Outcomes Prognostication for Septuagenarians and Octogenarians Diagnosed with Meningioma: Leveraging National Registry Data for the Development of a Prognostic Model for Survival Prediction in the Elderly Population

Introduction: With recent increases in average life expectancy, elderly patients are now more likely to present with symptomatic meningiomas. Although previous institutional and national database studies have described the increased risk associated with treating elderly, frail patients diagnosed with meningiomas, there is a need for further investigation of individualized prognoses and optimization of patient-specific treatment regimens. This is particularly true for elderly meningioma patients, as those >70 years old (septuagenarians and octogenarians) are more likely to harbor high-grade, progressive meningiomas than younger patients. In this study, we construct a machine learning (ML) calculator designed for online and remote mobile-device use that can predict individualized survival outcomes in elderly patients diagnosed with meningiomas.

Methods: Patients aged 65 years and older diagnosed with histopathologically confirmed meningioma were identified from the American College of Surgeons National Cancer Database (NCDB) and the Surveillance, Epidemiology, and End Results (SEER) database (2000-2019). Univariate and multivariate Cox proportional hazards regression was performed to identify risk factors for patient mortality. Additionally, an ML calculator was constructed for the estimation of 2-, 5-, and 10-year overall survival (OS) and median survival in months. Concordance indices in the form of areas under the curve (AUCs) were utilized to calibrate each survival model for optimal performance. Across all analyses, p = 0.05 was set as the threshold for statistical significance.

Results: In total, 56,485 elderly meningioma patients were identified from the NCDB and SEER databases. Eleven variables, including age, group, sex, race, ethnicity, anatomical location, laterality, histology, WHO grade, surgery status, radiotherapy, and chemotherapy, were used as risk factors to build the calculator based on the data reported by the NCDB and SEER databases (https://spine.shinyapps.io/elderly_meningioma/). The concordance indices of the final models were 0.782, 0.765, and 0.751 for 2-, 5-, and 10-year OS, respectively, suggesting that good calibration was achieved for each.

Conclusion: We present a prognostic ML calculator featuring three models capable of satisfactorily predicting accurate, individualized 2-, 5-, and 10-year OS for septuagenarians and octogenarians (elderly patients) diagnosed with meningiomas. While multi-institutional cohort studies are warranted as the next step in validating the present analysis, the ML calculator developed and reported herein holds promise for aiding clinicians in making well-informed decisions when treating elderly patients diagnosed with meningiomas.

Publikationsverlauf

Artikel online veröffentlicht:
07. Februar 2025