PDF
Citation
Print
Introduction
Glioblastoma (GBM) is the most aggressive and prevalent primary malignant brain tumor in adults, with incidence increasing with age. The age-adjusted incidence rate of GBM rises dramatically from 1.37 per 100,000 for adults aged 35-44 years to 8.11, 13.19, and 15.17 among those aged 55-64, 65-74, and 75-84 years, respectively [1]. Elderly patients with GBM (eGBM), comprising nearly half of all GBM cases, face a particularly poor prognosis with a median survival of approximately six months [2]. Currently, the standard treatment for GBM involves maximal safe resection followed by temozolomide (TMZ)-based chemoradiation, typically administered as conventionally fractionated radiotherapy (CFRT) at a total dose of 60 Gy in 30 daily fractions [3]. However, applying this regimen to eGBM patients remains controversial due to concerns about tolerance, given their high comorbidities, poor performance status, and increased vulnerability of aged brain tissues to radiation [4,5].
Several randomized trials have compared survival outcomes between CFRT and hypofractionated radiotherapy (HFRT) in eGBM patients, with HFRT delivering a total dose of 34-40 Gy over 2-3 weeks [6,7]. Although these trials demonstrated that HFRT can produce comparable outcomes to CFRT while reducing treatment-related sequelae, they were limited by being conducted in radiotherapy (RT)-alone setting without TMZ administration. More recently, research has established the survival benefit of TMZ-based chemoradiation compared to RT alone in eGBM patients, leading to the widespread adoption of this combination therapy [8,9]. Yet, despite this modern approach for eGBM patients, evidence guiding the choice between CFRT and HFRT in the context of TMZ-based chemoradiation remains scarce, with only retrospective and single-arm prospective studies available [5].
This study hypothesized that a more aggressive approach using CFRT plus TMZ may improve survival outcomes compared to HFRT plus TMZ in selected eGBM patients. We developed a graded prognostic assessment (GPA) model that integrates genomic characteristics of eGBM patients (eGBM-molGPA). Based on this model, we stratified patients into distinct risk groups and evaluated the survival outcomes of CFRT+TMZ, HFRT+TMZ, and RT alone within each group. The primary goal of this study was to provide evidence-based guidance for selecting the most appropriate treatment strategy for individual eGBM patients.
Materials and Methods
1. Study cohort
This multi-institutional retrospective study was conducted across four institutions in Korea (Yonsei Cancer Center, Seoul National University Hospital, Seoul National University Bundang Hospital, and SMG-SNU Boramae Medical Center). We identified elderly patients aged 65 years or older who were newly diagnosed with isocitrate dehydrogenase–wildtype GBM between January 2006 and December 2021. All patients underwent surgical resection or biopsy, followed by RT with or without concurrent and adjuvant TMZ. To ensure an accurate comparison of treatment efficacy between CFRT and HFRT, only patients who completed the planned course of RT were included in the analysis. Additionally, patients who were initially diagnosed with distant metastases or other malignancies, or who were followed up for less than 3 months were excluded.
2. Treatment and follow-up
Treatment and follow-up plan for every patient were developed by a multidisciplinary neuro-oncology team composed of radiation oncologists, neurosurgeons, medical oncologists, radiologists, and neurologists. After preoperative evaluation using magnetic resonance imaging (MRI), patients underwent maximal safe resection of the tumor. The extent of resection was categorized based on postoperative MRI within 48-72 hours as follows: gross total resection was defined as the absence of any enhancing tumor (< 1%); near-total resection as 1-5% residual tumor; subtotal resection as 5%-20%; partial resection as 20%-50%; and biopsy as greater than 50% residual tumor [10,11]. All patients were profiled for methylation of the O6-methylguanine–DNA methyltransferase (MGMT) promoter, as determined by methylation-specific polymerase chain reaction (PCR). Telomerase reverse transcriptase (TERT) promoter mutations were examined by pyrosequencing via PCR, focusing on C228T and C250T mutations. Next-generation sequencing (NGS) was performed using institution-specific panels: the TruSight Tumor 170 or TruSight Oncology 500 panel at Yonsei Cancer Center, the FIRST brain tumor panel at Seoul National University Hospital, and the Brain v1.0 panel at Seoul National University Bundang Hospital [12-14].
