MIRAGE
Year of Publication: 2024
Journal: Neuro-Oncology
Link: https://doi.org/10.1093/neuonc/noae053
PDF: https://academic.oup.com/neuro-oncology/.../9/1670/7631790
PDF: Clinical Question
Does adding marizomib to standard TMZ/RT→TMZ improve overall survival in newly diagnosed glioblastoma, particularly in MGMT-unmethylated tumors?
Study Overview
Objective
Marizomib added to standard temozolomide-based radiochemotherapy (TMZ/RT→TMZ) vs standard alone — to test whether the blood-brain barrier-penetrant pan-proteasome inhibitor marizomib improves overall survival in newly diagnosed glioblastoma.
Study Summary
- Marizomib plus TMZ/RT→TMZ did not improve overall survival in newly diagnosed glioblastoma (16.5 vs 17.0 mo; HR 1.04; p=0.64).
- No benefit even in MGMT-unmethylated tumors (15.1 vs 14.5 mo; HR 1.13; p=0.27), the prespecified high-need subgroup.
- Progression-free survival was also equivalent (6.3 vs 6.0 mo; HR 0.97; p=0.67).
- Grade 3/4 nervous system AEs 33% vs 20% and psychiatric AEs 14% vs 3% with marizomib.
- 8 treatment-related deaths in the marizomib arm (soft tissue necrosis, cerebral hemorrhage, leukoencephalopathy).
- First phase 3 trial of a proteasome inhibitor in GBM — pan-proteasome inhibition in combination with TMZ/RT is not the path forward.
Intervention
Marizomib 0.8 mg/m² IV on days 1, 8, 15, 29, 36 during radiotherapy and on days 1, 8, 15 of each 28-day cycle during maintenance TMZ, with up to 12 additional cycles after TMZ completion (maximum 18 cycles), added to standard TMZ 75 mg/m² daily with RT 60 Gy + maintenance TMZ 150-200 mg/m² days 1-5 of 28-day cycle.
Patients per Arm
Standard 374; Marizomib 375
Bottom Line
In 749 adults with newly diagnosed glioblastoma randomized 1:1 to marizomib plus standard TMZ/RT→TMZ vs standard alone (EORTC 1709/CCTG CE.8), median overall survival was 16.5 vs 17.0 months (HR 1.04; p=0.64) with no benefit in the MGMT-unmethylated subgroup (15.1 vs 14.5 months; HR 1.13; p=0.27). Marizomib added significant neurological and psychiatric toxicity (grade 3/4 nervous system AE 33% vs 20%; psychiatric 14% vs 3%) without survival benefit.
Major Points
- Multicenter (82 institutions, Europe/Canada/US) open-label phase 3 superiority RCT with futility rule (Roth 2024)
- N=749 patients (99.9% of planned 750); 1:1 randomization to standard or standard + marizomib
- Stratified by institution, age (≤55 vs >55), KPS (70/80 vs 90/100), extent of surgery (partial/biopsy vs gross total)
- Primary endpoint: OS in ITT and in MGMT-unmethylated subgroup (coprimary)
- OS ITT: 16.5 vs 17.0 months; stratified HR 1.04 (95% CI 0.87-1.22); p=0.64 — futile
- OS MGMT-unmethylated: 15.1 vs 14.5 months; HR 1.13 (95% CI 0.88-1.44); p=0.27 — also no benefit
- PFS ITT: 6.3 vs 6.0 months; HR 0.97 (95% CI 0.82-1.13); p=0.67
- No subgroup (age, sex, KPS, MMSE, steroids, surgery extent, MGMT status) showed treatment-by-subgroup interaction (all p>0.15)
- Toxicity: Grade 3/4 AEs 67% vs 48%; nervous system AEs 33% vs 20%; psychiatric AEs 14% vs 3% with marizomib
- 8 treatment-related deaths in marizomib arm (soft tissue necrosis, intestinal perforation, seizure, cerebral hemorrhage, leukoencephalopathy, meningitis, head injury) vs 1 in standard
- 32% of marizomib patients discontinued the drug for AEs; 36% required dose withdrawal
- Conclusion: pan-proteasome inhibition with marizomib adds toxicity without benefit — proteasome pathway as monotherapy not viable in GBM
Design
Study Type: Multicenter open-label phase 3 superiority RCT with futility stopping rule (EORTC 1709 / CCTG CE.8, NCT03345095)
Randomization: 1
Blinding: Open-label (no blinding)
Follow-up Duration: Median 29.1 mo (marizomib) / 27.5 mo (standard)
Sample Size: 749
Analyzed: 749
Analysis: Stratified Cox proportional hazards model; stratified log-rank; coprimary ITT + MGMT-unmethylated
Inclusion Criteria
- Newly diagnosed histologically confirmed glioblastoma (WHO grade 4)
- Age ≥18 years
- Karnofsky performance status ≥70
- No prior GBM treatment other than surgery
- Stable or decreasing corticosteroid dose ≥1 week before consent
- Post-operative brain MRI within 14 days of randomization
- IDH-mutant tumors excluded (testing recommended for age <55 or atypical features)
Exclusion Criteria
- IDH1 R132H-mutant tumors (known)
- Prior non-surgical GBM treatment
- Unstable or increasing corticosteroid requirement
- KPS <70
- Significant cardiac, hepatic, or renal dysfunction precluding TMZ or marizomib
- Pregnancy
Baseline Characteristics
