. 1). Other regions, such as basal ganglia, allocortex, hypothalamus, thalamus, and brainstem, showed no significant impact on OS. Further subgroup analyses and threshold evaluation results are presented in Supplementary File 1 (Supplement 1 Fig. 2) (Table 2).

Fig. 1.

Illustrative cohort cases demonstrating tumor involvement in the isocortex (a) and allocortex (b)

Table 2.

The effects of Clinical, Surgical, and pathological parameters on overall survival

Parameters	Effect to Overall Survival/Mean ± SD (month)	p-Value
Number of Compartments	Negative correlation	0.58
Age	Negative correlation	0.53
Ki-67	Negative correlation	0.74
Sex		0.041
Male	19.4 ± 15.2
Female	27.5 ± 20.1
Surgery Type		0.001
Excision	26.0 ± 18.2
Stereotactic Biopsy	14.4 ± 14.3
MSP		< 0.001
Methylated	32.4 ± 19.4
Unmethylated	17.0 ± 14.4
Multifocal	14.6 ± 9.0	0.032
Unifocal	25.1 ± 19.0
Isocortex		0.022
Involved	23.8 ± 18.0
Not involved	8.6 ± 6.4

Determination of the optimal cut-off value

To identify the optimal qMSP threshold reflecting MSP-defined methylation status, a receiver operating characteristic [3] analysis was performed using MSP as the binary reference variable. The analysis was restricted to cases with measurable qMSP values (n = 56). The ROC curve demonstrated strong discriminatory power in this cohort, with an area under the curve (AUC) of 0.875. The optimal cut-off point was determined as 0.242% by maximizing the Youden index, yielding 78% sensitivity, 86% specificity, 91% positive predictive value (PPV), and 70% negative predictive value (NPV) (Fig. 2a).

Fig. 2.

Diagnostic and prognostic relevance of the threshold of ≥ 0.242% qMSP methylation cutoff in GBM- IDHwt patients. a ROC analysis identified 0.242% as the optimal qMSP threshold, yielding an AUC of 0.875 with strong sensitivity and specificity for methylation classification. b Kaplan-Meier survival analysis based on the 0.242% cutoff shows significantly improved overall survival in the group above the cutoff (qMSP ≥ 0.242%) compared to the group below the cutoff (qMSP < 0.242%) (log-rank p = 0.0006). c Survival comparison between unmethylated and low-positive (0 < qMSP < 0.242%) cases revealed no significant difference (p = 0.722), suggesting biological similarity. d Kaplan–Meier analysis of the entire cohort shows that patients with qMSP ≥ 0.242 had significantly longer overall survival compared to those with qMSP < 0.242 or unmethylated (log-rank p = 0.00026)

To evaluate the prognostic significance of MGMT promoter methylation as determined by qMSP, we performed survival analyses using a predefined cut-off value of 0.242. Overall survival (OS) was calculated in months, and log-rank tests were applied to compare groups.Among 56 patients with quantitative qMSP data, 29 were classified as having methylation levels ≥ 0.242, and 27 were classified as < 0.242. Kaplan–Meier analysis demonstrated a significantly longer OS in the high methylation group compared with the low methylation group with measurable qMSP values (log-rank p = 0.0006) (Fig. 2b). This finding indicates that the qMSP cut-off reliably discriminates between prognostically favorable and unfavorable subgroups.We next compared patients with low positive qMSP values (0–0.242; n = 27) against those explicitly annotated as “unmethylated” by qMSP (n = 37). In this analysis, survival curves were largely overlapping, with no significant difference in OS between groups (log-rank p = 0.722) (Fig. 2c). This suggests that the biological and clinical behavior of “low positive” cases approximates that of truly unmethylated tumors.Finally, in the entire cohort, patients with qMSP ≥ 0.242 (n = 29) were compared against those with either qMSP < 0.242 or designated as unmethylated (n = 64). Kaplan–Meier analysis revealed a striking survival advantage for the high methylation group, with a highly significant difference in OS (log-rank p = 0.00026) (Fig. 2d). This reinforces the clinical utility of the 0.242 cut-off, as it clearly distinguishes patients with favorable outcome, even when “unmethylated” cases are incorporated into the lower methylation category.

Correlation of overall survival with clinical and molecular parameters

Multivariable Cox regression analysis identified several independent predictors of overall survival in GBM-IDHwt patients. High qMSP methylation (≥ 0.242%) was strongly associated with improved survival (HR ≈ 0.45, p < 0.001), while greater extent of resection (EOR ≥ 90%) also independently predicted longer survival (HR ≈ 0.30, p = 0.002). In contrast, TERT promoter mutations were associated with increased mortality risk (HR ≈ 1.9, p = 0.017). Age was not significantly associated with survival (HR ≈ 1.01, p = 0.367). Patients with high qMSP methylation, TERTp wild-type tumors, or ≥ 90% resection showed significantly prolonged survival. These results confirm the independent prognostic value of qMSP, TERTp status, and surgical (EOR) factors in GBM outcomes (Fig. 3).

