Considerations for the Clinical Implementation of MR-Guided ART for H&N Cancer

Aim: This work describes the steps taken to guide the clinical implementation of MR-guided adaptive radiotherapy for oropharyngeal (H&N) cancers on the MRL. Process: Healthy and patient volunteers were recruited to an ethically approved imaging study to aid in the commissioning of H&N cancers on the MRL. Volunteers were immobilised in thermoplastic shells to replicate treatment position. Image quality of vendor-provided T1 and T2 weighted images were assessed using visual guided assessments (VGAs), in which target volume and organs at risk (OARs) were graded on clarity by one oncologist and two radiographers for (n=10) patients. A multidisciplinary team (MDT) was identified, including therapeutic and diagnostic radiographers, clinical oncologists, and MR physicists. The MDT met alternate weekly to discuss considerations including immobilisation, treatment positioning, optimal imaging sequences for daily image registration, image registration methodology and an appropriate planning adaption strategy. Between meetings, further volunteers were imaged and five treatment plans were developed and assessed by clinicians and medical physics experts. Two adaptation strategies were considered: 1) a virtual couch shift (adapt-to-position or ATP) and 2) daily re-contouring based on target and OAR position (adapt-to-shape or ATS). Adapted plans were assessed clinically and compared to the original reference plan created on CT. MR-CT and MR-MR image registrations were performed by nine radiographers to identify a matching strategy, time taken to match and confidence of match using a 5-point Likert scale. Benefits/Challenges: MRL workflows allow for online re-planning and re-contouring of patients whilst eliminating extra radiation dose from additional CT and CBCTs. This is a benefit over conventional treatment, where offline re-planning can lead to gaps in treatment or continuation of treatment on suboptimal plans. Challenges included: ensuring minimal treatment time, patient discomfort and risk of claustrophobia. Image registrations were undertaken offline prior to service implementation to optimise radiographer speed and confidence at imaging matching. Dosimetric criteria were not met for ATP test plans, therefore a simplified ATS approach was taken, as the ATS optimisation algorithm was better at reproducing the intended dose. Original contours were rigidly propagated for each fraction, as opposed to complex structures being re-contoured daily. Contours were reviewed weekly offline by a clinician to determine if contour adaptation was necessary, and subsequently a full ATS workflow could be conducted if needed. Impact/Outcomes: We have successfully implemented MR guided H&N radiotherapy. The MDT identified T1 weighted sequences to be preferential. Further alternative sequences have been optimised and superior tumour and OAR visualisation has been recognised and communicated to the MRL Consortium for review. Multiparametric imaging is also being explored via a number of clinical trials within our institution to potentially aid in the facilitation of biological adaptive radiotherapy.