Abstract
Background
Glioblastoma (GBM) is the most aggressive primary brain tumor with dismal prognosis. Patients residing in rural and Appalachian regions encounter multiple barriers to timely diagnosis and access to multidisciplinary neuro-oncology care, but whether these barriers translate into worse survival remains uncertain. West Virginia (WV), a predominantly rural Appalachian state, is served by a primary academic medical center (WV University Ruby Memorial Hospital) with a dedicated neuro-oncology program. We performed a retrospective analysis of the electronic health record to characterize overall survival and prognostic factors among WV residents diagnosed with GBM.
Methods
Retrospective cohort of 380 adults (≥ 18 years) with pathologically confirmed Isocitrate Dehydrogenase-wildtype (IDH-WT) GBM (2015–2025). Rurality defined by ZIP-code RUCA. Cox proportional hazards models were used to assess the association between overall survival (OS) and age, sex, treatment, rurality, and O⁶-methylguanine-DNA methyltransferase (MGMT) promoter methylation status.
Results
Median OS was 12.7 months. Age ≥ 65 was associated with worse survival (15.9 vs. 9.1 months; p < 0.001). Rural and non-rural survival was equivalent (12.5 vs. 12.7 months; p = 0.87). Temozolomide (TMZ) use significantly improved OS (14.0 vs. 6.2 months; p < 0.001). Gross total resection and MGMT promoter methylation were both associated with significantly improved overall survival (p < 0.001 and p = 0.0005, respectively).
Conclusions
In a predominantly rural state served by a primary academic neuro-oncology program, median overall survival for GBM was 12.7 months indicating centralized care eliminates rural-urban survival gaps in GBM.
Keywords: Glioblastoma, Rural health disparities, Survival analysis, Temozolomide, Centralized care, Real-world evidence
Introduction
Glioblastoma (GBM) is the most common and lethal primary malignant brain tumor in adults, accounting for approximately 51.5% of all malignant central nervous system (CNS) tumors in the United States [1]. The age-adjusted incidence is about 4.50 per 100,000 population, with peak occurrence between 65 and 75 years [1]. Despite advances in molecular classification (e.g., Isocitrate Dehydrogenase-wildtype (IDH-WT), TERT promoter mutation, EGFR amplification), prognosis remains dismal [2]. The landmark protocol that includes maximal safe resection followed by radiotherapy (60 Gy in 30 fractions) with concurrent TMZ (75 mg/m² daily) and six cycles of adjuvant TMZ (150–200 mg/m²) has established a median overall survival (mOS) of about 14.6 months in clinical trial settings [3]. However, real-world cohorts consistently report median OS of 10–13 months, with 5-year survival below 10% [4, 5].
Treatment completion is a critical determinant of outcome. Gross total resection (GTR), when feasible, improves OS by 3–6 months compared to biopsy alone, while omission of TMZ or radiation independently worsens prognosis [6]. Yet, sociodemographic and geographic barriers significantly impair adherence. Rural patients, who comprise 19.3% of the U.S. population but 60–70% in states like West Virginia (WV), face prolonged travel distances (often > 100 miles), delayed specialist referral, lower clinical trial enrollment, and higher rates of uninsured or underinsured status [7, 8]. These factors contribute to treatment delays, incomplete multimodal therapy, and inferior survival across multiple cancer types [9].
In neuro-oncology, rural-urban disparities are less well-characterized. Surveillance, Epidemiology, and End Results (SEER) registry analyses suggest rural residence increases mortality risk in “brain and other CNS cancers” (HR 1.08–1.15), but aggregate data obscure GBM-specific effects [10]. A 2021 meta-analysis of 12 studies found rural patients with high-grade gliomas were 20% less likely to receive standard chemoradiation, though survival differences were inconsistent after adjustment for treatment receipt [11]. Notably, a single-institution study from an NCI-designated cancer center reported equivalent OS between rural and urban GBM patients when treated within a centralized multidisciplinary program, suggesting that access to high-volume neuro-oncology care may mitigate geographic inequity [12].
WV exemplifies rural healthcare challenges: 64% of its population resides in rural counties, the highest proportion in the eastern U.S., with persistent deficits in subspecialty access, broadband connectivity, and socioeconomic resources [13]. WVU Ruby Memorial Hospital, and the state’s neuro-oncology referral center, serves a catchment area of over 1.8 million across 55 counties. Mostly, all complex GBM cases, regardless of rurality, are referred to WVU Ruby Memorial Hospital and WVU Cancer Institute, enabling standardized diagnostic workup including molecular work up, treatment planning, and longitudinal follow-up.
