Customized Prognostic Assessment for Newly Diagnosed Glioblastoma Using Statistical Decision Tree
Overview
This study developed a prognostication tool for newly diagnosed glioblastoma (GBM) patients using clinical, radiological, and molecular factors to estimate individual survival probabilities. It analyzed a retrospective cohort from 2010 to 2021, excluding IDH-mutant tumors, and incorporated key prognostic variables such as age, tumor resection extent, MGMT promoter methylation, and tumor characteristics.
Background
Glioblastoma is the most aggressive glioma with limited overall survival despite treatment. Prognostic factors like younger age, better clinical status, tumor location, extent of resection, and MGMT promoter methylation influence survival outcomes. Although molecular classification has advanced, no comprehensive prognostic system integrating clinical, radiological, and molecular data has been established. This study aims to fill this gap by creating a personalized prognostic tool based on objective factors available early after diagnosis.
Data Highlights
The study included patients diagnosed with GBM from 2010 to 2021, excluding those with IDH mutations. Tumor resection was classified as gross total resection (≥95% removal) or subtotal resection, assessed by MRI 72 hours post-surgery. Clinical parameters such as age, Karnofsky performance status, and comorbidities (Charlson Comorbidity Index) were evaluated. Radiological tumor characteristics included tumor size, multifocality, and ventricular contact. Molecular markers (MGMT methylation, p53, Ki67) were available from 2016 onward. Adjuvant treatments mostly followed the Stupp protocol, with some patients enrolled in clinical trials.
Key Findings
Age, clinical status, tumor location, extent of resection, and MGMT promoter methylation are significant prognostic factors for GBM survival.
Gross total resection (≥95%) is associated with improved overall survival compared to subtotal resection or biopsy.
Tumor multifocality and ventricular involvement negatively impact prognosis.
The Charlson Comorbidity Index and Karnofsky performance status at diagnosis provide additional prognostic information.
Molecular markers, particularly MGMT promoter methylation, enhance individualized survival prediction when combined with clinical and radiological data.
The developed statistical decision tree approach enables personalized estimation of survival probability early after GBM diagnosis.
Clinical Implications
This prognostic tool can assist clinicians in tailoring treatment strategies and counseling patients by providing individualized survival estimates based on a combination of clinical, radiological, and molecular factors. Early identification of high-risk patients may guide more aggressive or experimental therapies, while better prognosis patients might benefit from standard treatment protocols. Incorporation of this model into clinical practice could improve decision-making and patient communication.
Conclusion
The study successfully developed a comprehensive prognostic model for newly diagnosed GBM patients integrating key clinical, radiological, and molecular factors. This personalized approach facilitates more accurate survival predictions and supports individualized patient management.
References
Stupp et al. 2005 -- Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma
WHO Classification of Tumors of the Central Nervous System 2021 -- Molecular diagnostics in glioblastoma
Charlson et al. 1987 -- A new method of classifying prognostic comorbidity in longitudinal studies
