The Shifting Prognostic Value of Performance Status in Aging Glioblastoma Patients: A Retrospective Cohort Study

Introduction

Glioblastoma (GBM) is the most common and aggressive primary malignant brain tumor in adults, with an incidence that rises sharply with age.¹ Patients aged 65 and older account for a substantial proportion of new diagnoses, yet their prognosis remains exceptionally poor, with a median overall survival (OS) of less than one year and a 5-year survival rate of only 5.3%.² The management of this vulnerable population is complicated by factors such as increased comorbidities, diminished physiological reserve, and heightened risk of treatment-related toxicity, creating a pressing need for refined prognostic tools to guide individualized therapeutic strategies.^3,4 Recent studies continue to emphasize that individualized treatment, guided by both clinical and molecular parameters, is essential for optimizing outcomes in this heterogeneous patient group.^5,6 While the 2021 WHO classification integrates molecular markers like IDH mutation and MGMT promoter methylation as central prognostic factors, clinical predictors remain indispensable, particularly when molecular testing is unavailable or inconclusive in real-world settings.

A patient’s functional status, commonly quantified by the Karnofsky Performance Status (KPS) score, is a cornerstone in clinical decision-making for GBM. It is widely used to determine eligibility for aggressive treatments, including extensive surgery and standard chemoradiation (the Stupp protocol).⁷ However, the prognostic utility of preoperative KPS in the elderly remains a subject of debate. While some studies support its predictive value,⁷ others have found no significant association between preoperative KPS and survival outcomes.^8–10 This inconsistency suggests that the impact of performance status may be modified by other factors, such as chronological age itself.

The distinction between the “young-elderly” (eg, 60–70 years) and the “older-elderly” (eg, >70 years) is increasingly recognized in oncology, as treatment tolerance and outcomes can differ significantly.^6,11 Geriatric oncology guidelines suggest that physiological decline accelerates non-linearly, necessitating finer age stratification.^12,13 It remains unclear whether preoperative Karnofsky Performance Status (KPS) retains its prognostic power across this entire spectrum, particularly given that age and functional status interact in complex ways and may not be exhaustive predictors on their own.⁶ Furthermore, the significance of postoperative KPS, reflecting the immediate impact of surgical intervention, may offer distinct prognostic information, as studies suggest that postoperative KPS ≥ 80 is a stronger predictor of survival than preoperative KPS in elderly glioblastoma patients.¹⁴ Therefore, this study was designed to dissect the prognostic role of both preoperative and postoperative KPS within specific age strata of elderly GBM patients, aiming to provide a more nuanced understanding to guide surgical and adjuvant therapy decisions in this challenging demographic.

Materials and Methods

Study Design and Patient Population

This single-center, retrospective cohort study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.¹⁵ We retrospectively identified all consecutive patients aged 60 years or older with newly diagnosed, histopathologically confirmed primary GBM (WHO Grade IV) at the Fourth Hospital of Hebei Medical University between January 1, 2017, and December 31, 2021. Patients with recurrent GBM or those who did not undergo any surgical procedure (biopsy or resection) were excluded. A final cohort of 89 eligible patients was included for analysis. A post-hoc power calculation indicated that with a sample size of 89 and the observed effect sizes, the study had >80% power to detect significant hazard ratios for the primary variable of postoperative KPS (alpha=0.05).

Ethical Considerations

The study protocol received full approval from the Institutional Committee on Human Research of the Fourth Hospital of Hebei Medical University (No. 2024KS147). The committee granted a waiver of the requirement for individual patient informed consent due to the retrospective nature of the research, the use of de-identified data, and the confirmation that the study posed no more than minimal risk to subjects. The study was conducted in accordance with the Declaration of Helsinki.

Data Collection and Variables

A comprehensive dataset was extracted from electronic medical records. Demographic data included age at diagnosis and gender. Clinical variables included tumor characteristics (laterality, location, number of lesions), presence of systemic comorbidities, and preoperative KPS score. Preoperative KPS was assessed by a consensus of two senior neurosurgeons upon admission to minimize inter-observer variability. Treatment-related variables included the extent of resection (EOR), postoperative KPS score, and details of adjuvant therapies. Postoperative KPS was assessed prior to discharge (median postoperative day 12, range 7–21) to reflect the immediate functional outcome of surgery. EOR was determined based on postoperative MRI scans within 72 hours of surgery and categorized as gross total resection (GTR, >95% resection), subtotal resection (STR, <95% resection), or biopsy only. In the absence of routine volumetric software, EOR was determined via comparison of pre- and post-operative contrast-enhanced T1-weighted images by a neuroradiologist and the operating surgeon. Adjuvant treatments were categorized as: early chemotherapy, concurrent chemoradiation (CCRT), and adjuvant chemotherapy. Standard CCRT followed the Stupp protocol: Radiotherapy consisted of a total dose of 60 Gy delivered in 30 fractions over 6 weeks. Concurrent chemotherapy consisted of temozolomide (75 mg/m² daily). All patients were followed until death or the study census date of June 30, 2022.