Fig. 3.

Prognostic significance of qMSP methylation, TERT promoter mutation, and extent of resection (EOR) in GBM-IDHwt patients

Of 101 glioblastoma patients, 92 were included in survival analysis after excluding 9 patients (5 with invalid qMSP data, 4 with missing survival data). Thirty-seven patients labeled “Unmethylated” were included by assigning qMSP = 0, placing them below the 0.242 methylation threshold.Patients were stratified into four groups: TERTp-mutant + MGMT methylated (n = 26), TERTp-mutant + MGMT unmethylated (n = 48), TERTp-wildtype + MGMT methylated (n = 6), and TERTp-wildtype + MGMT unmethylated (n = 12). Median overall survival varied significantly by group: TERTp-wildtype + MGMT methylated (42.0 months), TERTp-mutant + MGMT methylated (26.0 months), TERTp-wildtype + MGMT unmethylated (19.0 months), and TERTp-mutant + MGMT unmethylated (13.0 months). Twenty-four-month survival rates were 66.7%, 53.8%, 41.7%, and 19.1%, respectively. Kaplan-Meier analysis showed highly significant differences between groups (p = 0.0005). Pairwise analysis revealed significant survival differences between methylated and unmethylated patients within TERTp-mutant tumors (p = 0.0022), and significant differences in TERTp-wildtype tumors (p = 0.0314). TERTp status alone did not significantly impact survival within methylation groups (methylated groups: p = 0.1779; unmethylated groups: p = 0.3991). These findings establish MGMT methylation as the dominant prognostic factor in glioblastoma, with the survival benefit most pronounced when combined with TERTp-wildtype status (Fig. 4).

Fig. 4.

Kaplan-Meier survival curves stratified by TERTp mutation and MGMT methylation status in glioblastoma patients (n = 92).Patients were grouped by TERTp promoter mutation status and MGMT methylation determined by qMSP (threshold: ≥0.242 methylated, < 0.242 unmethylated). Patients labeled “Unmethylated” were assigned qMSP = 0 and included in the unmethylated group. Group sizes: TERTp-mut + MGMT ≥ 0.242 (n = 26), TERTp-mut + MGMT < 0.242 (n = 48), TERTp-wt + MGMT ≥ 0.242 (n = 6), TERTp-wt + MGMT < 0.242 (n = 12). Tick marks indicate censored observations. Multivariate log-rank test showed significant differences between groups (p = 0.0005). MGMT methylation status was the primary prognostic factor, with methylated patients showing superior survival regardless of TERTp status

Discussion

MGMT is a repair protein encoded by MGMT gene on chromosome 10 (10q26), which is overexpressed in gliomas. This protein counteracts the effect of TMZ by removing alkyl groups from the guanine, as a feature of intrinsic cellular defense mechanism [4]. MGMT promoter is consistent of 98 CpG islands. Among these 98 islands, methylation of DMR1 and DMR2 has been shown to be responsible for transcriptional silencing [9]. Epigenetic modifications at CpG islands, resulting in hypermethylation of the MGMT promoter gene, lead to insufficient DNA repair [2]. CpG positions relevant to survival described in the literature include CpG83, CpG84, CpG86, and CpG87, with CpG84 showing the strongest association with overall survival and having the highest impact on MGMT expression. In our clinical practice, we analyze 10 CpG islands in total and among these, there are three CpG islands (CpG84, CpG86, and CpG87) mentioned in the literature as having an impact on survival [10, 11].

A variety of techniques have been developed for MGMT methylation analysis, including methylation-specific PCR (MSP), pyrosequencing, and methylation arrays [12]. Among these, quantitative methylation-specific PCR (qMSP) is clinically feasible especially in settings where cost, speed, and formalin-fixed paraffin-embedded (FFPE) material compatibility are relevant concerns. qMSP enables sensitive and specific detection of methylated alleles by using bisulfite-converted DNA and primers/probes targeting CpG-rich regions of the MGMT promoter. qMSP has been implemented in several multicenter trials (e.g., NOA-08) as a standardized test for MGMT methylation evaluation, supporting its robustness in routine pathology workflows [8]. However, identifying an appropriate cut-off value in the qMSP approach has consistently been a challenge, and no universal standardization has been established regarding the choice of kits or threshold values. This study demonstrates that we applied the qMSP method accurately and that the cut-off we defined effectively stratified patients in a way that reflects its prognostic impact on overall survival.

Our findings reinforce the prognostic significance of MGMT promoter methylation in GBMs and contribute to the ongoing discussion regarding an optimal qMSP cutoff for clinical use. The improved survival associated with higher MGMT methylation highlights the importance of incorporating qMSP-based stratification into clinical decision-making, particularly in identifying patients who may benefit most from alkylating chemotherapy.

