ESTRO 2021 Abstract Book

S1498

ESTRO 2021

L. Masi 1 , R. Doro 1 , V. Di Cataldo 2 , L. Redapi 1 , G. Francolini 3 , L. Visani 2 , L. Livi 3,4 1 IFCA, Department of Medical Physics, Florence, Italy; 2 IFCA, Department of Radiation Oncology, Florence, Italy; 3 AOU Careggi, Radiotherapy Unit, Florence, Italy; 4 University of Florence, Clinical and Experimental Biomedical Sciences "Mario Serio", Florence, Italy Purpose or Objective A trial (NCT03520894-ROCK) enrolling early breast cancer patients for single fraction (SF) CyberKnife (CK) preoperative radiotherapy is currently running in our centre. The present study aims to evaluate for the enrolled patients: i) respiratory patterns ii) CK tracking uncertainties ii) dosimetric effects of tracking errors. Materials and Methods 12 ROCK enrolled patients were treated at 21 Gy SF using CK fiducial-based real time respiratory tracking (Synchrony). Log-files generated by Synchrony for each patient were analysed and data extracted along superior-inferior (SI), right-left (RL) and anterior-posterior (AP) directions. Tumor motion was calculated as fiducial center of mass modelled excursion (1%-99%). Correction for baseline drifts (BD) was performed using a moving average algorithm. Synchrony residual tracking errors, i.e., correlation and prediction models errors, were computed at each fraction from the corresponding log files. Synchrony uncertainties were compared with errors obtained if respiratory tracking was not used (No Synchrony): fiducial tracking without respiratory compensation was simulated for different imaging intervals (45s-60s). Dosimetric consequences of tracking uncertainties were simulated for 10 patients, using a 3 mm CTV-PTV margin. Each plan was modified adding the respective tracking errors to each beam robot’s position and the perturbed dose distribution calculated on the planning CT scan. Results After correction for BD, respiratory amplitude along the 3 directions was below 5.5 mm and higher in AP (3.1 ±0.9 mm), than in RL (1.1±0.7 mm) and SI (2.0±0.9 mm). BD above 3 mm were observed in 58% of patients with a maximum of 5.5 mm. Prediction model errors remained always below 0.9 mm and 95% of absolute values were within 0.25 mm. Correlation model errors over all patients along SI, RL and AP were respectively (-0.13±0.91 mm), (-0.10±0.77 mm) and (0.03±1.10 mm). For each patient, standard deviations (SDs) of Synchrony and No Synchrony fiducial tracking errors were comparable. Both sets of SDs remained below 2.5 mm. Values for each patient are shown as a function of respiratory excursion in Fig.1 along AP, where the largest excursion and errors were found. Perturbed dose distributions showed a CTV coverage above 95% for all but one patient (92%) (Fig.2). Skin constraints were violated in 1 patient but differences remained below 2

cc.

Conclusion Respiratory patterns indicated small breathing excursions but not negligible baseline drifts. Due to the small