(P 445) ISOTOXIC RADIOTHERAPY DOSE ESCALATION IN UNRESECTABLE PEDIATRIC EWING SARCOMA: A DOSIMETRIC FEASIBILITY STUDY

Objective: Radiotherapy (RT) is the definitive local therapy for unresectable localized or metastatic Ewing sarcoma (ES). A recent phase III trial testing standard-dose radiotherapy (SDRT) versus dose-escalated radiotherapy (DE-RT) showed improved local control (LC) with DE-RT. We hypothesized that utilization of a simultaneous integrated boost (SIB) to gross disease could enable isotoxic DE-RT for pediatric ES patients receiving definitive radiation therapy without increasing the number of radiation fractions, and that feasibility of this approach could support a phase I trial.

Methods: We retrospectively identified patients treated from 2012-2022 with pathologically confirmed ES, treated with definitive RT to a prescribed dose of 50.4 Gy to 55.8 Gy in 1.8 Gy fractions after induction chemotherapy. Patients who received surgery with gross residual disease were included. Standard target volumes included the pre-chemotherapy soft tissue and osseous extent for PTV1 (45 Gy in 1.8 Gy fractions), the post-chemotherapy soft tissue involvement and pre-chemotherapy osseous extent for PTV2 (5.4-10.8 Gy in 1.8 Gy fractions). The pre-chemotherapy osseous involvement was the subvolume selected for DE-RT (PTV3). OAR dose constraints were primarily adopted from AEWS1031. In addition, we contoured cauda equina and sacral plexus and converted AEWS1221 dose constraints to an isotoxic dose equivalent in 28 or 31 fractions (α/β = 2). Isotoxic DE-RT was assessed at prespecified levels of 58.8 Gy, 67.2 Gy, and 75.6 Gy in 28 fractions or 65.1 Gy, 74.4 Gy, and 83.7 Gy in 31 fractions, defined as levels 1, 2, and 3, respectively. OAR constraints were prioritized over SIB dose escalation; dose escalation success was defined as V95% ≥ 95% of target volume while respecting constraints. We recorded DE-RT level achieved, OAR D5cc/D3cc and D0.03cc, DE-RT target volume, and conformity index (CI), gradient index (GI) with respect to adjacent dose-limiting OAR (cauda equina, sacral plexus, and spinal cord), and homogeneity index (HI) for original RT and DE-RT replan levels in each patient. RT planning was performed in Eclipse version 15.6 (Varian, Palo Alto, CA), and differences in CI, GI, and HI between original RT and DE-RT were assessed with matched-pairs t-test (α = 0.05) in SPSS v27.0 (IBM, Armonk, NY).

Results: We identified ten patients meeting inclusion criteria, with a median age of 11 years (IQR 7-16). Unresected or residual ES was in the pelvis in five patients, lumbar spine and paraspinal musculature in three patients, and cervical/thoracic spine and paraspinal musculature in two patients. Two patients underwent partial resection prior to RT, and all patients received induction chemotherapy. Seven patients were successfully replanned to level 3 DE-RT, one to level 2, and one to level 1; one patient could not achieve DE-RT replan due to OAR (spinal cord) within the target volume. Median OAR and target doses at the highest DE-RT level achieved are reported in Table 1. Gradient indices with respect to each critical OAR were significantly increased with DE-RT at all dose levels excepting level 1 sacral plexus (Table 2). DE-RT replans displayed decreased homogeneity and increased conformity with respect to the SIB volume as compared to prescribed plans (Table 2). Median DE-RT target volume was 20.03cc (IQR 7.4-44.4).

Conclusion: Isotoxic DE-RT using SIB was feasible for the majority of pediatric ES patients. DE-RT using SIB resulted in elevated GI, increased conformity, and decreased homogeneity, and safe delivery would require closely monitored image guided RT. A phase I trial of isotoxic dose escalation using our prespecified dose levels is planned.