MGMT promoter methylation as a predictive biomarker for response to radiotherapy versus chemotherapy in malignant astrocytomas in the elderly: The NOA-08 trial

Reporter: Annemarie Fernandes
The Abramson Cancer Center of the University of Pennsylvania
Last Modified: June 3, 2012

Presenter: Wolfgang Wick
Presenter's Affiliation: University Hospital Heidelberg, Heidelberg, Germany


  • The incidence of glioblastoma has increased in elderly adults, and more than half of patients are more than 65-years-old. Increasing age is a strong predictor for poor prognosis with a median survival of 4-5 months in patients >70 years.
  • The current standard of care in elderly patients with glioblastoma (GBM) or anaplastic astrocytoma (AA) is resection or biopsy followed by involved-field radiotherapy (RT) (Keime-Guibert, NEJM 2007).
  • The role of chemotherapy in the elderly population is poorly defined.
  • The Stupp trial (NEJM, 2005) established the standard of care for treatment with post-operative radiotherapy with concurrent and adjuvant temozolomide, a DNA alkylating agent. This trial, however, excluded patients older than 70 years old.
    • In a subset analysis (Stupp, Lancet, 2009) of patients aged 61-70 years, there was no improvement in median survival using the combined modality approach compared to radiation therapy alone. However, there was a statistically significant prolongation in overall survival with adjuvant temozolomide compared to radiation therapy alone
    • Methylation of the promoter for methyl guanine methyl transferase (MGMT) was a major prognostic factor for improved survival and was predictive of benefit from chemotherapy
  • The Nordic Elderly Trial that was presented at ASCO 2010 randomized elderly (>60 years old) patients to radiation therapy with standard fractionation (60 Gy in 30 fractions) vs. a hypofractionated radiation therapy regimen (34 Gy in 10 fractions) vs. TMZ (200 mg/m2 for 5 days with cycles repeated every 28 days). Overall survival for the 60 Gy, 34 Gy, and TMZ treatment arms were 6.0, 7.5, and 8.3 months, respectively. The difference between chemotherapy versus the six-week course of RT, but not the shortened course of RT, was statistically significant.
  • Preliminary results of the NOA-08 trial comparing efficacy and safety of RT to temozolomide (TMZ) in elderly patients with newly diagnosed AA or GB were presented at ASCO 2010. Patients in the TMZ arm had an increased risk of death (HR=1.24 [95% CI: 0.94-1.63]). Furthermore, toxicity was more severe in those initially treated with chemotherapy.
  • The current study reports updated results from the NOA-08 trial and evaluates MGMT promoter methylation as a predictive biomarker for response to radiotherapy versus chemotherapy

Materials and Methods

  • This is a randomized control trial of elderly high-grade glioma patients (Anaplastic Astrocytoma (AA) or Glioblastoma Multiforme (GBM)) treated with immediate post-operative therapy with either TMZ vs. RT
  • Patients who progressed on the TMZ arm received salvage RT and patients who progressed on the RT arm received salvage TMZ.
  • Inclusion criteria: AA or GBM histology, Age > 65 years with a Karnofsky performance score > 60
  • Treatment
    • TMZ: 100mg/m2 PO/day for 7 days, every 14 days (one week on/one week off)
    • Radiotherapy: 54-60 Gy in 30 1.8-2 Gy fractions to the tumor bed with margin.
  • The primary endpoint was overall survival (OS). The trial sought to demonstrate the non-inferiority of TMZ compared with RT. Data is presented as an intent-to-treat analysis


  • 412 patients (39 AA, 373 GBM) were randomized and 373 patients were evaluated
  • Many patients received biopsy-only, rather than surgical resection
    • Biopsy: 36.5% RT and 41% TMZ
    • Incomplete Resection: 34.8% RT and 31.3% TMZ
    • Complete Resection: 28.6% RT and 27.2% TMZ
  • 48.5% of patients had tissue available for MGMT promoter analysis
  • Patients in the TMZ arm experienced higher rates of Grade 2-3 hematologic toxicity, fatigue and nausea/vomiting compared to the RT arm.
  • Overall Survival
    • Non-inferiority of TMZ compared with RT was significant (p<0.05).
    • When the preliminary results were published in 2010, patients in the TMZ arm had an increased risk of death. With longer follow-up, the survival curves cross and the patients in the RT arm have an increased risk of death (HR: 1.09, p=0.033).
  • Additionally, the event-free survival (EFS) curves also cross at about 7 months and result in overall improved EFS in the TMZ arm (HR: 1.15, p=0.043)
  • On multivariable analysis, extent of resection, but not age or diagnosis of AA were associated with prolonged EFS and OS.
  • MGMT
    • Overall, MGMT promoter methylation in tumor tissue was associated with prolonged OS (HR=0.62 [0.42-0.91], p=0.014) and EFS (HR=0.50 [0.36-0.68], p<0.001)
    • Patients with MGMT promoter methylation had longer EFS when treated with TMZ (8.4 months [5.5-11.7] versus RT (4.6 [4.2-5] months)
    • Patients without MGMT promoter methylation had longer EFS when treated with RT (4.6 [3.7-6.3] versus 3.3 [3-3.5] months). This effect persisted for OS.

Author's Conclusions

  • In high-grade glioma, TMZ is non-inferior to RT
  • Toxicity is an issue in some patients with TMZ treatment, but this may be due to the weekly alternative schedule of drug administration
  • MGMT is a strong predictive biomarker in elderly patients with malignant glioma
  • Despite difficulties with testing of MGMT, NOA-08 calls for testing to reduce toxicity, save costs and improve efficacy

Clinical Implications

  • Although only 50% of patients in the study had tissue available for analysis, the results show that assessment of MGMT is critical for determining treatment of elderly patients with high-grade glioma.
  • If MGMT status is unknown, patients should not be treated with TMZ upfront, as patients without MGMT methylation have poorer outcomes when treated with TMZ vs. RT.
  • It is important to realize that most patients in this study were treated with both therapies (RT and TMZ) as patients switched treatment with local progression. However, it is unclear how many patients completed both treatment modalities. This treatment schema may contribute to the crossing of the survival curves demonstrated with prolonged follow-up.
  • The rates of surgical resection in this study were low, but multivariable analysis demonstrates a clear correlation between extent of resection and OS/EFS. Elderly patients with good performance status should be offered maximal surgical resection when determined to be safe. When appropriate, we should avoid under-treating our elderly patients and recognize any inherent bias when discussing treatment options.