We conducted a retrospective cohort study of 380 adults with pathologically confirmed IDH-WT GBM treated at WVU Ruby Memorial Hospital from 2015 to 2025 to evaluate survival differences by treatment (surgery, radiation, TMZ), and demographic factors (age, sex). Kaplan–Meier (KM) estimates and Cox proportional-hazards models were used to assess independent predictors of OS. We hypothesize that the survival benefits observed with centralized neuro-oncology care, are driven predominantly by uniform treatment use, could eliminate previously reported rural disparities in GBM outcomes, with important implications for regional referral networks and health policy in underserved regions.
Methods
Study design and population
This retrospective cohort study included adults (≥ 18 years) with pathologically confirmed IDH-WT GBM treated at WVU Ruby Memorial Hospital (Morgantown, WV) between January 1, 2015, and August 31, 2025. Patients were identified from the hospital medical records under ICD-10 code C71 (malignant neoplasm of brain). Of 1,015 patients initially screened, 380 had pathologically confirmed IDH-WT GBM (Fig. 1). Additionally, we conducted individual chart validation of GBM status and treatment details. Rural residence was defined using ZIP-code-level Rural-Urban Commuting Area (RUCA) codes version 3.10 (USDA Economic Research Service), with codes 4–10 classified as rural [14].
Fig. 1.
Flow diagram of patient selection for the study cohort of IDH-wildtype glioblastoma. Patients were initially identified from institutional records using ICD-10 code C71 (malignant neoplasm of brain) between the study period, yielding a total of 1015 cases. All cases were then assessed for confirmation of IDH-wildtype glioblastoma (WHO grade 4) based on integrated histopathologic, molecular, and clinical criteria (including IDH1/IDH2 sequencing or immunohistochemistry demonstrating wild-type status). A total of 380 patients met the inclusion criteria for confirmed newly diagnosed IDH-wildtype glioblastoma and were included in the final study population. The remaining 635 cases were excluded as they represented non-glioblastoma entities (e.g., other gliomas, metastases, non-malignant tumors, or IDH-mutant gliomas). This selection process ensured a homogeneous cohort focused exclusively on IDH-wildtype glioblastoma for subsequent survival and prognostic analyses. Numbers at each step indicate the sample size retained or excluded
Data sources and variables
Data were abstracted from electronic health records (Epic Systems). Demographic variables included age at diagnosis, sex, and residential ZIP code. Clinical variables encompassed surgery, radiation therapy, and TMZ (TMZ) use.
Overall survival (OS) was defined as the time from diagnosis of GBM to death from any cause. Patients known to be alive were followed until the date of last contact or August 31, 2025, whichever occurred first.
Geospatial analysis
Cancer centers were geocoded to their zip code centroids to display their position across the state. Cancer centers were color coded based on WVU affiliation. Patients were geocoded to their zip code centroids and displayed using heat density. A second map was created showing RUCA rurality by zip code with the cancer center locations for comparison to patient densities. All maps were created using ESRI ArcGIS Pro version 3.3.2 (2025).
Statistical methodology for the survival curves
Descriptive statistics were used to summarize the patients’ characteristics. Categorical data were described using contingency tables with counts and percentages. Continuous variables were summarized using median or mean with standard deviation. In the data analysis of the survival outcome, the KM method and log-rank test were used in the univariate data analysis, and a Cox proportional hazards regression model was used in the multivariate data analysis. Statistical inferences were based on two-sided tests at a significance level of p < 0.05.
Results
Patient cohort and baseline characteristics
Between January 1, 2015, and August 31, 2025, a total of 1,015 patients were initially identified in the WVU Ruby Memorial Hospital (Morgantown, WV) using the ICD-10 code C71 (malignant neoplasm of brain). After inclusion of IDH-WT GBM pathology, the final analytic cohort consisted of 380 adult patients with newly diagnosed, pathologically and molecularly confirmed GBM.
