Phase III trial of CCNU/temozolomide (TMZ) combination therapy vs. standard TMZ therapy for newly diagnosed MGMT-methylated glioblastoma patients: the CeTeG/NOA-09 trial presented by Ulrich Herrlinger for the Neurooncology Working Group (NOA) of the German Cancer Society.
The summary below written by SW, who was in attendance at the presentation by Ulrich Herrlinger and took photos of the slides presented.
The CeTeG trial (also known as NOA-09) is a randomized phase 3 trial for newly diagnosed glioblastoma with methylated MGMT promoter, testing CCNU (lomustine) combined with temozolomide (TMZ) versus TMZ alone. This trial was conducted at 17 centers in Germany and was a follow-up to a non-randomized phase 2 trial which had results published in 2006 and 2009. The CeTeG trial was relatively small for a phase 3 trial, with a sample size calculation of 128 patients total, and this sample size was based on expectations of a significant increase in survival rate at 2 years as seen in the phase 2 trial compared to historical controls.
Patients in this trial had relatively good prognosis, with high rates of complete resection and high average KPS. Overall the arms were well balanced, with the only significant imbalance between the two arms being gender, which was not prognostically relevant.
In the combination arm receiving CCNU + TMZ, cycles were 6 weeks in length, with 100 mg/m2 oral CCNU given on day 1 of each cycle and TMZ on days 2-6 of each cycle, with a starting TMZ dose of 100 mg/m2 and possible escalation up to 200 mg/m2 in later cycles. Cycle 1 starts at the same time as radiation.
In the control arm of TMZ alone, cycles were 4 weeks in length, and used the standard TMZ schedule (daily at a dose of 75 mg/m2 during radiation, and 150 mg/m2 on days 1-5 of the first adjuvant cycle and possible escalation up to 200 mg/m2 in later cycles.
Importantly, this trial achieved its primary endpoint of increased overall survival. Survival in the CCNU + TMZ arm was statistically superior to TMZ alone, with a p value of 0.049. Hazard ratio for death from any cause was 0.6 in the CCNU + TMZ arm.
Median reported survival was 46.9 months for CCNU + TMZ versus 30.4 months for TMZ alone, a difference of 16.5 months. As seen in the Kaplan-Meier survival estimates, the curves did not separate until after the 2 year mark. 1, 2, 3, 4, and 5 year survival rate was 88.8, 71.4, 57.4, 48.8 and 34% in the CCNU + TMZ arm versus 84.4, 65.4, 42.3, 31.4 and 27.7% in the TMZ arm. Differences in the survival rate between the two arms were greatest at the 3 and 4 year mark, with 15% more patients surviving to 3 years and 17.4% more patients surviving to 4 years in the combination arm versus the TMZ alone arm.
Given the significant overall survival differences, it's surprising to note that progression-free survival was not significantly different between the two arms (p=0.41), although the progression-free survival curves separated somewhat at about 2 years (a phenomenon also seen in the overall survival curves), after which time CCNU + TMZ shows a slight superiority over TMZ alone. Some potential explanations given by the authors for the lack of a strong PFS signal included: "problems with PFS assessment according to RANO?" including potential undetected pseudoprogressions; and "long-term effects of CCNU?", noting that in studies of low grade gliomas treated with CCNU, responses were sometimes seen months or years after the end of therapy.
The percentage of patients receiving further lines of therapy after progression was similar in both arms (59.1% in the CCNU + TMZ arm, 63.5% in the TMZ arm). More patients underwent re-resection in the CCNU + TMZ arm (31.8% versus 22.2%), and more patients received re-irradiation in the TMZ alone arm (23.8% versus 18.2%). More patients in the TMZ alone arm received further chemo or targeted agent therapy (60.3% versus 48.5%). The percentage of patients receiving bevacizumab after progression was similar in both arms (27% in the TMZ alone arm, 30.3 % in the TMZ + CCNU arm). The authors concluded that differences in treatment after progression are not an explanation for the superior survival in the TMZ + CCNU arm.
Combination therapy with TMZ + CCNU approximately doubled the rate of low-grade (but not high-grade) hematoxicity (including neutropenia and thrombocytopenia) and nausea. However no deaths due to treatment toxicity were observed, and no severe infections, liver failure, or lung fibrosis. More brain edema was observed in the combination arm, and more low-grade alopecia (patchy hair loss).
The authors concluded by noting that acute toxicity of the combination treatment was rare, and importantly, "the primary aim of CeTeG/NOA-09 was achieved: the OS superiority of CCNU/TMZ for MGMT promoter methylated newly diagnosed GBM could be demonstrated".
This may well be the most significant trial outcome reported at the 2017 SNO conference. Since TMZ was approved for newly diagnosed glioblastoma in 2005, it has been exceedingly rare for a phase 3 trial in the newly diagnosed GBM setting to achieve statistically significant prolongation of survival with a novel regimen. More work needs to be done to explain why progression-free survival benefit was more modest than overall survival benefit in this trial. A potential hypothesis for the improved survival results for the combination therapy is a possible synergistic interaction between TMZ and CCNU, whereby cells escaping sensitivity to TMZ through mismatch repair defects are thereby rendered more sensitive to the CCNU treatment (Stritzelberger et al. 2017). Now that the results of CeTeG have been reported to the international neuro-oncology community, the mechanisms behind the improved outcomes with the combination chemotherapy will likely become the subject of more intense investigation.
