Work

Independent external validation of the foramen magnum meningioma risk score: limited discrimination and overestimation of absolute risk

Year: 2026

Type: article

Abstract: Objective: Foramen magnum meningiomas (FMMs) cause substantial operative morbidity because of their close relationship to the brainstem, lower cranial nerves, vertebral artery (VA), and upper cervical spinal cord. The Foramen Magnum Meningioma Risk Score (FMMRS) is the only published preoperative, MRI-based model for estimating postoperative complication risk. However, it has not been independently validated, and its calibration is unknown. We performed independent external validation, focusing on absolute risk transportability. Methods: from preoperative MRI by a neurosurgeon blinded to outcomes, with independent re-scoring by a second neurosurgeon. The primary outcome was any postoperative complications; a major complication was a prespecified secondary outcome. Discrimination was assessed using the area under the receiver operating characteristic curve (AUC) and calibrated the primary outcome using the calibration slope and intercept, the Brier score, and observed-vs-expected rates across the published risk groups. Results: Complications occurred in 14/40 patients (35.0%) with major complications in 7/40 (17.5%). Point estimates suggested limited discrimination for both endpoints (AUCs of 0.64 [95% CI 0.46-0.79] and 0.67 [95% CI 0.48-0.84]), with confidence intervals that did not exclude chance. Observed complication rates were lower than predicted in the intermediate- and high-risk groups, indicating an overestimation of absolute risk. The calibration slope was 0.23 (95% CI - 0.04 to 0.57), the intercept -0.90 (95% CI - 1.45 to -0.57), and the Brier score (0.333) exceeded that of the null model (0.228). The largest discrepancy was in the high-risk group, where the expected rate was 96.4% vs. an observed rate of 42.9% (95% CI 21.4-67.4%). In exploratory threshold analysis, FMMRS ≥6 captured all major complications, whereas the published high-risk threshold of ≥8 missed approximately half. Conclusion: In this independent external cohort, the FMMRS retained only limited value as an ordinal stratifier for complication risk and overestimated absolute complication probabilities, particularly for high-risk patients. Its published probabilities should not guide individualized preoperative counseling without recalibration; larger multicenter studies with patient-level data are needed to refine and validate a clinically reliable preoperative prediction model.

Source: Frontiers in Neurology

DOI: 10.3389/fneur.2026.1926742