Males predominated (n = 221, 58.2%), consistent with the known epidemiology of GBM, and 183 patients (48.2%) were older than 65 years at the time of diagnosis. Receipt of standard-of-care therapies was as follows: TMZ (TMZ) in 254 patients (66.8%), any surgical resection in 336 patients (88.4%), and radiation therapy in 226 patients (59.5%). Detailed patient characteristics including age, sex, rural versus non-rural residence, treatment variables, and corresponding univariate survival estimates are presented in Table 1.
Table 1.
Baseline characteristics and univariate Kaplan–Meier survival estimates stratified by key clinical and demographic variables (n = 380)
| Variable | Level | N | Events | Median OS, mo (95% CI) | 1-yr OS % (95% CI) | 2-yr OS % (95% CI) | Log-rank p |
|---|---|---|---|---|---|---|---|
| All patients | 380 | 230 | 12.69 (11.61, 14) | 0.53 (0.47, 0.59) | 0.25 (0.2, 0.32) | ||
| Sex | 0.61 | ||||||
| Female | 159 | 92 | 13.57 (11.38, 18) | 0.56 (0.48, 0.65) | 0.24 (0.17, 0.35) | ||
| Male | 221 | 138 | 12.13 (10.92, 14) | 0.5 (0.43, 0.59) | 0.26 (0.2, 0.34) | ||
| Age, years | < 0.0001 | ||||||
| ≤ 65 | 197 | 117 | 15.9 (13.57, 19.57) | 0.63 (0.56, 0.71) | 0.35 (0.28, 0.44) | ||
| > 65 | 183 | 113 | 7.67 (5.38, 10.39) | 0.39 (0.32, 0.49) | 0.13 (0.08, 0.22) | ||
| RUCA rural | 0.87 | ||||||
| 1–3 | Non-rural | 204 | 121 | 12.69 (11.51, 16.03) | 0.54 (0.46, 0.62) | 0.26 (0.19, 0.35) | |
| 4–10 | Rural | 176 | 109 | 12.52 (10.2, 16.36) | 0.52 (0.44, 0.61) | 0.25 (0.18, 0.34) | |
| Surgery | 0.31 | ||||||
| No (biopsy) | 44 | 16 | 12.2 (10.2, NA) | 0.55 (0.38, 0.8) | 0.31 (0.15, 0.62) | ||
| Yes (GTR/STR) | 336 | 214 | 12.69 (11.57, 14.82) | 0.52 (0.46, 0.59) | 0.25 (0.2, 0.31) | ||
| Radiation | 0.19 | ||||||
| No | 154 | 75 | 11.18 (7.38, 15.57) | 0.46 (0.37, 0.57) | 0.25 (0.17, 0.38) | ||
| Yes | 226 | 155 | 13.54 (11.97, 16.52) | 0.56 (0.5, 0.64) | 0.26 (0.2, 0.34) | ||
| Temozolomide | < 0.0001 | ||||||
| No | 126 | 58 | 6.66 (3.61, 12.2) | 0.37 (0.27, 0.51) | 0.22 (0.13, 0.36) | ||
| Yes | 254 | 172 | 14 (12.72, 16.89) | 0.58 (0.52, 0.65) | 0.27 (0.22, 0.35) |
P-values reflect log-rank tests
Survival outcomes
The mOS for the entire cohort was 12.69 months (95% CI: 11.61–14.00), with one-year and two-year OS rates of 53% (95% CI: 47–59%) and 25% (95% CI: 20–32%), respectively (Fig. 2). These figures are consistent with contemporary population-based GBM cohorts treated in the TMZ era.
Fig. 2.
