Tumour and Dosimetric Parameters Associated with Radionecrosis in Ependymoma Treated with Proton Beam Therapy

Abstract

Objectives

Proton beam therapy (PBT) potentially reduces treatment-related neurocognitive decline. This study aimed to assess factors associated with neurocognitive changes in ependymoma patients treated with PBT.

Methods

All cranial ependymomas treated with PBT from 01/2019 with baseline and 2-year neurocognitive assessment before 02/2024 were included. Neurocognitive outcomes were correlated with patient, tumour and treatment characteristics using Mann-Whitney U and Kruskal-Wallis tests for categorical variables and Spearman Correlation test for continuous variables.

Results

Twenty-one patients were included. Median age at PBT was 7 years (3-19). Prescription dose was 59.4Gy(RBE) in 33 fractions. Median pre-operative gross tumour volume (GTVpre-op) was 48.74cm3 (7.5-306.6) and clinical target volume (CTV) 54.2cm3 (11.6-232.8).

Processing speed (Symbol search) at 2-years was significantly better for posterior fossa (PF) (n=10) than supratentorial tumours (ST) (p=0.018) while PF location was associated with greater change (improvement) in motor skills over 2 years (p=0.033). There was a significant correlation between tumour laterality and change in verbal comprehension index (VCI) over 2 years (p=0.028), with all left-sided tumours (n=6) experiencing decrease, all right-sided tumours (n=5) experiencing increase and midline tumours (n=10) tending to show little change.

Hydrocephalus at diagnosis (n=8) (p=0.042) and systemic anticancer therapy (SACT) (n=15) (p=0.030) were correlated with lower perceptual fluid reasoning (PFR) at 2-year assessment.

At baseline, GTVpre-op was correlated with processing speeds (Coding) (PS) (p=0.0002), VCI (p=0.0145), Full-Scale IQ (FSIQ) (p=0.042). At 2-year assessment, GTVpre-op was correlated with VCI (p=0.0054), FSIQ (p=0.017), and decline from baseline in PFR (p=0.0083), PS (p=0.045), and motor skills (p=0.043), while CTV was correlated with PS (p=0.040), VCI (0.025), FSIQ (p=0.013) and PFR decline from baseline (p=0.010).

Conclusion

This study suggests GTVpre-op, tumour location, laterality, hydrocephalus and SACT impact neurocognitive outcomes in ependymoma patients receiving PBT. Further studies are warranted to assess effect of dosimetric parameters on neurocognitive outcomes.