EORTC 22033-26033/NCIC-CTG/TROG/MRC-CTU
ForClinical high-risk WHO grade 2 glioma, first-line, molecularly classified
No significant difference
PFS and OS both ns between RT and TMZ arms; effect size not reported in source
TL;DRMature phase III: RT (50.4Gy/28fx) vs dose-dense TMZ in high-risk WHO grade 2 LGG, no PFS or OS difference in any molecular subtype.
The RT-relevant read is what this does NOT license: single-modality TMZ never beat RT, so deferring 50.4Gy/28fx to spare late toxicity buys nothing on PFS or OS. IDH wild-type was the one split (OS 2.5 vs 4.7 yrs, HR 0.47 [0.27-0.82], P=.0068), post hoc and n=64, and those tumors now grade as GBM.
In molecularly classified high-risk WHO grade 2 glioma being planned for first-line therapy, this supports neither leading modality over the other and does not address the combined-modality RT plus alkylator approach that is standard for IDH-mutant astrocytoma.
50.4Gy in 28fx as sole therapy was not beaten by TMZ, so deferring RT to postpone late neurocognitive cost does not buy an outcome advantage. The trial says nothing about RT dose, target volume or the combined-modality regimen a high-risk grade 2 astrocytoma now receives, so it does not move the modern RT decision, only closes the omission argument.
Dose-dense TMZ 75 mg/m² 21/28d for up to 12 cycles as sole first-line therapy is equivalent, not superior, to RT, and the schedule was never tested against the combined-modality standard. The IDH wild-type OS advantage (2.5 vs 4.7 yrs, HR 0.47 [0.27-0.82]) is post hoc in n=64 tumors that now classify as GBM.
8 details 2 trials watching
Randomised phase III, four-group intergroup trial (EORTC/NCIC-CTG/TROG/MRC-CTU), N=478, comparing two single-modality first-line strategies. This is the mature analysis; a post hoc reclassification to 2021 WHO criteria was possible in 73% (351/478) with analyzable tissue.
Clinical high-risk WHO grade 2 low-grade glioma, treatment-naive. Post hoc molecular strata: IDHmt/1p19q non-codeleted astrocytoma n=178, IDHmt/1p19q codeleted oligodendroglioma n=109, IDH wild-type n=64.
Standard fractionation, 28 × 1.8 Gy (50.4 Gy), delivered as the entire assigned treatment. No concurrent or adjuvant systemic therapy in the RT arm, which is the gap between this trial and how grade 2 glioma is treated now.
Dose-dense temozolomide 75 mg/m² once daily, 21 of 28 days, up to 12 cycles, given as sole first-line therapy with RT withheld.
Primary: progression-free survival. Overall survival and differential response by molecular marker were the mature-analysis questions.
No significant difference in PFS or OS between arms, and the null held across every molecular subtype except IDH wild-type, where OS favored TMZ.
| Subtype (n) | RT | TMZ | HR |
|---|---|---|---|
| IDHmt astrocytoma, non-codel (n=178) | 6.6-6.7 yrs (either arm) | 6.6-6.7 yrs (either arm) | 0.67-1.44, P=.93 |
| IDHmt oligodendroglioma, 1p/19q codel (n=109) | 12.9 yrs (9.4-NR) | 14.9 yrs (10.1-NR) | 0.88 (0.52-1.49), P=.63 |
| IDH wild-type (n=64) | 2.5 yrs (1.8-3.3) | 4.7 yrs (2.2-7.2) | 0.47 (0.27-0.82), P=.0068 |
RTOG 9802 established RT followed by PCV as the high-risk grade 2 benchmark, and combined-modality therapy is now standard for IDH-mutant astrocytoma. EORTC 22033 tested neither arm of that question: it asked which single modality to lead with, and answered that it does not matter.
Molecular strata are post hoc in 73% of the enrolled population, so subtype comparisons carry both selection and multiplicity risk. The IDH wild-type finding rests on n=64, and under 2021 criteria those tumors would not be enrolled as grade 2 at all.
The durable contribution is prognostic rather than therapeutic: median OS of 12.9-14.9 years in codeleted oligodendroglioma against 6.6-6.7 years in IDH-mutant astrocytoma quantifies how far molecular class outweighs modality choice. The finding that pts ≥40 fared better than <40 undercuts the age-40 cutoff that still gates high-risk definitions.
Mature randomised phase III, primary endpoint negative in both arms and across molecular subtypes; the IDHwt survival difference is post hoc, n=64.
- Optimal sequencing of RT and alkylator in IDH-mutant astrocytoma
- Whether IDH inhibitors displace upfront RT in low-risk grade 2 glioma n=90 · primary completion 2027-10 · PROs: vorasidenib vs RT vs surveillance in grade 2 IDH-mutn=150 · primary completion 2031-11 · real-world vorasidenib, grade 2, no immediate RT need
- Late neurocognitive cost of combined-modality vs single-modality
📚 Sources · 📄 1 paper
Abstract
The longer read
The negative primary endpoint is the least interesting thing about this readout, and it was already known from the 2016 interim publication. What mature follow-up adds is a decade-scale view of how little the initial modality choice mattered once molecular class is fixed, and that is the finding worth carrying into a tumor board.
Read against RTOG 9802, this trial answers an adjacent question rather than a competing one. RTOG 9802 showed that adding PCV to radiotherapy extended survival substantially in high-risk grade 2 glioma, and it is the reason combined-modality therapy became the benchmark. EORTC 22033 asked instead whether one modality could be swapped for the other up front, and the answer is that neither leading with RT nor leading with dose-dense TMZ changes progression-free or overall survival. Those two results are consistent: the benefit in this disease appears to come from receiving both, not from the sequence you start with. That leaves the reader with an important negative, which is that withholding radiotherapy to defer late neurocognitive toxicity does not come for free in outcome terms only because TMZ is doing equivalent work, and it certainly does not license single-modality TMZ as a strategy now that combined therapy is standard for IDH-mutant astrocytoma.
The post hoc molecular reclassification is where confidence should be handled carefully in both directions. On the one hand, tissue was analyzable in only 351 of 478 pts, so every subtype comparison is drawn from a subset selected by tissue availability rather than by randomisation, and three strata tested against a null primary invites multiplicity. On the other hand, the direction and magnitude of the subtype separations track what the field has independently established: codeleted oligodendroglioma at 12.9 and 14.9 years versus IDH-mutant astrocytoma at 6.6 to 6.7 years is a prognostic gradient nobody would now dispute, so its appearance here is corroborative rather than surprising.
The IDH wild-type result deserves the most skepticism, and not because HR 0.47 (0.27-0.82, P=.0068) is fragile on its own terms. The problem is what that group is. Under 2021 criteria, an IDH wild-type hemispheric glioma in an adult with the relevant molecular features is classified as glioblastoma, so these 64 pts, whose median survival of 2.5 and 4.7 years reflects that biology, would not be enrolled on a low-grade glioma protocol today. A reader cannot carry a TMZ-favoring signal from that group into any decision, because the decision for those pts is already made by a different standard entirely.
The age finding is the one that should change how a reader reads eligibility criteria rather than how they treat. Age under 40 has functioned as a favorable factor and as a component of high-risk definitions for decades, and here pts 40 and older fared better once tumors were molecularly classified. The plausible explanation is that younger pts in this cohort were enriched for the aggressive molecular subtypes, meaning age was standing in for biology it can now be replaced by. For a trial designer that argues against carrying an age cutoff into future high-risk stratification. For the clinic it argues only that molecular classification should precede any risk statement, which is already how these tumors should be worked up.