Overall survival in 380 adults with newly diagnosed, pathologically confirmed glioblastoma treated at WVU Ruby Memorial Hospital, 2015–2025. Kaplan–Meier estimate of overall survival from date of histologic diagnosis. mOS was 12.69 months (95% CI: 11.61–14.00). One-year survival probability was 53% (95% CI: 47–59%) and two-year survival probability was 25% (95% CI: 20–32%). Tick marks indicate censored observations. Number at risk is shown below the x-axis at 6-month intervals
Univariate KM analyses confirmed age, receipt of TMZ, GTR and MGMT promoter status as the dominant prognostic factors, whereas sex, and rural status did not significantly influence survival in this dataset (full univariate metrics are provided in Table 1). Patients aged ≤ 65 years (n = 197) experienced a mOS of 15.9 months (95% CI: 13.57–19.57) compared to only 7.67 months (95% CI: 5.38–10.39) for those older than 65 years (n = 183; log-rank p < 0.0001; Fig. 3); one-year OS rates were 63% versus 39% in the younger and older groups, respectively. Patients who received TMZ (n = 254) achieved a median OS of 14.0 months (95% CI: 12.72–16.89) versus 6.66 months (95% CI: 3.61–12.20) among non-recipients (n = 126; log-rank p < 0.0001; Fig. 4), with corresponding one-year OS rates of 58% versus 37%. Finally, survival did not differ significantly by sex (males: median OS 12.13 months, 95% CI: 10.92–14.00, n = 221; females: 13.57 months, 95% CI: 11.38–18.00, n = 159; log-rank p = 0.61; Fig. 5). Rurality did not significantly affect survival. mOS was 12.5 months for rural patients versus 12.7 months for non-rural patients (log-rank p = 0.87), indicating comparable outcomes between these groups (Fig. 6). KM analysis demonstrated superior survival in the GTR group, with a median overall survival of 20.1 months (95% CI: 17.5–24.0) versus 12.3 months (95% CI: 10.5–14.8) in the non-GTR including subtotal resection or biopsy group (log-rank p < 0.0001). This pronounced benefit underscores GTR as a key modifiable prognostic factor in this molecularly defined population (Fig. 7). MGMT promoter methylation was strongly associated with improved overall survival. KM revealed superior survival in the MGMT-methylated group compared to the MGMT-unmethylated group, with a mOS of 16.8 months (95% CI: 14.5–19.5) versus 11.9 months (95% CI: 10.2–13.8) (log-rank p = 0.0005). These findings affirm MGMT promoter methylation as a key predictive biomarker for alkylating agent response and an independent favorable prognostic factor in IDH-wildtype glioblastoma (Fig. 8).
Fig. 3.
Kaplan–Meier estimates of overall survival stratified by age at diagnosis (≤ 65 years, n = 197 vs. >65 years, n = 183) in 380 patients with newly diagnosed, pathologically confirmed glioblastoma. mOS was 15.9 months (95% CI: 13.57–19.57) for patients ≤ 65 years and 7.67 months (95% CI: 5.38–10.39) for those > 65 years (log-rank p < 0.0001). Shaded regions represent 95% confidence intervals. Tick marks indicate censored observations. The number at risk is displayed below the x-axis at 6-month intervals
Fig. 4.
Kaplan–Meier estimates of overall survival stratified by TMZ receipt (yes, n = 254 vs. no, n = 126) in 380 patients with newly diagnosed, pathologically confirmed glioblastoma. MOS was 14.0 months (95% CI: 12.72–16.89) for patients who received TMZ and 6.66 months (95% CI: 3.61–12.20) for those who did not (log-rank p < 0.0001). Tick marks indicate censored observations. The number at risk is displayed below the x-axis at 6-month intervals
Fig. 5.
Kaplan–Meier estimates of overall survival stratified by sex (female, n = 159 vs. male, n = 221) in 380 patients with newly diagnosed, pathologically confirmed glioblastoma. MOS was 13.57 months (95% CI: 11.38–18.00) for females and 12.13 months (95% CI: 10.92–14.00) for males (log-rank p = 0.61). Tick marks indicate censored observations. The number at risk is displayed below the x-axis at 6-month intervals
Fig. 6.
Kaplan–Meier estimates of overall survival stratified by rural residence (non-rural, n = 204 vs. rural, n = 176) in 380 patients with newly diagnosed, histologically confirmed glioblastoma. Median overall survival was 12.69 months (95% CI: 11.51–16.03) for non-rural patients and 12.52 months (95% CI: 10.20–16.36) for rural patients (log-rank p = 0.87). Tick marks indicate censored observations. The number at risk is displayed below the x-axis at 6-month intervals
Fig. 7.
Kaplan–Meier estimates of overall survival stratified by extent of resection (GTR = Yes vs. GTR = No) in patients with IDH-wildtype glioblastoma. The red curve represents patients who achieved gross total resection (GTR = Yes), while the black curve represents those with subtotal resection or biopsy (GTR = No). Overall survival probability is plotted against time in months. Median overall survival was 20.1 months (95% CI: 17.5–24.0) for the GTR group versus 12.3 months (95% CI: 10.5–14.8) for the non-GTR group (log-rank p < 0.0001). Tick marks indicate censored observations. The number at risk is displayed below the x-axis at selected intervals (e.g., every 10–20 months). Error bars represent 95% confidence intervals at key time points
Fig. 8.
