Protocol compliance in a multicentric phase III trial investigating scheduled adaptive radiotherapy and dose painting in head and neck cancer

Purpose: To report on quality assurance (QA) and protocol adherence (PA) in a multicentre phase III trial for head and neck cancer, evaluate patterns of protocol deviations and investigate the effect of PA on study outcomes. Methods: All 221 patients from the ARTFORCE trial (NCT01504815) were included in this study. Pre- and pertreatment QA measures included protocol guidelines, a dummy run, early case reviews and trial meetings. FDG-PET-guided dose painting and scheduled adaptive radiotherapy were reviewed in patients in the experimental arm (eRT). Patient and disease characteristics, as well as institutes' accrual rate and timing were examined for correlation with PA. Cox regression was used to determine the impact of PA on outcome. Results: The dummy run was completed in all nine institutes and early case reviews were completed in five out of nine institutes that contributed 190 out of 221 patients. Among all patients randomized to eRT, 64 % had at least one deviation of the experimental trial components. Protocol deviations were significantly correlated with the institute patients were treated at (Cramer'sV 0.34-0.48). Despite early identification of institute-specific deviations in QA, these continued during the trial. No significant associations were seen between deviations and accrual timing or rate (P >= 0.26). Within eRT, no significant relation was observed between experimental PA and locoregional control (LRC), the primary endpoint of the trial (P >=.15). Conclusions: Despite QA, protocol deviations persisted during the trial, which were mostly institute-specific. However, deviations of the experimental treatment strategy did not significantly impact LRC and therefore the trial conclusion.