Abstract: The nomogram effectively predicts GJT recurrence, validated internally and externally, aiding clinical risk stratification.
Introduction
Glomus jugulare tumors (GJT), or tympanojugular paragangliomas, are rare, highly vascularized lesions located at the skull base. Despite their typically slow growth, they exhibit local aggression, often involving critical structures such as the internal carotid artery (ICA) and various cranial nerves. Surgical resection remains the primary treatment modality; however, postoperative recurrence is a significant clinical challenge. Current guidelines, including those from the NCCN, lack specific protocols for GJT management, particularly regarding the role of adjuvant radiotherapy (RT). This study addresses the need for a reliable, data-driven tool to assist clinicians in individualized risk assessment and treatment planning.
Methodology and Cohort Characteristics
Researchers conducted a retrospective analysis of 318 patients treated at a single tertiary medical center over a 20-year period (2003–2022). The cohort was stratified into a training set (70%), an internal validation set (30%), and a separate temporal test set of 41 recent patients (2021–2022) to evaluate the model's performance on a fresh cohort.
The predictive model was developed using univariate analysis followed by multivariate logistic regression to isolate independent predictors of recurrence. A bidirectional stepwise regression approach ensured the final nomogram was parsimonious, incorporating only the most statistically impactful variables.
Study Cohort and Anatomical Analysis
The dataset included comprehensive demographics, radiological findings, and intraoperative parameters. A critical component of the analysis involved mapping tumor invasion sites, as anatomical involvement often dictates both surgical difficulty and recurrence risk.
Key Insights and Predictive Factors
The study identified several critical clinical and biological markers that significantly influence recurrence probability:
- Biological Markers: Negative S-100 expression (indicating a loss of sustentacular supporting cells) and a higher Ki-67 proliferation index were strong predictors of aggressive tumor behavior.
- Demographics: Younger patient age was correlated with higher recurrence rates, suggesting more aggressive biological phenotypes in earlier-onset cases.
- Anatomical Red Flags: Invasion of the internal carotid artery (ICA) and the facial nerve were identified as independent predictors of recurrence, serving as markers for the tumor's invasive potential.
Results and Model Performance
The resulting nomogram demonstrated robust predictive accuracy across all cohorts. The model achieved an Area Under the Curve (AUC) of 0.863 in the training set, 0.711 in the validation set, and 0.784 in the temporal test set. Calibration plots confirmed a high degree of consistency between the predicted probability of recurrence and actual clinical outcomes. The overall recurrence rate observed in this study was 21.4%, with a mean follow-up of 8.7 years.
Clinical Implications and Future Directions
This nomogram provides a practical tool for clinicians to identify high-risk patients who may benefit from adjuvant radiotherapy or more intensive surveillance. While the study is limited by its retrospective, single-center design and the absence of certain genetic markers like SDHB mutations, it establishes a foundation for future multi-center prospective validation. Incorporating genetic sequencing data in future iterations could further refine the model's risk stratification capabilities.
Summary updated