Kaplan–Meier estimates of overall survival stratified by MGMT promoter methylation status (methylated vs. unmethylated) in patients with IDH-wildtype glioblastoma. The red curve represents patients with MGMT promoter methylated tumors, while the black curve represents those with MGMT promoter unmethylated tumors. Overall survival probability is plotted against time in months. Median overall survival was 16.8 months (95% CI: 14.5–19.5) for the methylated group versus 11.9 months (95% CI: 10.2–13.8) for the unmethylated group (log-rank p = 0.0005). Tick marks indicate censored observations. The number at risk is displayed below the x-axis at selected intervals (e.g., every 10–20 months). Error bars represent 95% confidence intervals at key time points
Geospatial distribution and access
Figure 9 illustrates the geospatial distribution of GBM patient density per 1,000 population across WV ZIP Code Tabulation Areas (ZCTAs), calculated using 380 incident IDH-wildtype GBM cases diagnosed at WVU Ruby Memorial Hospital from 2015 to 2025 and 2023 U.S. Census ZCTA population estimates. Patient density is represented by a heat map gradient from sparse (blue) to dense (red/yellow). Cancer treatment centers are overlaid by affiliation: WVU-affiliated centers (green dots), non-affiliated centers (blue dots), and the Mary Babb Randolph Cancer Center at J.W. Ruby Memorial Hospital in Morgantown (yellow star). Cases were predominantly concentrated in central and northern counties, with the highest densities observed in proximity to the WVU hub in Morgantown, reflecting centralized referral patterns to WVU-affiliated facilities, which treated the vast majority of patients. Non-affiliated centers managed a smaller proportion of cases. Figure 10 complements this by depicting RUCA classification by ZCTA, with rural areas shaded blue and non-rural areas shaded light blue, overlaid with the same cancer center markers. The map demonstrates that the majority of WV is classified as rural, yet WVU-affiliated centers are distributed across both rural and non-rural zones, facilitating statewide access despite geographic challenges. The concentration of cases near Morgantown (as shown in Fig. 9) highlights the dominant role of the central academic center in GBM care within a predominantly rural state.
Fig. 9.
Choropleth map of glioblastoma patient density by ZCTA in West Virginia. Heatmap shows case density (380 incident IDH-wildtype GBM cases diagnosed at WVU Ruby Memorial Hospital, 2015–2025; 2023 Census ZCTA estimates), with color gradient from sparse (blue) to dense (red/yellow). Cancer centers overlaid: green dots = WVU-affiliated, blue dots = non-affiliated, yellow star = Mary Babb Randolph Cancer Center at J.W. Ruby Memorial Hospital (Morgantown). WV border outlined in black. Highest densities cluster in central/northern counties near the WVU hub
Fig. 10.
Choropleth map of RUCA rural-urban classification by ZCTA in West Virginia with cancer center overlay. Rural areas (RUCA-based) shaded blue; non-rural shaded light blue. Cancer centers geocoded by ZIP centroid: green dots = WVU-affiliated, blue dots = non-affiliated, yellow star = Mary Babb Randolph Cancer Center at J.W. Ruby Memorial Hospital. WV border outlined in black. Map illustrates predominant rural classification statewide, with WVU-affiliated centers distributed across rural and non-rural zones
Discussion
This retrospective cohort study of 380 adults with pathologically confirmed IDH-WT GBM treated at WVU Ruby Memorial Hospital from 2015 to 2025 demonstrates that centralized neuro-oncology care in a predominantly rural state like WV can achieve equitable survival outcomes. Our findings support our hypothesis that standardized referral to a high-volume center mitigates geographic disparities, with TMZ adherence and GTR emerging as the primary modifiable prognostic factor.
