Reirradiation as part of a salvage treatment approach for progressive non-pontine pediatric high-grade gliomas: preliminary experiences from the German HIT-HGG study group

Abstract

Doc number: 177

Abstract

Background and purpose: The aim of the present analysis was to assess the feasibility, toxicity, and the tumor control of reirradiation as a salvage treatment for progressive pediatric non-pontine high-grade gliomas (HGG).

Patients and methods: The database of the Reference Center for Radiation Oncology of the German HIT (HIT = German acronym for brain tumor) treatment network for childhood brain tumors was screened for children who were reirradiated for progressive non-pontine HGG.

Results: We identified eight patients (WHO grade III: n = 5; WHO grade IV: n = 3) who underwent reirradiation between April 2006 and July 2012. Median age was 13.5 years at primary diagnosis and 14.8 years at first progression. All patients initially underwent surgery (incomplete resection, n = 7; biopsy, n = 1) followed by radiochemotherapy. Relapses occurred inside (n = 2), at the margin (n = 4), and outside of the preirradiated area (n = 2). In all patients, reirradiation was tolerated well without significant acute toxicity. Temporary clinical improvement and tumor regression on magnetic resonance imaging (MRI) following reirradiation was reported (n = 3). However, all patients finally died by disease progression. Median survival time was 26.2 months from initial diagnosis and 11.4 months after first progression. Median time interval between initial radiotherapy and first reirradiation was 9.0 months. In six patients, all macroscopic tumor deposits were reirradiated. In these patients, median progression-free (overall) survival from the start of reirradiation was 2.4 (4.6) months.

Conclusion: Our analysis, although based on a limited patient number, suggests that reirradiation of progressive non-pontine HGG is feasible in children. Benefit in terms of quality of life and/or survival needs to be assessed in a prospective and ideally in a randomized manner.