How molecular profiling turns the standard Stupp protocol into a decision the tumor's biology actually informs.
Before a single chemotherapy decision is made, the tumor must be molecularly classified. The WHO 2021 CNS classification now mandates an integrated molecular diagnosis - meaning the tumor's DNA profile is as important as how it looks under the microscope. This classification determines which chemotherapy backbone even enters consideration.
Think of molecular classification as sorting the mail before delivering it. If the mail goes to the wrong address, it doesn't matter how fast the courier runs. Classification ensures every treatment decision that follows is aimed at the right disease.
Flowchart showing molecular classification of adult diffuse gliomas: IDH status branches into IDH-mutant and IDH-wildtype, with IDH-mutant further dividing by 1p/19q co-deletion into oligodendroglioma and astrocytoma, and a separate branch for H3-altered diffuse midline glioma.
MGMT promoter methylation is the most important chemotherapy biomarker in glioblastoma because it predicts whether temozolomide-induced DNA damage is likely to persist or be repaired. But here is the critical detail most patients are never told: MGMT methylation is an epigenetic event, not a genetic mutation - which means standard genomic sequencing panels will not detect it. It requires dedicated testing.
Think of MGMT as a tiny repair worker that lives inside your tumor cells. Its only job is to fix DNA damage - specifically, the kind of damage that chemotherapy drugs like temozolomide are designed to cause.
Methylation is like putting that repair worker to sleep. When the MGMT gene is "methylated," the repair worker is switched off. That means when chemotherapy damages the tumor's DNA, the damage sticks - and the cancer cell dies.
When MGMT is not methylated (unmethylated), the repair worker is wide awake. It quickly fixes the damage that chemotherapy causes, making the treatment less effective - like mopping up water while someone keeps pouring.
MGMT Methylated
When the MGMT promoter is methylated, the tumor is less efficient at repairing alkylation damage, making temozolomide more effective.
Expected signal
Median overall survival is often quoted around 21.7 months with methylated disease.
MGMT Unmethylated
When MGMT remains active, the tumor can reverse TMZ-induced DNA injury more efficiently, which is why chemotherapy strategy often needs to broaden.
Why it changes the plan
Outcomes are often closer to 12.7 months in unmethylated disease, making alternative schedules, agents, or trial options more important to discuss.
Methylation-specific PCR (MSP)
The most widely used clinical method. Uses bisulfite-converted DNA with primers specific to methylated vs. unmethylated sequences. Gives a binary result (methylated or unmethylated).
Pyrosequencing
Quantitative method that gives a percentage methylation value across specific CpG sites. More granular than MSP, allows clinicians to see how strongly the gene is silenced.
Genome-wide methylation arrays (Heidelberg Classifier)
The gold standard in research. Illumina EPIC/850K arrays profile the entire methylome, giving MGMT status plus a full tumor classification plus copy number profile in a single assay.
Immunohistochemistry (IHC) for MGMT protein
Detects protein expression rather than methylation directly. Less reliable because protein loss can occur through mechanisms other than promoter methylation. Generally considered insufficient as a standalone test.
If direct MGMT methylation testing is unavailable, certain genomic alterations from standard sequencing can serve as indirect predictors. Here is what your existing molecular report may already tell you:
| Genomic Finding | What It Suggests About MGMT | Confidence Level |
|---|---|---|
| IDH1/2 mutation present | Very likely methylated - IDH mutations produce the oncometabolite 2-HG, which causes the Glioma CpG Island Methylator Phenotype (G-CIMP), driving global hypermethylation including at the MGMT locus. | High (>80–90%) |
| IDH-mutant + TP53 mutation + ATRX loss | Very likely methylated - this combination defines the astrocytoma, IDH-mutant lineage, which is driven by G-CIMP. | High |
| 1p/19q co-deletion | Almost certainly methylated - defines oligodendroglioma, which is by definition IDH-mutant. | Very high (>95%) |
| IDH-wildtype + TERT promoter mutation + chromosome 10q loss | Cannot determine - approximately 40–50% chance either way. | Low - direct testing is mandatory |
| No informative molecular markers | Cannot determine. | Direct testing is mandatory |
Genomic Finding
IDH1/2 mutation present
What It Suggests About MGMT
Very likely methylated - IDH mutations produce the oncometabolite 2-HG, which causes the Glioma CpG Island Methylator Phenotype (G-CIMP), driving global hypermethylation including at the MGMT locus.
