MR-Guided Radiotherapy (MRgRT) for Laryngeal Cancer With Real-Time Visualization of Intrafraction Larynx Motion.

PURPOSE/OBJECTIVE(S) Although early larynx cancers require treatment of only the vocal cord(s), radiation for early larynx cancer historically consisted of opposed lateral portals that included the entire larynx. Treatment techniques have advanced to 3D conformal and intensity modulated radiotherapy (IMRT), but the treatment target has remained the entire larynx. One concern preventing volume reduction in these patients is periodic swallowing that occurs during treatment delivery due to normal patient physiology. Such intrafraction swallowing can displace the treatment target up to 2 cm and could lead to tumor underdosing. Treatment with real-time target tracking could greatly mitigate this concern. Herein, we describe our initial experience using MR-guided radiotherapy (MRgRT) to visualize and understand the impact of intrafraction larynx motion. MATERIALS/METHODS Three patients with cT1-2 laryngeal cancer were treated using an MRgRT Linac. Treatment volume included the entire larynx to evaluate feasibility of setup, immobilization, patient tracking, and treatment planning. A commercial MR-compatible head-and-heck immobilization system was developed and utilized, consisting of an indexable baseplate with integrated MR receiver coils affixed to a head cradle, with an open-face thermoplastic mask and head-mold cushion. Treatment plans were inversely planned utilizing 5 anterior IMRT fields to eliminate entrance dose to posterior organs at risk (OARs). The tracking structure consisted of patients' intrinsic larynx muscles delineated by a head and neck radiation oncologist on a 35cm x 35cm x 0.70cm sagittal MR plane. Patients were instructed not to swallow during setup imaging and during treatment. Continuous real-time tracking was performed at 4 frames per second with automatic beam hold if > 5% of the area of the tracking structure exceeded the 3mm isotropic boundary limit. RESULTS Images were taken during patient treatment, and movement of the intrinsic muscles of the larynx was apparent in all patients due to periodic deglutition. At the time of analysis, 63 fractions have been delivered. The average total in-room time for the three patients in the study was 27:22 minutes (range 18:00 - 45:00 minutes), while the average beam-on time was 8:44 minutes (range 4:00 - 30:00 minutes). CONCLUSION MR real-time visualization of the patient's intrinsic laryngeal muscles during MRgRT provides a method of gating treatment during patient swallowing. Robust intrafraction target tracking appears to enable higher confidence in reducing the treatment volume for early larynx cancers such that improvements in preservation of voice quality and deglutition capabilities are possible without sacrificing local control. A clinical trial is planned to examine such volume reduction and its effect on patient quality of life and function.