The absence of survival differences aligns with prior single-institution data from NCI-designated centers, where equivalent outcomes were reported between rural and urban GBM patients under multidisciplinary care [12]. Similar equity has been observed in other tertiary rural catchment analyses, including a Vermont single-center study where urban and rural GBM patients had comparable presentation, treatment, and survival due to centralized access [15]. In contrast, population-based SEER analyses have documented modest mortality risks for rural brain tumor patients (HR 1.08–1.15) [10], attributed to barriers such as travel distance, delayed referrals, and lower treatment adherence [11]. County-level disparities in GBM care further highlight rural deficits in neurosurgeon availability and adjuvant therapy receipt [16, 17], yet high-volume facilities consistently attenuate these effects [18]. Our cohort’s centralized structure, wherein nearly all complex GBM cases are referred to WVU Ruby Memorial, the state’s primary provider of advanced brain tumor services possible accounts for this equity. Geospatial mapping confirmed case clustering facilitating timely multimodal therapy. This referral pattern underscores the value of hub-and-spoke models in underserved regions, where rural patients comprise 46.3% of cases but face equivalent access to standard care, consistent with studies linking high facility volumes to improved GBM survival [19, 20].
TMZ’s prognostic dominance (median OS 14.0 vs. 6.66 months in recipients vs. non-recipients) reinforces its established role in the Standard of care regimen [3], even in real-world settings where median OS is 2–4 months shorter than trial benchmarks [4, 5]. Non-receipt, observed in 33.2% of patients, may reflect toxicity, comorbidities, or logistical challenges, particularly in rural subsets with prolonged commutes. A meta-analysis of TMZ adherence in GBM confirms 20–30% OS gains with completion, with benefits persisting beyond 6 cycles in select cohorts [21, 22]. Although univariate analysis did not stratify TMZ by molecular markers, the effect size suggests broad applicability. Age > 65 years conferred a 2-fold survival detriment (7.67 vs. 15.9 months), mirroring CBTRUS epidemiology where incidence peaks in this group [1]. Notably, neither surgical extent nor radiation completion independently influenced OS, possibly due to high resection rates (88.4%) and selection bias in a referral center where biopsy-only cases (n = 44) were limited to inoperable tumors.
Furthermore, KM analysis revealed that MGMT promoter methylation was strongly associated with prolonged overall survival (p = 0.0005), consistent with its established role as a key prognostic and predictive biomarker in IDH-WT GBM, enhancing sensitivity to temozolomide. Similarly, GTR conferred a highly significant survival advantage (p < 0.0001), underscoring the importance of maximal safe cytoreduction even in real-world cohorts with high resection rates (88.4%).
These results advance understanding of rural GBM care by providing state-specific evidence from WV. Strengths include comprehensive capture via institutional medical records, minimizing loss to follow-up, and geospatial integration to contextualize access. However, limitations temper generalizability: reliance on aggregate data precluded granular assessment of treatment discontinuation reasons and the single-center design may not reflect decentralized systems. Future studies will incorporate multivariable modeling and explore interventions like tele-neuro-oncology to enhance rural adherence, particularly given persistent reporting gaps in GBM trials that may exacerbate disparities in real-world populations [23].
Limitations
This study has limitations inherent to its retrospective design. Although treatment completion was initially ascertained at the aggregate level from electronic health records, individual chart review allowed for more precise assessment of surgical extent, molecular markers (IDH-WT confirmation per WHO criteria and MGMT promoter methylation status), and other clinical details. Consequently, granular details, such as reasons for treatment discontinuation (e.g., toxicity, patient preference, or logistical barriers), could not be evaluated. Second, Second, although surgical intervention was initially identified using ICD-10-PCS/CPT/procedure codes that did not permit clear distinction between biopsy and resection extents, individual chart review mitigated this limitation and enabled analysis of GTR versus lesser extents, which showed significant univariate survival benefit. Third, loss to follow-up or incomplete treatment records could not be fully assessed, potentially underestimating attrition in rural cohorts with greater travel burden. Although chart review improved data granularity for key prognostic factors (e.g., MGMT methylation, extent of resection), preoperative Karnofsky Performance Status (KPS) was not uniformly documented, limiting multivariable adjustment for performance status. Despite these constraints, the centralized referral pattern at WVU Ruby Memorial Hospital, serving nearly all complex GBM cases statewide, minimizes selection bias related to access to neuro-oncology expertise and supports the generalizability of findings within similar rural referral networks. However, the single-center data may not capture patients treated entirely outside the WVU network or those unable to access referral due to barriers (e.g., extreme rurality or socioeconomic factors), introducing potential selection bias toward patients with better access and possibly overestimating rural-urban outcome equity.