The decision regarding RT dose-fractionation schedules was made by the treating radiation oncologist. Patients treated with a daily dose of 1.8-2.0 Gy and total doses ranging from 59.4 Gy to 70 Gy were classified as receiving CFRT. Conversely, those treated with a daily dose of 2.5-3 Gy and total doses ranging from 39 Gy to 45 Gy were classified as receiving HFRT. During concurrent chemoradiation, TMZ was administered orally once daily at 75 mg/m2, commencing on the first day of RT and continuing throughout the treatment duration. Following a 4-week break, adjuvant TMZ was administered orally at 150-200 mg/m2 once daily for five consecutive days every 28 days, for up to six cycles or until disease progression occurred.
Regular follow-up assessments were conducted every 3 months for the first 2 years, and then every 4-6 months thereafter, in accordance with each institution’s policy. Tumor progression was defined according to the Response Assessment in the Neuro-Oncology (RANO) Working Group guidelines.
3. Statistical analysis
Overall survival (OS) was calculated from the date of resection to the date of death. Progression-free survival (PFS) was calculated from the date of resection to the date of progression or death. Surviving patients were censored at the last follow-up date. Survival outcomes were estimated using the Kaplan-Meier method. To enhance the robustness of our survival analysis, we extracted individual patient-level ‘pseudodata’ from published Kaplan-Meier curves of two landmark randomized trials: the German NOA-08 (NCT01502241) [15,16] and the Nordic trial (ISRCTN81470623) [7], using previously described methodologies [17]. Utilizing these datasets, we reconstructed Kaplan-Meier curves for the respective treatment groups and plotted them alongside the survival curves from our cohort. This approach enabled us to perform an integrated comparison of survival outcomes, including those in the TMZ monotherapy group.
Univariate and multivariate analyses were conducted using a Cox proportional hazards model. Clinical variables included Karnofsky performance scale (KPS) score, Charlson comorbidity index, subventricular zone (SVZ) involvement, temporalis muscle thickness (TMT), and surgical extent. Two radiation oncologists (H.I.L. and J.K.) measured TMT on preoperative contrast-enhanced T1-weighted MRI, following Furtner et al.’s methodology [18]. Patients were classified as having narrow or normal TMT status based on sex-specific mean TMT cut-off values (narrow: ≤ 6.3 mm for men, ≤ 5.2 mm for women; normal: > 6.3 mm for men, > 5.2 mm for women). For patients with available NGS data, we collected molecular profiles for epidermal growth factor receptor, cyclin-dependent kinase inhibitor 2A/2B, phosphatase and tensin homologue, α-thalassemia/mental retardation syndrome X-linked, TERT, and tumor protein p53 (TP53), as well as chromosomal arms 7 and 10. Based on these variables, we performed univariate and multivariate analyses and identified significant prognostic factors associated with OS. We then developed the eGBM-molGPA by assigning scores of 0, 0.5, and 1.0 proportional to the corresponding hazard ratio (HR) in the multivariable model, categorizing patients into three risk groups. All statistical analyses were performed using STATA software ver. 15.1 (StataCorp LP).
Results
1. Patient and treatment characteristics
Patient characteristics are summarized in Table 1. A total of 573 patients with eGBM were included in this study. The median age was 71 years (range, 65 to 86 years), and 22.2% of the patients were older than 75 years. Most patients underwent TMZ-based chemoradiation (n=491, 85.7%), of whom 328 (57.2%) received CFRT+TMZ and 163 (28.4%) received HFRT+TMZ. The remaining 82 patients (14.3%) received RT alone. The median total RT dose was 60 Gy (range, 59.4 to 70 Gy) for the CFRT+TMZ group, 45 Gy (range, 39 to 45 Gy) for the HFRT+TMZ group, and 45 Gy (range, 39 to 60 Gy) for the RT alone group. Molecular profiling revealed that 228 patients (39.8%) had methylation of the MGMT promoter. NGS data were available for 275 patients (48.0%). Of these, 152 (44.7%) had TERT promoter mutations and 94 (34.2%) had TP53 mutations.