One of the strengths of this study is the extensive clinical follow-up data, which allowed for a more comprehensive assessment of the interplay between MGMT methylation, surgical approach, and anatomical tumor characteristics.In our cohort, we found no statistically significant difference in overall survival between “low positive” and unmethylated cases. This suggests that tumors with low-level methylation behave biologically and clinically in a manner similar to truly unmethylated tumors. Based on this evidence, we propose that a qMSP value of < 0.242 can be reasonably classified as “unmethylated.” This threshold not only provides a biologically meaningful distinction but also offers practical guidance for harmonizing molecular results with clinical decision-making.

Our findings further demonstrate that gross total resection significantly enhances survival outcomes, particularly in patients with high MGMT methylation levels. This supports previous work indicating that the extent of resection is a major prognostic factor in glioblastoma, irrespective of molecular subtype [13, 14]. However based on determined cutoff value (0.242), stereotactic biopsy patients with high qMSP values were associated with a longer median survival compared to low qMSP values, though the statistical significance was borderline, likely due to the small sample size. This finding further indicates that the identified cutoff value may hold clinical significance. In addition, TERT promoter mutations indicate poor prognosis in gliomas, whereas MGMT promoter methylation is a favorable prognostic marker. Both are key molecular indicators in clinical decision-making [15]. In our cohort, TERT promoter mutations were associated with shorter survival, and this association remained significant in multivariable Cox regression analysis using our predefined methylation cutoff, reinforcing the adverse impact of TERT mutations compared to the beneficial role of MGMT methylation. Consistently, another study also demonstrated prolonged survival in patients with high MGMT methylation and TERT promoter wild-type status, further supporting our findings [16].

Haque et al. [17] showed that multifocal GBM patients have worse progression-free survival (PFS) and overall survival (OS) compared to unifocal GBM patients. However, maximal safe resection significantly improves survival outcomes even in multifocal GBM cases. In our study based on the determined qMSP cut-off value of 0.242, survival outcomes were analyzed across various anatomical regions and surgical groups. We demonstrated that multifocality at the time of diagnosis as a poor prognostic factor and surgical excision has superiority to stereotactic biopsy regarding overall survival. Each distinct component of the brain, including phylogenetic compartments (such as the isocortex, allocortex, and diencephalon), different sectors of the white matter (association, projection, and commissural tracts), and specific germinal zones (such as the supraventricular zone) hosts different progenitor populations that may serve as the tumor stem cells of GBMs. Our previous study demonstrated that tumors invading different phylogenetic compartments were associated with worse survival [18]. Building on this, we sought to identify specific anatomical patterns that might correlate with survival. The present analysis indicated that GBMs located in the isocortex exhibit more favorable survival characteristics compared with those arising in other compartments. The absence of statistically significant effects on overall survival in regions such as the basal ganglia, allocortex, hypothalamus, thalamus, brainstem, long association fibers, commissural bodies, projection tracts, and subventricular zones is consistent with previously published observations [19] (Supplement 1. Table 1).

These findings align with prior reports that highlight the role of MGMT methylation in glioblastoma prognosis. Hegi et al. demonstrated that patients with MGMT promoter methylation exhibited improved survival following temozolomide treatment compared to unmethylated cases [2]. Subsequent studies have further confirmed that methylation status is a robust predictor of chemotherapy response [20]. Additionally, research suggests that the extent of methylation across CpG islands within the MGMT promoter region may further refine its prognostic utility, indicating that not all methylation events contribute equally to therapeutic outcomes [11, 21].

Despite its strengths, this study has limitations. Its retrospective design introduces potential selection bias, and although the cohort size is robust, external validation in larger multicenter studies is needed. Moreover, the limited sample sizes in certain subgroup analyses warrant cautious interpretation of these results, and the implications for clinical decision-making should therefore be interpreted as hypothesis-generating rather than definitive. Inter-laboratory variability in qMSP assays may also affect reproducibility across institutions. Future studies should aim to integrate multi-institutional data to enhance the generalizability of these findings and explore additional factors influencing MGMT methylation’s prognostic role.

Conclusion

In conclusion, this study identified an optimal cutoff value for the qMSP method based on patient survival outcomes. The validity of the identified cutoff was verified within the same patient group by considering factors such as age, gender, tumoral involvement, and extent of resection. Our results support the internal validity of qMSP-based stratification using the 0.242 cutoff in glioblastoma patients. However, external and multicenter studies are needed before recommending this threshold as a routine clinical standard.

Authors’ contributions

Z. H and A. E. D were responsible for the study design and wrote the main manuscript text. Z.H, E.U and F. K. A conducted the data analysis. M. A.I, T. D, G. G, S. E, K. O and N. P reviewed the manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

All data generated or analysed during this study are included in this published article and its supplementary information files.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki, and the study was approved by the Medical Research Ethics Committee of Acıbadem University Institutions (No: ATADEK-2022-04/109). Informed consent was obtained from all participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.