ESTRO 2021 Abstract Book

S496

ESTRO 2021

In postoperative radiotherapy (RT) for pediatric ependymoma the target is often in close vicinity to sensitive organs at risk (OARs). A balance has to be found between target coverage and sparing of OARs. The aim of this study was to quantify inter-center variations in RT treatment plans of pediatric ependymoma generated following a common guideline. Materials and Methods Nine centers from the NOPHO-RT group participated in this study. Four proton and four photon plans were generated for three pediatric ependymoma cases based on a common dataset (planning CT and structure set), following the RT treatment guideline of the European SIOPE Ependymoma II study (NCT02265770). Case 1 (supratentorial) and 2 (infratentorial) were radically resected tumors with a prescription dose of 59.4 Gy to the clinical target volume (CTV). Case 3 (infratentorial) presented with residual tumor at time of RT. The treatment was planned in three phases – 54 Gy to CTV1, additional 5.4 Gy to CTV2 (CTV1-spinal cord) and a boost of 8 Gy/2 fractions in the residual tumor (Fig. 1). For all plans, relevant dose/volume metrics were compared, neglecting fractionation differences for case 3.

Results For case 1, all plans achieved excellent target coverage with a V95%>100%. Proton plans reduced the irradiated whole brain volume, with a median brain Dmean of 8.9 Gy vs. 16.8 Gy for the photon plans. This advantage was less clear in the infratentorial plans, with a photon vs. proton median brain Dmean of 10.7 Gy vs. 8.7 Gy for case 2; 12.5 Gy vs. 10.5 Gy for case 3. For case 2 and 3, all centers regardless of treatment modality had to make compromises between target coverage and brainstem/spinal cord sparing. The median value of all brainstem metrics were lower in the proton plans, but the range was large for all modalities. The spinal cord C1 was better spared with photons in case 2, with a median D2% of 50.9 Gy vs. 52.8 Gy, while C2/C3 median D2% was lower with protons in case 3. For case 2, the range in target coverage V95% was similar between modalities, i.e. [96.5%-99.9%] (photons), [95.5%-99.9%] (protons). For case 3 the target coverage of CTV1 was similar for both modalities but not satisfactory for some of the proton plans for CTV2/CTVboost. In case 2, Dmean to the pituitary gland was lower with protons whereas the hypothalamus was better spared with photons. In case 3, the contralateral cochleas and hippocampi were better spared in the proton plans. No clear patterns were identified for other OARs (Table 1).