2. Survival outcomes and prognostic factors
The median follow-up period was 12.3 months (range, 1.4 to 149.9 months) for all patients and 17.4 months (range, 3.3 to 149.9 months) for survivors. In patients treated with CFRT+TMZ, the median OS was 18.7 months, and 1- and 3-year OS rates were 77.1% and 25.5%, respectively. Patients treated with HFRT+TMZ had significantly lower survival outcomes, with a median OS of 15.1 months, and 1- and 3-year OS rates of 64.9% and 11.5%, respectively (CFRT+TMZ vs. HFRT+TMZ: HR, 1.52; p < 0.001). Patients treated with RT alone had the lowest survival outcomes, with a median OS of 10.4 months, and 1- and 3-year OS rates of 38.4% and 6.2%, respectively (CFRT+TMZ vs. RT alone: HR, 2.52; p < 0.001 and HFRT+TMZ vs. RT alone: HR, 1.64; p=0.001). PFS was highest in the CFRT+TMZ group, with a median PFS of 12.3 months, followed by the HFRT+TMZ group with a median PFS of 8.5 months, and the RT alone group with a median PFS of 8.0 months (CFRT+TMZ vs. HFRT+TMZ: HR, 1.17; p < 0.001; CFRT+TMZ vs. RT alone: HR, 1.82; p < 0.001; HFRT+TMZ vs. RT alone: HR, 1.24; p=0.129). Fig. 1 presents the combined survival curves by treatment group in the current study, as well as survival estimates from the NOA-08 and Nordic trials. Notably, regardless of MGMT promoter status, the CFRT+TMZ group in our cohort demonstrated the highest survival rate among all treatment groups across the three studies.
The univariate and multivariate Cox proportional hazard models for OS are presented in Table 2. In the multivariate analysis, the survival benefits of CFRT (p=0.013) and TMZ (p < 0.001) remained significant after adjusting for all clinical factors. KPS, surgical extent, and MGMT promoter status were identified as strong prognostic factors, with estimated HRs greater than 1.5 and p-values less than 0.001. Additionally, SVZ involvement, TMT, TERT promoter status, and TP53 status were identified as secondary prognostic factors, demonstrating estimated HRs between 1.3 and 1.5, with p-values ranging from 0.005 to 0.05.
3. GPA for elderly glioblastoma patients
Table 3 details the scoring criteria and elements of the eGBM-molGPA, along with a worksheet for calculating the total score. The three most significant factors (KPS, surgical extent, and MGMT promoter status) were assigned scores of 0 and 1.0 in the eGBM-molGPA. Secondary significant factors (SVZ involvement, TMT, TERT promoter status, and TP53 status) were assigned scores of 0 and 0.5. We categorized total scores into three risk groups: high (0.0-1.0), intermediate (1.5-2.5), and low (3.0-4.5). Survival outcomes for these risk groups are shown in Table 4 and Fig. 2. The median OS was 10, 16, and 24 months in the high-, intermediate-, and low-risk groups, respectively. All adjacent groups in this model demonstrated significantly different hazard functions (p < 0.001 for high- vs. intermediate-, high- vs. low-, and intermediate- vs. low-risk groups).
We then compared survival outcomes among treatment options within each risk group (Fig. 3, S1 Table). In the low- and intermediate-risk groups, CFRT+TMZ resulted in significantly higher OS compared to HFRT+TMZ or RT alone. For the low-risk group, the median OS was 27.7 months with CFRT+TMZ versus 18.7 months with HFRT+TMZ (p=0.023), while data for RT alone was unavailable. In the intermediate-risk group, the median OS was 17.5 months for CFRT+TMZ, 15.2 months for HFRT+TMZ, and 11.7 months for RT alone (p < 0.001). However, the high-risk group showed no significant OS differences among treatment options, with the median OS of 10.7 months for both CFRT+TMZ and HFRT+TMZ, and 9.4 months for RT alone (p=0.770).
Discussion
In this multi-institutional study, we analyzed treatment outcomes of different adjuvant regimens in eGBM patients. We found that CFRT+TMZ outperformed HFRT+TMZ or RT alone in both the current cohort and the combined cohort including two randomized trials. To identify suitable candidates for this aggressive treatment approach, we developed a novel eGBM-molGPA model incorporating KPS, surgical extent, MGMT promoter status, SVZ involvement, TMT, TERT promoter status, and TP53 status. This model revealed that CFRT+TMZ provided significant OS benefits to patients in the low- and intermediate-risk groups, but not in the high-risk group. The strength of our study lies in its comprehensive dataset from four participating institutions and two major randomized trials focusing on eGBM patients, providing an unprecedented depth of analysis for this patient subgroup.
