A multi-centre analysis of treatment procedures and error components in dynamic tumour tracking radiotherapy.

Abstract

This study aimed to compare procedures for dynamic tumour tracking (DTT) using a gimbal-mounted linac between centres in Japan (KU-IBRI) and Belgium (UZB), to quantify tracking error (TE), and to estimate tumour-fiducial uncertainties and PTV margins. Twenty-two patients were evaluated. TE was divided into components originating from the patient, fraction, segment, and residuals. KU-IBRI applied DTT to lung cancer, while UZB treated both the lung and liver. Patients from UZB were younger and had a higher body mass index. DTT procedures differed in the use of body fixation, correction for set-up error, type of fiducial markers, and goodness of fit of correlation model. TE was larger at UZB in the intra-fraction components, whereas the tumour-fiducial uncertainties were estimated to be larger at KU-IBRI. These results ultimately led to similar PTV margins at both centres (2.1, 4.2, and 2.6 mm for KU-IBRI; 2.4, 3.6, and 2.0 mm for UZB in LR, AP, and SI, respectively, for 99% coverage of patients). Several differences in procedures and patient characteristics were observed that affected TE and tumour-fiducial uncertainties. This analysis confirmed similar accuracy in DTT delivery and adequate PTV margins in the different centres based on their local specific workflows.