Background: Clinical outcomes in isocitrate dehydrogenase (IDH)-wildtype grade 4 diffuse glioma remain highly heterogeneous despite standard multimodal therapy. We evaluated the prognostic significance of pretreatment platelet-to-albumin ratio (PAR) and compared its performance with established inflammatory biomarkers. Methods: This retrospective cohort study included 166 patients with histopathologically confirmed IDH-wildtype grade 4 diffuse glioma treated between 2017 and 2025. Pretreatment PAR, neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), pan-immune inflammation value (PIV), C-reactive protein-to-albumin ratio (CAR), and lactate dehydrogenase-to-albumin ratio (LAR) were evaluated. Discriminative performance was assessed using classical and inverse probability of censoring weighting (IPCW)-adjusted time-dependent receiver operating characteristic (ROC) analyses. Survival outcomes were evaluated using Kaplan–Meier analyses and predefined baseline-adjusted multivariable Cox proportional hazards regression models. Internal validation was performed using 1000 bootstrap resampling iterations. Results: PAR demonstrated the highest discriminative performance for 12-month overall survival (OS), with an area under the curve of 0.853. Using an optimal cutoff value of 79.459, patients with elevated PAR experienced significantly shorter OS (median, 6.2 vs. 16.7 months; p < 0.001) and progression-free survival (PFS) (median, 5.9 vs. 12.3 months; p < 0.001). In baseline-adjusted multivariable analyses, elevated pretreatment PAR remained independently associated with inferior OS (hazard ratio [HR], 3.287; 95% confidence interval [CI], 2.196–4.921; p < 0.001) and PFS (HR, 3.791; 95% CI, 2.501–5.749; p < 0.001). These findings were supported by sensitivity analyses and bootstrap internal validation. Conclusions: Pretreatment PAR was independently associated with survival outcomes and demonstrated favorable discriminative performance relative to other inflammatory biomarkers. PAR may represent an accessible biomarker reflecting tumor–host interactions in IDH-wildtype grade 4 diffuse glioma. The proposed PAR cutoff should be considered exploratory and requires external validation before routine clinical application.

Background/Objectives: Diffuse gliomas can reach grade 4 disease through markedly different temporal trajectories, yet prognosis after transformation is generally assessed from the characteristics of the transformed tumor itself. Whether the duration of the preceding disease course retains prognostic information after grade 4 transformation remains incompletely understood. Methods: We conducted a retrospective cohort study of patients with an antecedent WHO grade 2 or 3 diffuse glioma who subsequently underwent histopathologically confirmed grade 4 transformation. Transformation interval was defined as the time between the antecedent diagnosis and grade 4 transformation and was modelled continuously per 12-month increments. The primary endpoint was overall survival (OS) from transformation; progression-free survival (PFS) was secondary. Multivariable Cox models were adjusted for age, Eastern Cooperative Oncology Group performance status, antecedent grade and IDH status. Functional form, molecular context, landmark robustness and incremental prognostic information were additionally examined. Results: Fifty-one patients were included; 33 (64.7%) had antecedent grade 2 and 18 (35.3%) grade 3 disease. Median transformation interval was 51.6 months. Median post-transformation OS and PFS were 20.9 and 14.1 months, respectively. Each additional 12 months before grade 4 transformation was associated with lower mortality in univariable analysis (hazard ratio [HR], 0.845; 95% CI, 0.755–0.946; p = 0.003) and remained independently associated with OS after multivariable adjustment (HR, 0.858; 95% CI, 0.748–0.983; p = 0.028). No evidence of nonlinearity was observed (p = 0.407). The corresponding association with PFS was weaker and not statistically significant (adjusted HR, 0.943; 95% CI, 0.840–1.060; p = 0.327). Exploratory interaction analysis indicated that the inverse transformation interval–OS association was concentrated in IDH-mutant tumours (adjusted HR per 12 months, 0.758; 95% CI, 0.630–0.912), whereas no significant inverse association was demonstrated in IDH-wildtype tumours (adjusted HR, 1.225; 95% CI, 0.923–1.626; interaction p = 0.007). Adding transformation interval improved overall model fit but did not materially improve optimism-corrected discrimination. Conclusions: In this retrospective, single-centre cohort, a longer transformation interval was associated with improved post-transformation OS after adjustment for selected clinical and molecular factors, whereas the association with PFS remained inconclusive. The exploratory IDH interaction and the absence of meaningful improvement in optimism-corrected discrimination warrant cautious interpretation. These findings are hypothesis-generating and require validation in larger, molecularly defined longitudinal cohorts before clinical prognostic utility can be established.

Background and Objectives: Clinical outcomes remain markedly heterogeneous in glioblastoma, isocitrate dehydrogenase (IDH)-wildtype, despite contemporary molecular classification and established clinicopathological risk factors. We investigated whether the pretreatment lactate dehydrogenase-to-albumin ratio (LAR), an accessible circulating index integrating systemic disease- and host-related information, provides complementary prognostic information. Materials and Methods: We retrospectively evaluated 134 patients with glioblastoma, IDH-wildtype, central nervous system (CNS) World Health Organization (WHO) grade 4. An exploratory, cohort-specific LAR threshold was derived for 12-month overall survival (OS) and retained without re-optimization across subsequent analyses. Associations with OS and progression-free survival (PFS) were evaluated using Cox regression, and incremental prognostic performance was assessed against a baseline clinicopathological model. Temporal and continuous associations with survival, internal model validation, and the relationship between LAR and treatment response during adjuvant temozolomide therapy were also examined. Results: The cohort-derived LAR threshold was 5.718. High pretreatment LAR was associated with shorter median OS (5.5 vs. 16.0 months) and PFS (5.2 vs. 10.6 months; both p < 0.001) and remained independently associated with OS (adjusted hazard ratio [HR], 4.51; 95% CI, 2.84–7.18) and PFS (adjusted HR, 3.60; 95% CI, 2.33–5.56). Adding LAR to the clinicopathological model increased Harrell’s C-index from 0.699 to 0.760 for OS and from 0.690 to 0.759 for PFS (both likelihood ratios, p < 0.001), with optimism-corrected C-indices of 0.743 and 0.742, respectively. Prognostic discrimination and adjusted effect estimates were greatest early in the follow-up period and attenuated over time, whereas continuous LAR showed nonlinear associations with both endpoints. Among the 104 patients who received adjuvant temozolomide, a high LAR was independently associated with progressive disease (adjusted odds ratio, 3.72; 95% CI, 1.28–10.81). Conclusions: Pretreatment LAR identified patients with IDH-wildtype glioblastoma who experienced poorer survival, and the prognostic effect was concentrated predominantly in the early disease course. Its incremental contribution beyond conventional clinicopathological characteristics supports the further evaluation of LAR as an accessible component of baseline risk characterization. However, independent external validation is required before clinical application.