Conclusion
Centralized care eliminates rural GBM survival disparities, with chemotherapy, gross total resection and MGMT promoter methylation as the positive prognostic factors.
By enabling routine biomarker testing (e.g., MGMT, IDH) and comprehensive guideline-concordant care, even for rural patients, centralized systems achieve equivalent survival between rural and non-rural cohorts. These findings advocate for expanded regional referral networks and policy support for rural oncology infrastructure to narrow persistent inequities in high-grade glioma outcomes.
Acknowledgements
The work has benefited from detailed external peer review by Dr. Mark Gilbert (Scientist Emeritus, National Institutes of Health) and Mr. David Arons (CEO, National Brain Tumor Society) as their comments significantly strengthened its real-world relevance.
Author contributions
Conceptualization: Sonikpreet Aulakh. Methodology: Sonikpreet Aulakh, Sijin Wen. Formal analysis and investigation: Sonikpreet Aulakh, Timothy Shaun Dotson, Morgan W. Denney, Sijin Wen, Matthew Armistead. Writing – original draft preparation: Sonikpreet Aulakh, Emily Pack. Writing – review and editing: all authors. Supervision: Sonikpreet Aulakh. All authors read and approved the manuscript.
Funding
The author(s) received no specific funding for this work.
Data availability
The data that support the findings of this study are available from West Virginia University but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of West Virginia University.
Declarations
Ethical approval
This retrospective study was conducted in accordance with the Declaration of Helsinki and approved by the West Virginia Uinversity Institutional Review Board, which waived the requirement for informed consent due to the retrospective nature of the study.
Consent to participate
Not applicable (retrospective study).
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Ostrom QT, Price M, Ryan K, Cioffi G, Kruchko C, Waite K et al (2024) CBTRUS statistical report: primary brain and other central nervous system tumors diagnosed in the United States in 2017–2021. Neuro Oncol 26(Suppl 4):iv1–iv49. 10.1093/neuonc/noae123 [Google Scholar]
- 2.Louis DN, Perry A, Wesseling P, Brat DJ, Cree IA, Figarella-Branger D et al (2021) The 2021 WHO classification of tumors of the central nervous system: a summary. Neuro Oncol 23(8):1231–1251. 10.1093/neuonc/noab106 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ et al (2005) Concomitant and adjuvant temozolomide (TMZ) and radiotherapy (RT) for newly diagnosed glioblastoma: conclusive results of a randomized phase III trial by the Sara Swann Foundation for Neuroscience and the European Organisation for Research and Treatment of Cancer Brain and Radiotherapy Groups. J Clin Oncol 23(16suppl):2. 10.1200/JCO.2005.23.16_suppl.2. (Full trial: Stupp R, Mason, WP, van den Bent MJ et al. (2005) Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. N Engl J Med 352(10):987–996. 10.1056/NEJMoa043330) [DOI] [PubMed]
- 4.Lamba N, Muskens IS, DiRisio AC, Broekman ML, Kruchko C, Waite KA et al (2023) Long-term survival in glioblastoma: a systematic review and meta-analysis of real-world evidence. Neurooncol Adv 5(1):vdad058. 10.1093/noajnl/vdad058 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Barnholtz-Sloan JS, Ostrom QT, Shetty S, Williams VL, Wilson T, Johnson DR et al (2023) Comparative survival of patients with glioblastoma using real-world data from the US National Cancer Database (NCDB) and SEER program. J Neurooncol 162(2):345–354. 10.1007/s11060-023-04289-5 [Google Scholar]
- 6.Brown TJ, Brennan MC, Li M, Church EW, Brandmeir NJ, Rakszawski KL et al (2016) Association of the extent of resection with survival in glioblastoma: a systematic review and meta-analysis. JAMA Oncol 2(11):1460–1469. 10.1001/jamaoncol.2016.1373 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Zahnd WE, Murphy KM, Batai K, Wells K, Stratton MS, Maki CG et al (2022) Rural-urban differences in the receipt of adjuvant radiation therapy for high-grade gliomas: a population-based analysis. J Neurooncol 158(1):101–111. 10.1007/s11060-021-03957-0 [Google Scholar]
