Session Item

Ependymomas are the 3rd cause of children’s intracranial tumours and achieve a 5-years local control rate of 50% with a standard post-operative dose of 59.4 Gy. Proton beam therapy (PBT) is the best alternative technic so far to decrease the dose to organs at risk. Recently, retrospective studies reported higher rates of radionecrosis and imaging modifications of the brainstem after PBT than photon radiation therapy (XRT). This led several centers to decrease the prescribed dose to 54 Gy in PBT for posterior fossa ependymomas under 3 years old. The benefit-risk’s balance of the dose reduction is unclear especially since the real incidence of clinically relevant brainstem’s toxicity appears low. This prompted us to compare the incidence of clinical and imaging brainstem toxicity between XRT and PBT and to correlate the MRI changes to the dosimetric data in the retrospective PEPPI French cohort.

From 2001 to 2021, we studied the MRI brainstem’s modification and clinical brainstem toxicity for patients ≤ 22 years old from 5 reference French centers treated with surgery and radiation therapy (RT) for a localized intracranial ependymoma. Patients with relapse at the first follow-up MRI were excluded. Pre operative, post operative and follow-up MRI during 18 months after RT (or until relapse if it appeared sooner) were analyzed. Post-gadolinium T1, T2, FLAIR and diffusion weighted imaging were used. Survival rates estimated by the Kaplan-Meier method are presented with their 95% confidence intervals (CIs).

Eighty-three patients were included in the study, 42 patients were treated with PBT, 37 with XRT and 4 with both. Median follow-up was 5.6 years [95%CI: 4.8-6.2], The 5-years overall survival and progression free survival rates were respectively 86.1% [95%CI: 75.7-92.3] and 68.7% [95%CI: 56.4-78.3]. First relapse occurred for 60% in the field of treatment. Neither clinical or MRI brainstem radionecrosis nor new or progressive brainstem symptoms were found.  Four patients, all with posterior fossa tumour, presented a punctiform brainstem enhancement (3/4 also had a FLAIR hypersignal), with a median time of onset of 3.5 months [3.0-9.4] after RT and a median time of disappearance of 7.6 months [3.7-7.9] for 3 on 4 patients. Two of them had received PBT, one was treated with XRT and one with mixed XRT-PBT. Median prescribed dose and median brainstem mean dose were respectively 57.6 Gy [54-59.4] and 47.2 Gy [45.3-53.2]. None of the 4 patients suffered from new or progressive brainstem symptoms after completion of RT. For patients with posterior fossa tumour without MRI abnormality, median of brainstem mean dose was 48.5 Gy [17.6-55.3].

Brainstem imaging change occurred in 4,8% of children in a large series including PBT. No radionecrosis or symptomatic MRI change was found within the brainstem. The pursue of the study with the others French centers will strengthen our results.