The concept, intention, and evaluation of the term treatment-refractory meningioma

Abstract

Background: Treatment-refractory meningioma is a widely used term but lacks standardized criteria, impairing research comparability and treatment evaluation. The aim of this study was to assess the heterogeneity of patient populations labeled as treatment-refractory and to explore recommendations for better consistency.

Methods: We systematically reviewed 69 studies published before 2024 and analyzed individual participant data from 15 cohorts (n = 211) that included treatment-refractory patients who underwent experimental therapy with somatostatin receptor (SSTR)-targeted therapies. A reference cohort (n = 102) with newly diagnosed WHO-3 meningiomas was used descriptively for comparison. Progression and death were the primary endpoints. Hazard rate ratios were estimated via Poisson regression, and inter-study heterogeneity was quantified using I² statistics.

Results: Definitions of treatment-refractory varied substantially across previous studies. WHO-1 patients showed high statistical inter-study variability, particularly for the long-acting SSTR-analogues group when assessing progression (I² = 81.7%) and death (I² =74.8%). Patients with treatment-refractory WHO-2 and WHO-3 meningioma exhibited more consistency across endpoints and SSTR-targeted therapies (I² percentages ≤ 16.0%). Risk of progression and death differed significantly by WHO grade, regardless of SSTR-targeted therapy.

Conclusions: Our findings demonstrate an inconsistent use of the term treatment-refractory and substantial variability of effect estimates dependeing on the individual cohorts. Pooling patients across WHO grades is unfeasible for assessing treatment effects. Based on the present study and prior evidence, we outline recommendations to improve consistency in future trial design and enable more meaningful comparisons across studies. The recommendations are grouped into four categories: radiographic evaluation, endpoints, clinical core elements, and molecular classification.

Keywords: Individual participant data; Progressive meningioma; Recommendations; Retrospective cohort; Treatment-refractory.

Fig. 1

Flow chart of the study cohort, which was divided into two treatment groups (Group A and Group B) and the Reference group.

Fig. 2

The risk of progression presented as hazard rate ratios for each of the included studies when compared with the Reference group. Patients stratified for type of treatment and WHO-grade. A fixed-effects model on log(hazard ratios) and corresponding standard errors were undertaken to render I²-percentages. Each square reflects the weight given to the individual study (inverse-variance). Variance is derived from the square standard error of the log(hazard ratio). Inverse-variance reflect 1/variance, i.e., smaller weight is given to studies with a larger standard error, typically directly related to fewer participants. *Cardona et al. and Graillon et al. combined SSA with Everolimus [20, 23]. A: Peptide receptor radionuclide therapy. B: Long-acting somatostatin analogues

Fig. 3

The risk of death presented as hazard rate ratios for each of the included studies when compared with the Reference group. Patients stratified for type of treatment and WHO-grade. A fixed-effects model on log(hazard ratios) and corresponding standard errors were undertaken to render I²-percentages. Each square reflects the weight given to the individual study (inverse-variance). Variance is derived from the square standard error of the log(hazard ratio). Inverse-variance reflect 1/variance, i.e., smaller weight is given to studies with a larger standard error, typically directly related to fewer participants. *Cardona et al. and Graillon et al. combined SSA with Everolimus [20, 23]

Fig. 4

A: Aalen-Johansen estimates - risk of progression. B: Kaplan-Meier estimates – overall survival probability. *PRRT: peptide receptor radionuclide therapy; *SSA: long-acting somatostatin analogues