The definition of “elderly” in GBM patients lacks consensus, with age cut-offs ranging from 60 to 75 years across studies. Among these, the most commonly used threshold is 65 years, which was adopted in our study. This threshold of 65 years is consistent with major landmark trials, including the NOA-08 and the Nordic trial (both included in our analysis), as well as the study by Perry et al. [7,8,15]. Elderly GBM patients are known to have worse prognoses than younger patients due to reduced treatment tolerance, poorer response to treatment, increased toxicity, and possibly altered tumor biology [2]. Considering that the median age at GBM diagnosis is 64 years, and elderly patients comprise more than half of the GBM population, establishing optimal treatment strategies for this specific subgroup is crucial [1]. However, the absence of standardized protocols has led to varied practice across institutions, highlighting the necessity for robust evidence for eGBM patients.
In this study, the CFRT+TMZ group demonstrated significant improvements in both OS and PFS compared to the HFRT+TMZ and RT alone group. While our reported OS was longer than that observed in the NOA-08 and Nordic trials, this disparity may be attributed to our exclusion of patients with incomplete treatment, a criterion not applied in the other two randomized trials. The optimal dose-fractionation regimen for adjuvant RT in eGBM patients has been controversial for the past several decades. Although a Canadian randomized controlled trial found no difference in survival outcomes between CFRT and HFRT [6], other studies have reported encouraging survival outcomes following more aggressive RT strategies in eGBM patients [19]. In a prospective trial involving 32 eGBM patients, Minniti et al. [20] demonstrated that CFRT+TMZ could be a feasible treatment option for this population. Concerns persist regarding higher treatment-related toxicity in MGMT-unmethylated eGBM patients receiving combined modality treatment. However, the findings from the EORTC 26062 study align with our results, indicating improved survival rates with this combined approach regardless of MGMT status [21].
Our analysis identified KPS score, surgical extent, and MGMT promoter status as the strongest prognostic factors, with SVZ involvement, TMT, TERT promoter, and TP53 status as additional prognostic indicators. These findings corroborate numerous previous studies. KPS score, surgical extent, and MGMT promoter status are well-established, robust prognostic factors in GBM patients [22,23]. SVZ involvement and narrow TMT have been recognized as novel prognostic biomarkers for GBM patients in recent years [24,25]. Additionally, TERT and TP53 mutations, frequently detected in GBM patients, have been consistently associated with unfavorable prognosis in previous reports [26,27]. The eGBM-molGPA model proposed in this study incorporated these prognostic factors and demonstrated its clinical utility by effectively stratifying survival outcomes among three risk groups. To our knowledge, this is the first clinically applicable, comprehensive GPA model for eGBM that integrates both clinicopathological and molecular parameters. Furthermore, the eGBM-molGPA model demonstrated that low- and intermediate-risk patients benefited most from CFRT+TMZ, whereas high-risk patients did not derive OS benefit from the treatment approach. It is important to note, however, that the limited number of patients in the high-risk group may have reduced the statistical power of this subgroup analysis. In the intermediate-risk group, CFRT+TMZ showed significant superiority to HFRT+TMZ, although the absolute survival benefit was modest at 2.3 months. Nevertheless, given the poor prognosis associated with eGBM, even such a modest survival gain should not be underestimated [28].
This study has several limitations. First, we did not collect data on treatment-related toxicities, such as hematologic and neurological toxicities, which previous trials identified as factors contributing to the inferiority of CFRT+TMZ compared to HFRT+TMZ or TMZ alone in eGBM patients [29]. Second, our combined analysis incorporated survival data from the current study, the NOA-08 trial, and the Nordic trial. However, the treatment periods differed significantly across these studies. During this time period, advancements in RT techniques, surgical procedures, brain imaging technologies, molecular marker knowledge, and overall patient care may have influenced treatment outcomes, complicating direct comparisons between studies. These factors should be carefully considered when interpreting the combined results. Despite these limitations, this study represents the largest, multi-institutional cohort focusing on eGBM patients, a subgroup for whom there is a paucity of data in the modern era. We developed a comprehensive model incorporating clinicopathological and molecular parameters, which could guide evidence-based decisions for eGBM patients. Although the eGBM-molGPA model requires further validation, we believe that our robust finding shed light on this promising approach and contribute to establishing optimal treatment strategies for this population.
In conclusion, a more aggressive treatment approach of CFRT+TMZ was associated with improved OS in carefully selected eGBM patients. The novel eGBM-molGPA model can serve as a clinical tool for choosing wisely between treatment regimens. Further prospective studies are warranted to validate our findings and establish the optimal treatment strategies for eGBM patients.
XML Download