ESTRO 2023 - Abstract Book

S920

Digital Posters

ESTRO 2023

Patients treated with frameless SRS were included in the retrospective study. The three immobilization devices were: CIVCO’s trUpoint ARCH ™ SRS/SRT system, BrainLAB mask fixation system, and POCL cradle mask. All the patients were treated with single isocentre LINAC SRS equipped with cone beam computed tomography (CBCT). The workflow included an initial CBCT to assess translational errors in 3 dimensions and matching. A subsequent CBCT was undertaken to evaluate residual errors before treatment delivery.

Results 143 fractions from 36 patients treated between January 2021 and June 2022 were included in the analysis. 65 and 78 fractions were delivered for benign tumors and brain metastases, respectively. Single, 3, and 5 fractions were done in 17%, 17%, and 64%, respectively. BrainLAB, CIVCO, and POCL were used for 71, 47, and 25 fractions, respectively. The mean shifts obtained from index CBCT were 1.05 mm (antero-posterior, AP), 1.74 mm (supero-inferior, SI) and 1.15 mm (medio lateral, ML). Following the application of shifts with repeat CBCT, the shifts were 0.19 mm, 0.24 mm, and 0.19 mm. Residual shift data were available for 83 fractions. Using the Van Herk and Stroom formula, the recommended PTV margins are 1.15, 1.15, 0.62 mm (AP, SI, ML) and 1.41, 1.39, 0.76 mm (AP, SI, ML), respectively, for the overall cohort. It was least for CIVCO 0.38 mm (Van Herk) and 0.47 mm (Stroom) (maximum in any direction). The mean vectors initially were 2.3, 2.7, and 3.1 mm for CIVCO, BrainLAB, and POCL mask, respectively. The corresponding values for residual shifts were 0.13, 0.73, 1.87 mm, respectively, for CIVCO, BrainLAB, and POCL. In terms of mean residual vector, CIVCO and BrainLAB were significantly better than POCL (p=0.00 and p=0.049, respectively). CIVCO was significantly better than BrainLAB (p=0.01). The mean 3D residual vector was 0.56 mm for the first fraction vs 0.71 mm for subsequent fractions (p=0.66). Conclusion CIVCO provided the best immobilization, followed by BrainLAB. With the use of CBCT and online corrections, a PTV margin of 1 mm can be used in appropriate cases for CIVCO and BrainLAB, while 2mm should be used for POCL. 1 The Christie NHS Foundation Trust, Networked Services- Proton beam therapy, Manchester, United Kingdom; 2 The Christie NHS Foundation Trust, Proton Clinical Outcomes Unit, Manchester, United Kingdom Purpose or Objective Ependymomas constitute around 20% of all primary spinal cord tumours. Surgery may be followed by post-operative radiotherapy (RT) depending on extent of resection, histopathology and tumour grade. We retrospectively compared acute toxicity with post-operative proton and photon RT. Materials and Methods We included all patients receiving proton RT at the Christie Hospital from 2018-2021 and all adults receiving photon RT from 2013-2018. For analysis, to account for 3 photon plans being 2-phase (initial larger volume and smaller boost volume), as well as differences in proportion of spine irradiated, we calculated Averaged Dose (Gy) = D1+(PTV2/PTV1)D2, and Normalised Dose = Averaged Dose × (PTV1/SL), where D1=phase1 dose (Gy), D2=phase2 dose (Gy), SL=spine length (cm), and PTV1 and PTV2 are the length (cm) of the Planning Target Volume (PTV) in phase1 and 2, respectively. We collected acute RT toxicity data including haematological toxicity from patient notes. Linear regression models examined the effect of RT type on blood parameter nadirs (%), with a significance level of 5%. Box plots examined incidence of toxicities by RT type. Results Of 12 patients, 6 were female, and 6 treated with protons. Median ages were 42 years for photon RT and 13 years for proton RT. The Total Dose (D1+D2) for proton plans ranged from 50.4Gy to 54Gy in 28 to 30 fractions, and for photon plans 45Gy to 50.4Gy in 25 to 30 fractions (including phase2 doses from 9Gy to 15Gy in 5 to 10 fractions in 3 cases) (Table 1). The PO-1150 Spinal ependymoma: Is treatment with proton therapy superior to photon therapy? D.H. Kannikanti 1 , A. France 2 , S. Pan 1 , G. Whitfield 1