Statistical Analysis

Statistical analyses were performed using IBM SPSS Statistics, Version 23.0 (IBM Corp., Armonk, NY, USA). Descriptive statistics were used to summarize patient characteristics. Continuous variables were presented as means with ranges, and categorical variables as frequencies and percentages. The primary survival analysis was conducted using the Kaplan-Meier method, and differences between survival curves were assessed using the Log rank test. To enhance visual interpretation, number-at-risk tables were generated and displayed below the survival curves. To evaluate the age-dependent effect of KPS, patients were stratified into three groups: 60–64 years, 65–69 years, and ≥70 years. This stratification was chosen a priori to distinguish between “young-elderly,” “intermediate-elderly,” and “older-elderly” subgroups, consistent with geriatric oncology frameworks that identify age 65 and 70 as critical inflection points for frailty.¹⁶ Univariable Cox proportional hazards regression was performed to identify potential prognostic factors for OS. Covariates with a P-value <0.20 in the univariable analysis were subsequently included in a multivariable Cox proportional hazards model using a forward stepwise method to identify independent prognostic factors. Hazard ratios (HR) and their 95% confidence intervals (CI) were calculated. A two-sided P-value <0.05 was considered statistically significant for all analyses.

Results

Patient Demographics and Treatment Characteristics

A total of 89 patients met the inclusion criteria. The cohort had a nearly balanced gender distribution (49.4% male) and a mean age at diagnosis of 66.4 years. The cohort was stratified into three age groups: 35 patients (39.3%) were 60–64 years old, 33 (37.1%) were 65–69, and 21 (23.6%) were 70 years or older. All patients underwent a surgical procedure: GTR was achieved in 64 patients (71.9%), STR in 18 (20.2%), and biopsy in 7 (7.9%). Following surgery, 74 patients (83.1%) received CCRT, and 68 (76.4%) proceeded to adjuvant chemotherapy. Molecular marker status (IDH mutation, MGMT methylation) was unavailable for the majority of patients (87.6%) due to the historical nature of the cohort (Supplementary Table S1). The detailed clinicopathological characteristics are presented in Table 1.

Table 1 Patient and Treatment Characteristics (N=89)

Overall Survival Outcomes

As of the final follow-up, the mean OS for the entire cohort was 14.8 months. A trend of decreasing survival with increasing age was observed: the mean OS was 17.2 months for the 60–64 age group, 14.3 months for the 65–69 group, and 9.8 months for the ≥70 group. However, this trend did not reach statistical significance in regression analysis (Table 2).

Table 2 Univariable Cox Regression Analysis for Overall Survival

Prognostic Factor Analysis for Overall Survival

Univariable Cox regression analysis identified several factors significantly associated with OS (Table 2). Male gender (HR=2.223, P=0.002) and bilateral tumor growth (HR=4.003, P=0.001) were associated with worse outcomes. In contrast, GTR (vs biopsy, HR=0.210, P=0.001), higher postoperative KPS (HR=0.984, P=0.008), receiving CCRT (HR=0.199, P<0.001), and receiving adjuvant chemotherapy (HR=0.348, P<0.001) were protective. Notably, preoperative KPS as a continuous variable was not significantly associated with OS in the overall cohort (P=0.871).

In the subsequent multivariable analysis (Table 3), four factors remained independent predictors of OS. Male gender was confirmed as an independent risk factor for mortality (HR=3.787, 95% CI: 2.111–6.794, P<0.001). GTR (vs biopsy, HR=0.187, 95% CI: 0.074–0.470, P<0.001), higher postoperative KPS score (HR=0.981, 95% CI: 0.967–0.994, P=0.006), and receipt of CCRT (HR=0.200, 95% CI: 0.099–0.404, P<0.001) were independently associated with prolonged survival.

Table 3 Multivariable Cox Regression Analysis for Overall Survival

Age-Stratified Prognostic Value of KPS

Given the non-significant finding for preoperative KPS in the overall cohort, we performed a planned age-stratified survival analysis to explore its nuanced role (patient distribution across subgroups is detailed in Supplementary Table S2). This analysis revealed a striking divergence. In the “young-elderly” group (60–64 years), preoperative KPS was a strong and significant predictor of OS, with higher scores correlating with better survival (log-rank P=0.003; Table 4 and Figure 1A). However, this association was lost in patients aged 65–69 years (log-rank P=0.798; Figure 1B) and those aged ≥70 years (log-rank P=0.354; Figure 1C).

Table 4 Log Rank Test for Equality of Survival Distributions by Preoperative KPS

Figure 1 Kaplan-Meier analysis of overall survival stratified by preoperative Karnofsky Performance Status (KPS). Number-at-risk tables are provided below the X-axis. Note: 95% confidence bands were omitted to maintain visual clarity given the multiple KPS strata. (A) Significant separation in 60–64 group. (B and C) No significant separation in older groups.