Confidence Level
High (>80–90%)
Genomic Finding
IDH-mutant + TP53 mutation + ATRX loss
What It Suggests About MGMT
Very likely methylated - this combination defines the astrocytoma, IDH-mutant lineage, which is driven by G-CIMP.
Confidence Level
High
Genomic Finding
1p/19q co-deletion
What It Suggests About MGMT
Almost certainly methylated - defines oligodendroglioma, which is by definition IDH-mutant.
Confidence Level
Very high (>95%)
Genomic Finding
IDH-wildtype + TERT promoter mutation + chromosome 10q loss
What It Suggests About MGMT
Cannot determine - approximately 40–50% chance either way.
Confidence Level
Low - direct testing is mandatory
Genomic Finding
No informative molecular markers
What It Suggests About MGMT
Cannot determine.
Confidence Level
Direct testing is mandatory
The critical clinical reality: the one group where MGMT status most changes chemotherapy decisions - IDH-wildtype glioblastoma - is precisely where these genomic surrogates cannot give you a reliable answer. If your molecular report shows IDH-wildtype disease and does not include MGMT methylation status, push for dedicated testing before committing to a temozolomide-heavy protocol.
Since 2005, the Stupp Protocol has defined first-line chemotherapy for most newly diagnosed glioblastoma patients: concurrent temozolomide with focal radiation, followed by adjuvant temozolomide. The reason this protocol became foundational is simple - adding temozolomide to radiation improved median survival from roughly 12 months to 14.6 months. That gain matters, but it also highlights the limits of a uniform chemotherapy strategy.
In the adjuvant phase, temozolomide is commonly delivered for 6 cycles on a 5-days-on, 28-day schedule, although real-world duration can change based on tolerance, imaging, and physician preference. The Stupp Protocol is the starting point for most GBM patients - but as you'll see in the sections that follow, molecular profiling can tell you whether this standard approach is well-matched to your tumor's biology, or whether the strategy needs to be adapted from the outset.
Concurrent phase
TMZ + RT
Adjuvant phase
Usually 6 cycles
Typical cadence
5 days every 28
Temozolomide is the backbone, but it is not the whole pharmacy. At recurrence, in MGMT-unmethylated disease, or in specific molecular subtypes, other cytotoxic agents enter the discussion.
Metronomic chemotherapy uses lower, more continuous dosing rather than widely spaced high-dose pulses. Dose-dense temozolomide takes a related approach by increasing exposure frequency, such as 21 days on / 28 days or 7 days on / 7 days off. The rationale extends beyond hitting proliferating tumor cells: it also keeps pressure on MGMT-mediated repair, angiogenesis, and immune signaling inside the tumor microenvironment.
This is where precision oncology and integrative thinking intersect. Chemotherapy in GBM is about more than cell kill; it is about reshaping the conditions that let glioblastoma persist. These schedules are not universal upgrades. They are strategic options that make the most sense when biology, prior tolerance, marrow reserve, and treatment goals all point in the same direction.
Markers of proliferative tempo, including Ki-67 and related pathology features, can influence whether a standard, metronomic, or dose-dense schedule is biologically reasonable.
At glioblastoma.center, powered by Art of Healing Cancer, we believe that wherever possible, the dosage of chemotherapy should be lowered rather than maximised. Higher doses do not always mean better outcomes - they often mean more toxicity, deeper immunosuppression, and reduced quality of life without a proportional survival benefit. Our approach prioritises the lowest effective dose that maintains meaningful anti-tumor pressure while preserving the patient's bone marrow reserve, immune function, and overall well-being. This is not about under-treating - it is about treating smarter, guided by molecular profiling, functional testing, and the individual biology of each patient's tumor.
Once molecular subtype and MGMT status are established, deeper genomic profiling reveals specific alterations that refine which drugs are selected, added, or avoided. Think of this as the precision layer - the mutations, amplifications, deletions, and fusions in your tumor that either create a vulnerability to a specific drug or predict resistance to one.
