ESTRO 2024 - Abstract Book

S2034

Clinical - Paediatric

ESTRO 2024

(DFS), overall survival (OS), and median follow-up were retrieved. The 5-year EFS, PFS, RFS and DFS were pooled and are henceforth referred to as PFS. Random effects models were used to pool the data and used to generate forest plots of the 5-year PFS and OS. Association with PFS and OS and radiation dose was tested using logistic regression. Statistical analyses were performed using SPSS (v28) and p-values <0.05 were considered statistically significant.

Figure 1 PRISMA flowchart

Results:

A total of 24 publications [2-25] on 17 studies including 2080 patients published between 1995 and 2022 were analyzed. The PRISMA flowchart is presented in figure 1. Included were four randomized trials, ten prospective studies, one partly prospective and partly retrospective study, one feasibility study and one study on quality assurance. All studies reported on use of chemotherapy (12 yes, 4 no, 1 both), fractionation (13 standard fractionation, 3 hyperfractionation, 1 both), and boost volume (7 posterior fossa, 4 tumor bed, 6 both). Fourteen studies reported on treatment modality (8 photons, 1 electrons* , 2 photons or electrons*, 2 photons or protons, 1 photons or electrons* or protons). All studies reported information on the prescribed dose. For hyperfractionated treatment, all studies reported a CSI and tumor bed boost dose of 36 and 68 Gy, respectively (1 Gy/fr twice a day). For standard fractionation, eleven studies reported a CSI dose of 23.4 Gy, two studies 36 Gy, two studies 25 Gy, and one study 18 Gy. All studies reported a boost dose of 54.0-55.8 Gy. Fifteen studies reported median follow-up, and the median value (range) was 76 months (23-194). The 5-year PFS and OS were retrieved from 15 and 13 studies, respectively. The pooled 5-year PFS and OS were 77% (95% confidence interval [CI] 74 81%) and 84 % (95% CI 82-86%), respectively. The introduction of chemotherapy was the only treatment characteristics significantly affecting the PFS (figure 2). Logistic regression revealed no association with radiation dose. The random effects model suggests boost to the primary tumor vs. whole posterior fossa to be equivalent, in line with Michalski et al. [13]. In this analysis, two of the studies reported on survival for molecular subgroups [5, 13].