Stereotactic reirradiation in the treatment of head and neck cancers: A retrospective study on the long-term experience of the Oscar Lambret Center.

BACKGROUND AND PURPOSE: Stereotactic radiotherapy potentially treats unresectable recurrences of previously irradiated head and neck (H&N) cancer. This study aimed to assess its efficacy and safety and evaluate prognostic factors. MATERIALS AND METHODS: We conducted a large retrospective series that included 110 patients who had undergone 36-Gy, six-fraction stereotactic reirradiation (CyberKnife®) for recurrent/secondary H&N cancer between 2007 and 2020 at the Oscar Lambret Center. Patient characteristics and toxicities were assessed. Overall survival (OS) and progression-free survival (PFS) were estimated using the Kaplan-Meier method. RESULTS: Median follow-up time was 106.3 months. The 2-year OS rate was 43.8 % (95 % confidence interval, 95 % CI, 34.3-52.9) and the median survival was 20.8 months (95 % CI, 16.5-26.3). The cumulative 2-year local-recurrence, regional-recurrence, and distant-metastasis rates were 52.2 % (95 % CI, 42.4-61.1 %), 12.8 % (95 % CI, 7.4-19.8 %), and 11 % (95 % CI, 6.0-17.6 %), respectively. 73 patients received concomitant cetuximab, and it was not significantly beneficial (HR = 1.34; 95 % CI, 0.80-2.26; p = 0.26). The cumulative incidences of grade ≥ 2 late toxicity was 42 % (CI95%: 33-51) at 24 months. Two grade 4 bleedings and no treatment-related deaths were reported. CONCLUSION: In a large retrospective series of SBRT reirradiation for recurrent or second primary H&N cancers, we observed a median OS of 20.8 months, with a cumulative incidence of grade ≥ 2 late toxicity of 42 % at 24 months. Such a treatment is feasible. However, local recurrence rates remain non-negligible, warranting further research. Radiosensitizer use is currently under study. Therefore, establishing a balance between therapeutic modifications and toxicity is essential.

Evaluation of surface image guidance and Deep inspiration Breath Hold technique for breast treatments with Halcyon.

PURPOSE: To evaluate the accuracy/agreement of a three-camera Catalyst Surface Guided Radiation Therapy (SGRT) system on a closed-gantry Halcyon for Free-Breathing (FB) and Deep Inspiration Breath Hold (DIBH) breast-only treatments. METHODS: The SGRT positioning agreement with Halcyon couch and cone-beam computed tomography (CBCT) was evaluated on phantom and by evaluation of 2401 FB and 855 DIBH breast-only treatment sessions. The DIBH agreement was evaluated using a programmable moving support. Dose agreement was evaluated for manual SGRT-assisted beam interruption and Halcyon arc beam interruption. RESULTS: Geometrical phantom agreement was < 0.4 mm. Couch and SGRT agreement for an anthropomorphic phantom resulted in 95% limits of agreement in Right-Left/Feet-Head/Posterior-Anterior (RL/FH/PA) directions of respectively ± 0.4/0.8/0.5 mm and ± 1.1/1.1/0.6 mm in the virtual and real isocenter. FB-SGRT-assisted patient positioning compared to CBCT positioning resulted in RL/FH/PA systematic differences of -0.1/0.1/2.0 mm with standard deviations of 2.7/2.8/2.4 mm. This mean systematic difference had three origins: a) couch sag/isocenter difference of ≤ 0.5 mm. b) Average reconstructed FB-CBCT images do not visually represent the average respiratory position. c) CBCT-based positioning focused on the inner thoracic interface, which can introduce a mean positioning difference between SGRT and CBCT. Manual SGRT-assisted beam interruption and arc interruptions resulted in mean gamma passing rates > 97% (0.5%/0.5 mm) and mean absolute differences < 0.3%. CONCLUSIONS: Accuracy was comparable with breast-only C-arm SGRT techniques, with different tradeoffs. Depending on the patient's morphology, real-time tracking accuracy in the real isocenter can be reduced. This study demonstrates possible discordances between SGRT and CBCT positioning for breast.

Abstract ID: 222 Clinical implementation of a Monte Carlo based QA platform for validation of Tomotherapy and Cyberknife treatment plans.

INTRODUCTION: This work describes the clinical implementation of a Monte Carlo based platform for treatment plan validation for Tomotherapy and Cyberknife, including a semi-automatic plan evaluation module based on dose constraints for organs-at-risk (OAR). METHODS: The Monte Carlo-based platform Moderato [1] is based on BEAMnrc/DOSXYZnrc and allows for automated re-calculation of doses planned with Tomotherapy and Cyberknife techniques. The Prescription/Validation module generates a set of dose constraints based on the anatomical region and fractionation scheme considered. Upon achievement of the planning, dose results are displayed with visual warnings in case of constraint violation. The system was tested on 83 patient cases in order to evaluate the influence of difference in calculation algorithms on OAR constraints. RESULTS: The first results with the Tomotherapy plans allowed for detecting and correcting a problem with the CT Hounsfield units when using a large reconstruction diameter (a CT artifact that lead to air voxels with an overestimated density). The Cyberknife results also showed some dose differences associated with different energy thresholds between Moderato and the Monte Carlo algorithm used in the Treatment Planning Station. Regarding OAR constraints, re-calculation generated few violations in thoracic, pelvic and abdominal cases. However, in spinal and head cases, significant differences can appear (-11% to +6%) on optic pathways and spinal cord, leading to doses above the limits. CONCLUSIONS: The Moderato platform constitutes a promising tool for the validation of plan quality, offering both dose re-calculation and OAR constraints evaluation. First results show the importance of this verification for some specific regions. Further work is ongoing to optimize the quantity and relevance of the information displayed, before fully introducing the system in clinical routine.