| Genomic Alteration | Drug(s) Indicated | Why It Works |
|---|---|---|
| MGMT promoter methylated | Temozolomide, Lomustine (CCNU), Carmustine (BCNU), Procarbazine | Alkylation damage goes unrepaired because the MGMT repair enzyme is silenced. |
| 1p/19q co-deletion | PCV regimen (Procarbazine + CCNU + Vincristine) | Exceptional and durable chemosensitivity - likely related to IDH-driven metabolic vulnerability plus impaired DNA damage response. |
| IDH1/2 mutation | Alkylating agents generally; Vorasidenib (IDH1/2 dual inhibitor, now FDA-approved for grade 2 IDH-mutant gliomas) | Altered metabolic state creates collateral vulnerabilities; Vorasidenib directly inhibits the mutant enzyme. |
| BRAF V600E | Dabrafenib + Trametinib | Direct kinase inhibition of mutant BRAF plus downstream MEK blockade. |
| NTRK1/2/3 fusions | Larotrectinib, Entrectinib | Direct inhibition of the oncogenic TRK fusion protein. |
| ALK fusions or amplification | Lorlatinib (best CNS penetration), Alectinib, Crizotinib | Kinase inhibition with varying degrees of BBB penetration. |
| FGFR-TACC fusions | Erdafitinib, Futibatinib, Infigratinib | FGFR kinase inhibition targeting the oncogenic fusion. |
| MET amplification / METex14 skipping | Capmatinib, Tepotinib | MET kinase inhibition. |
| High TMB / Mismatch Repair Deficiency | Checkpoint inhibitors (Pembrolizumab, Nivolumab) | High neoantigen load enables immune recognition of tumor cells. |
| CDKN2A homozygous deletion | CDK4/6 inhibitors (Ribociclib has best CNS penetration) | Loss of the p16 tumor suppressor leads to unchecked CDK4/6 activity, creating a druggable dependency. |
| PTEN loss / PI3K pathway mutation | mTOR inhibitors (Everolimus, Temsirolimus); PI3K inhibitors | Constitutive PI3K/AKT/mTOR pathway activation becomes a targetable vulnerability. |
| PDGFRA amplification | Dasatinib, Avapritinib | PDGFR kinase inhibition. |
| H3K27M alteration | ONC201 / Dordaviprone | Targets DRD2 and is showing signal specifically in H3K27M-mutant diffuse midline gliomas. |
| PTCH1 mutation (Hedgehog pathway) | Vismodegib, Sonidegib | Hedgehog pathway inhibition. |
Genomic Alteration
MGMT promoter methylated
Drug(s) Indicated
Temozolomide, Lomustine (CCNU), Carmustine (BCNU), Procarbazine
Why It Works
Alkylation damage goes unrepaired because the MGMT repair enzyme is silenced.
Genomic Alteration
1p/19q co-deletion
Drug(s) Indicated
PCV regimen (Procarbazine + CCNU + Vincristine)
Why It Works
Exceptional and durable chemosensitivity - likely related to IDH-driven metabolic vulnerability plus impaired DNA damage response.
Genomic Alteration
IDH1/2 mutation
Drug(s) Indicated
Alkylating agents generally; Vorasidenib (IDH1/2 dual inhibitor, now FDA-approved for grade 2 IDH-mutant gliomas)
Why It Works
Altered metabolic state creates collateral vulnerabilities; Vorasidenib directly inhibits the mutant enzyme.
Genomic Alteration
BRAF V600E
Drug(s) Indicated
Dabrafenib + Trametinib
Why It Works
Direct kinase inhibition of mutant BRAF plus downstream MEK blockade.
Genomic Alteration
NTRK1/2/3 fusions
Drug(s) Indicated
Larotrectinib, Entrectinib
Why It Works
Direct inhibition of the oncogenic TRK fusion protein.
Genomic Alteration
ALK fusions or amplification
Drug(s) Indicated
Lorlatinib (best CNS penetration), Alectinib, Crizotinib
Why It Works
Kinase inhibition with varying degrees of BBB penetration.
Genomic Alteration
FGFR-TACC fusions
Drug(s) Indicated
Erdafitinib, Futibatinib, Infigratinib
Why It Works
FGFR kinase inhibition targeting the oncogenic fusion.
Genomic Alteration
MET amplification / METex14 skipping
Drug(s) Indicated
Capmatinib, Tepotinib
Why It Works
MET kinase inhibition.
Genomic Alteration
High TMB / Mismatch Repair Deficiency
Drug(s) Indicated
Checkpoint inhibitors (Pembrolizumab, Nivolumab)
Why It Works
High neoantigen load enables immune recognition of tumor cells.
Genomic Alteration
CDKN2A homozygous deletion
Drug(s) Indicated
CDK4/6 inhibitors (Ribociclib has best CNS penetration)
Why It Works
Loss of the p16 tumor suppressor leads to unchecked CDK4/6 activity, creating a druggable dependency.