In stark contrast, postoperative KPS demonstrated robust and consistent prognostic power across all age strata (Table 5). A higher postoperative KPS was significantly associated with superior OS in the 60–64 age group (log-rank P<0.001; Figure 2A), the 65–69 age group (log-rank P=0.001; Figure 2B), and the ≥70 age group (log-rank P<0.001; Figure 2C).

Table 5 Log Rank Test for Equality of Survival Distributions by Postoperative KPS

Figure 2 Kaplan-Meier analysis of overall survival stratified by postoperative KPS. Higher postoperative KPS significantly predicted survival across all age groups. Number-at-risk tables are provided below the X-axis. (A) 60–64 years; (B) 65–69 years; (C) ≥70 years.

Discussion

The optimal management of elderly patients with GBM remains a significant clinical challenge, requiring a delicate balance between treatment efficacy and potential toxicity. Our study provides important insights into prognostic stratification in this population, confirming several established predictors while uncovering a novel, age-dependent role for performance status. The primary findings are threefold: (1) male gender is an independent negative prognostic factor, whereas maximal safe resection and CCRT are robust positive prognostic factors; (2) preoperative KPS is a significant predictor of survival only in the “young-elderly” (60–64 years), losing its prognostic utility in patients aged 65 and older; and (3) postoperative KPS is a powerful and consistent prognostic indicator across the entire elderly spectrum.

Our confirmation of EOR and CCRT as key determinants of survival aligns with the current standards of care and extensive existing literature.^4,17,18 These findings reinforce the principle that aggressive, multi-modal therapy should be offered to appropriately selected elderly patients, as it confers a substantial survival benefit. The unexpected finding of male gender as a strong independent risk factor warrants further investigation. The finding of male gender as a risk factor should be interpreted with caution. While consistent with some literature,² it may be influenced by unmeasured confounding variables such as molecular subtypes (eg, MGMT methylation distribution), which were not available for this analysis. This finding also aligns with recent epidemiological data suggesting a survival advantage for females in GBM.^19–22 Potential mechanisms include the neuroprotective effects of estrogen, which may inhibit glioma cell proliferation,^23–25 and sex-specific differences in immune regulation within the tumor microenvironment.^26,27

The most compelling finding of our study is the differential prognostic impact of preoperative KPS based on age. The observation that preoperative KPS is prognostic for patients aged 60–64 but not for those ≥65 provides a potential explanation for the conflicting reports in the literature, which often group all elderly patients together.^9,28 We hypothesize that in the “young-elderly,” preoperative KPS accurately reflects underlying functional reserve. However, in older patients (≥65), the prognostic signal of KPS may become obscured by the cumulative burden of age-related physiological decline and comorbidities, which themselves become dominant drivers of outcome. This finding has direct clinical implications: for patients over 65, a poor preoperative KPS alone should perhaps not be an absolute contraindication to considering aggressive surgery. Instead, surgical candidacy should be based on the potential for functional preservation rather than preoperative status alone, incorporating frailty indices and comprehensive geriatric assessments.^29,30

In contrast to its preoperative counterpart, postoperative KPS emerged as a universally powerful prognostic marker. This underscores that a patient’s functional status following surgical recovery is a critical determinant of their ability to tolerate and benefit from subsequent adjuvant therapies, ultimately influencing their survival trajectory.^31,32 Therefore, the surgical goal in the elderly must balance cytoreduction with the preservation of eloquent function. This emphasizes the importance of advanced surgical techniques, such as awake craniotomy, intraoperative neuromonitoring, and brain mapping, alongside meticulous perioperative care and early integration of neuro-rehabilitation services.³³ The RANO resect group recently validated the pivotal role of postoperative KPS, alongside residual tumor and molecular markers, in a novel risk model, further cementing its status as a cornerstone of postoperative prognostication.³⁴

Limitations

This study has several important limitations that must be acknowledged. First, its retrospective, single-center design is susceptible to selection bias and limits the generalizability of our findings. Second, the modest sample size of 89 patients, when stratified into age and KPS subgroups, results in small numbers within certain strata, which can reduce statistical power and increase the risk of spurious findings. While our results for the KPS subgroup analysis were highly significant, they require validation in a larger, multicenter cohort. Third, and most significantly, our analysis lacks data on IDH1/2 mutation and MGMT promoter methylation status for the majority of patients (Supplementary Table S1), as these tests were not standard of care in our region during the early study period. Consequently, our results should be interpreted as reflecting the prognostic landscape based on clinical factors alone.

Conclusions

In conclusion, this study reaffirms the central roles of maximal safe resection and concurrent chemoradiation in prolonging survival for elderly patients with GBM. Our key contribution is the demonstration of an age-dependent divergence in the prognostic utility of KPS. While preoperative KPS appears to be a reliable prognosticator for patients younger than 65 years, its value diminishes thereafter. Postoperative KPS, however, stands out as a robust and universally applicable prognostic indicator across all elderly age groups. These findings suggest that surgical decision-making in the older-elderly should not be based solely on preoperative presentation but rather on the likelihood of achieving a favorable postoperative functional state.