| Characteristic | Control | Active |
|---|---|---|
| N | 374 | 375 |
| Median age | 58.5 (21-79) | 58.0 (20-79) |
| Male | 255 (68.2%) | 233 (62.1%) |
| Gross total resection | 190 (50.8%) | 191 (50.9%) |
| KPS 90/100 | 249 (66.6%) | 252 (67.2%) |
| MGMT unmethylated | 233 (59.6%) | 217 (57.9%) |
| MGMT methylated | 116 (31%) | 122 (32.5%) |
| Baseline steroids | 159 (42.5%) | 150 (40%) |
Arms
| Field | Control | TMZ/RT→TMZ + Marizomib |
|---|---|---|
| N | 374 | 375 |
| Intervention | RT 60 Gy in 30 fx + concurrent TMZ 75 mg/m²/day × 6 weeks, then 4-week break, then maintenance TMZ 150-200 mg/m² days 1-5 of 28-day cycles × 6 cycles | Standard treatment + marizomib 0.8 mg/m² IV 10-min infusion on days 1/8/15/29/36 during RT, then days 1/8/15 of each 28-day cycle during TMZ maintenance, continuing as single agent for up to 12 additional cycles (maximum 18 marizomib cycles total) |
| Duration | ~7-8 months | Up to 18 cycles marizomib (~18 months) |
Outcomes
| Outcome | Type | Control | Intervention | HR / OR / RR | P-value |
|---|---|---|---|---|---|
| Overall survival (ITT); coprimary in MGMT-unmethylated subgroup | Primary | Median OS 17.0 mo (95% CI 15.9-18.6); 12-mo KM 71.9% | Median OS 16.5 mo (95% CI 15.4-17.6); 12-mo KM 71.1% | Stratified HR 1.04 | p=0.64 (log-rank) |
| OS MGMT-unmethylated (coprimary) | Secondary | 14.5 mo (95% CI 13.5-15.7) | 15.1 mo (95% CI 13.4-15.7) | HR 1.13 (95% CI 0.88-1.44) | p=0.27 |
| OS MGMT-methylated | Secondary | 25.5 mo (95% CI 21.1-31.3) | 29.4 mo (95% CI 20.7-32.1) | HR 0.86 (95% CI 0.60-1.24) | p=0.41 |
| PFS ITT | Secondary | 6.0 mo (95% CI 5.2-6.4) | 6.3 mo (95% CI 5.9-7.7) | HR 0.97 (95% CI 0.82-1.13) | p=0.67 |
| PFS MGMT-unmethylated | Secondary | 5.1 mo (95% CI 4.4-6.0) | 5.9 mo (95% CI 4.6-6.4) | HR 0.96 (95% CI 0.78-1.16) | p=0.64 |
| 12-month survival | Secondary | 71.9% | 71.1% | Equivalent | |
| Subgroup treatment-by-interaction test (MGMT) | Secondary | Reference | HR varies 0.91-1.16 across MGMT strata | p=0.665 for interaction | |
| Age <55 vs >55 interaction | Secondary | Reference | HR 0.97 vs 1.09 | p=0.495 | |
| Deaths on study (ITT) | Secondary | 267/374 (71.4%) | 271/375 (72.3%) | Equivalent | |
| Grade 3/4 any AE | Adverse | 174/362 (48%) | 250/371 (67%) | Markedly higher | |
| Grade 3/4 nervous system AE | Adverse | 20% | 33% | Anticipated BBB penetration | |
| Grade 3/4 psychiatric AE (confusion, hallucinations, anxiety) | Adverse | 3% | 14% | Characteristic marizomib toxicity | |
| Ataxia (G3/4) | Adverse | 0.3% | 5.1% | BBB-penetrant effect | |
| Hallucinations | Adverse | 0.3% G3/4 | 6.7% G3/4 | Novel AE | |
| Treatment-related deaths | Adverse | 1 | 8 (soft tissue necrosis, cerebral hemorrhage, leukoencephalopathy, etc.) | Higher with marizomib | |
| Dose withdrawal | Adverse | 12% | 36% | Tolerability-limiting | |
| Marizomib discontinuation for AE | Adverse | NA | 32% | High attrition |
Subgroup Analysis
The forest plot showed no subgroup with a treatment-by-interaction p<0.15. Hazard ratios clustered tightly around 1.0 across age, sex, KPS, baseline steroids, MMSE, surgical extent, and MGMT status. The prespecified hypothesis that marizomib would specifically benefit MGMT-unmethylated tumors (based on phase 1/2 signals) was not supported: HR 1.08 in that subgroup versus 0.91 in methylated tumors (interaction p=0.665). No biomarker or clinical feature identified a subpopulation of responders.
Criticisms
- Open-label design may have influenced subjective AE reporting and crossover to second-line therapy decisions
- Per-protocol exclusion of 52% of marizomib patients (largely for missing ECG evaluations after a protocol amendment) complicated sensitivity analysis
- CNS AEs indicate brain penetration, but quantitative tissue concentration of marizomib was not measured — cannot exclude insufficient intratumoral drug
- Trial predated the 2021 WHO CNS classification reclassifying some GBM histologies; IDH testing was recommended but not mandatory
- 8 treatment-related deaths in marizomib arm raise serious risk-benefit concerns even independent of efficacy
- No exploration of alternative proteasome inhibitors, lower doses, or combination with immunotherapy/TTF — proteasome pathway closure is based on this one trial
Funding
EORTC, Canadian Cancer Trials Group, Triphase Accelerator Corporation (marizomib supply), partial support from NCI grants
Based on: MIRAGE (Neuro-Oncology, 2024)
Content summarized and formatted by NeuroTrials.ai.