Genomic Alteration
PTEN loss / PI3K pathway mutation
Drug(s) Indicated
mTOR inhibitors (Everolimus, Temsirolimus); PI3K inhibitors
Why It Works
Constitutive PI3K/AKT/mTOR pathway activation becomes a targetable vulnerability.
Genomic Alteration
PDGFRA amplification
Drug(s) Indicated
Dasatinib, Avapritinib
Why It Works
PDGFR kinase inhibition.
Genomic Alteration
H3K27M alteration
Drug(s) Indicated
ONC201 / Dordaviprone
Why It Works
Targets DRD2 and is showing signal specifically in H3K27M-mutant diffuse midline gliomas.
Genomic Alteration
PTCH1 mutation (Hedgehog pathway)
Drug(s) Indicated
Vismodegib, Sonidegib
Why It Works
Hedgehog pathway inhibition.
A single genomic alteration is a hypothesis, not a guarantee. The presence of a target does not always mean the drug will work - and the absence of a known target does not mean no options exist. The layers that follow explain how to validate these genomic signals before they reach the patient.
Genomic mutations are necessary but not sufficient. A gene may be mutated but transcriptionally silent, or a pathway may be active through non-mutational mechanisms. Transcriptomic data - the measurement of which genes are actually being expressed as RNA - resolves this by answering one critical question: Is the target that genomics identified actually switched on in this tumor?
Target validation
EGFR amplification on DNA needs to be confirmed by high EGFR mRNA expression. If the gene is amplified but not being actively transcribed, targeting EGFR is unlikely to work despite the genomic finding.
Pathway activity confirmation
PTEN loss on DNA should correlate with actual mTOR pathway activation at the expression level. Transcriptomics confirms this through readouts of downstream effectors (phospho-S6K, 4EBP1 expression signatures).
Immune microenvironment profiling
Expression of PD-L1, T-cell infiltration signatures, and Treg/MDSC suppressive signatures determines whether immunotherapy has any real probability of working beyond what tumor mutational burden alone suggests.
Drug efflux pump expression
ABCB1 (P-glycoprotein) and ABCG2 (BCRP) expression levels predict whether drugs are being actively pumped out of tumor cells regardless of whether the target is present. High efflux pump expression can render an otherwise well-matched drug ineffective.
Molecular subtype refinement
GBM expression subtypes (Classical, Mesenchymal, Proneural) have different treatment response patterns. Mesenchymal subtype, for example, has higher NF-κB activity and may benefit from NF-κB pathway targeting adjuncts.
Think of genomics as reading the blueprint of a building, and transcriptomics as walking through the building to see which rooms are actually occupied. The blueprint might show a room exists, but only a walkthrough tells you if anyone is home.
This is the final validation layer before a drug reaches the patient. Functional drug sensitivity testing takes living tumor cells - either from circulating tumor cells in the blood or from fresh surgical tissue - and exposes them directly to candidate drugs in the laboratory. The result is a direct measurement: does this specific patient's tumor actually die when exposed to this specific drug?
A drug passes the selection filter only when there is convergence across all layers:
Any single layer alone can mislead. Convergence across all three provides the highest-confidence basis for drug selection.
Genomics tells you which drugs should work in theory. Functional testing tells you which drugs actually work against your tumor. The combination of both is what separates precision oncology from educated guessing.
Everything above converges here. This section maps how chemotherapy decisions differ by molecular subtype - because an IDH-wildtype glioblastoma, an IDH-mutant astrocytoma, an oligodendroglioma, and a diffuse midline glioma are biologically distinct diseases that call for different treatment strategies.
Molecular classification
IDH status → 1p/19q → H3 status → WHO integrated diagnosis
MGMT methylation status
Direct testing (MSP / pyrosequencing) - determines alkylating agent backbone: yes or no?
Actionable genomic alterations
Targeted panel or WES/WGS → identifies fusions, amplifications, mutations, deletions that match specific drugs
Transcriptomic confirmation
RNA expression confirms target activity and pathway state
Functional sensitivity testing
RGCC / ChemoID / organoid screening → validates candidate drugs against the patient's actual tumor cells
Protocol design
Select agents with convergent evidence across all layers. Sequence them based on synergy, toxicity, BBB penetration, and scheduling logic.
This five-layer convergence model is what separates empiric chemotherapy prescribing - where every GBM patient receives the same protocol regardless of molecular reality - from precision-guided drug selection, where the treatment is matched to the tumor's actual biology.