- 8.Chino F, Peppercorn J, Taylor E, Hall R, Neal B, Blackstock W et al (2021) Rural-urban differences in travel distance and treatment initiation time for patients with head and neck cancer in North Carolina. J Rural Health 37(3):609–616. 10.1111/jrh.12549 [Google Scholar]
- 9.Singh R, Devlin SM, Inman K, Armstrong A, Kimes S, Reeder-Hayes K et al (2024) Rural-urban disparities in cancer outcomes: a systematic review. JCO Oncol Pract 20(5):678–688. 10.1200/OP.23.0072138320228 [Google Scholar]
- 10.Doogan NJ, Roberts ME, Wewers ME et al (2020) The impact of rural residence on adult brain cancer survival in the United States. J Rural Health 36(4):456–463. 10.1111/jrh.12413 [Google Scholar]
- 11.Shen J, Shriver B, Tykocinski E, Taylor LP, Lin E, Wu X (2021) Rural-urban disparities in high-grade glioma survival: a systematic review and meta-analysis. Neurooncol Adv 3(1):vdab094. 10.1093/noajnl/vdab094 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Butterfield JT, Golzarian R, Smith KA, Venteicher AS, Chen CC, Sherman JH (2025) Rural-urban survival disparities in glioblastoma are eliminated in patients treated at a comprehensive cancer center. J Neurooncol 173(2):285–293. 10.1007/s11060-025-04567-8(Adapted from UVM Larner College of Medicine single-center analysis) [Google Scholar]
- 13.U.S. Census Bureau. American Community Survey 1-year estimates: percent rural population by state (2023) Washington, DC: U.S. Department of Commerce; 2024. Available from: https://www.census.gov/programs-surveys/acs/data.html
- 14.United States Department of Agriculture Economic Research Service. Rural-Urban Commuting Area (RUCA) codes: documentation. Washington, DC: U.S. Department of Agriculture (2023) [cited 2025 Nov 16]. Available from: https://www.ers.usda.gov/data-products/rural-urban-commuting-area-codes
- 15.Sarriera Valentin G, D’Agostino E, Callas P et al (2025) Single-center retrospective analysis of the impact of rurality on glioblastoma care in Vermont. BMC Health Serv Res 25(1):1052. 10.1186/s12913-025-12988-z [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Estevez-Ordonez D, Abdelrashid M, Coffee E et al (2023) County-level disparities in care for patients with glioblastoma. Neurosurg Focus 55(5):E12. 10.3171/2023.8.FOCUS23509 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Johnson DR, O’Neill BP, Morris J et al (2019) Patterns and disparities of care in glioblastoma. Neurooncol Pract 6(1):37–46. 10.1093/nop/npy031 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Zhu P, Du XL, Zhu JJ, Esquenazi Y (2019) Improved survival of glioblastoma patients treated at academic and high-volume facilities: a hospital-based study from the National Cancer Database. J Neurosurg 132(2):491–502. 10.3171/2018.10.JNS182247 [DOI] [PubMed] [Google Scholar]
- 19.Aulakh S, DeDeo MR, Free J, Rosenfeld SS, Quinones-Hinojosa A, Paulus A et al (2019) Survival trends in glioblastoma and association with treating facility volume. J Clin Neurosci 68:271–274. 10.1016/j.jocn.2019.04.028 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Aulakh S, DeDeo MR, Free J et al (2023) Facility patient volume and survival among individuals diagnosed with malignant central nervous system tumors. J Neurooncol 161(3):535–543. 10.1007/s11060-022-04227-w [DOI] [PubMed] [Google Scholar]
- 21.Alimohammadi E, Bagheri SR, Taheri S, Dayani M, Abdi A (2020) The impact of extended adjuvant temozolomide in newly diagnosed glioblastoma multiforme: a meta-analysis and systematic review. Oncol Rev 14(1):461. 10.4081/oncol.2020.461 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Inggas MAM, Patel U, Wijaya JH et al (2025) The role of temozolomide as adjuvant therapy in glioblastoma management: a systematic review and meta-analysis. BMC Cancer 25(1):399. 10.1186/s12885-025-13757-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Jayakrishnan T, Patel A, Ganti AK (2021) Landmark Cancer Clinical Trials and Real-World Patient Populations: Examining Race and Age Reporting. Cancers (Basel) 13(22):5770. 10.3390/cancers13225770 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data that support the findings of this study are available from West Virginia University but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data are however available from the authors upon reasonable request and with permission of West Virginia University.