Visual biofeedback on the Unity MR-linac as driver for treatment efficiency
Pim Borman,
The Netherlands
PO-1718
Abstract
Visual biofeedback on the Unity MR-linac as driver for treatment efficiency
Authors: Pim Borman1, Jasmijn Westerhoff1, Neil Winchester2, Chris Knox2, Jan Kok1, Bas Raaymakers1, Martin Fast1
1University Medical Center Utrecht, Department of Radiotherapy, Utrecht, The Netherlands; 2Elekta, AB, Stockholm, Sweden
Show Affiliations
Hide Affiliations
Purpose or Objective
Visual biofeedback systems based on online MRI may facilitate the dosimetrically optimal delivery of gated radiotherapy through instructed breath-holding. Providing legible instructions and geometrically correct guidance to patients is key to the success of such a system. In this study, we present a novel visual biofeedback setup for the 1.5 T Unity MR-linac (Elekta AB, Stockholm, Sweden). We quantify subject compliance with breathing instructions and physiological impact of (repeat) breath-holding.
Material and Methods
A visual biofeedback system was developed for Unity (Fig. 1). The system hardware consists of a 32" MRI-compatible screen (CRS Ltd, Rochester, UK) mounted at the end of the MRI bore, and an in-bore mirror assembly that facilitates on-couch screen visibility. Three healthy volunteers were scanned with a sagittal bFFE 2D-cine MRI sequence (4 Hz, voxel resolution 2.5x2.5x10 mm3). Using in-house software, 2D-cine images were deformably registered to an end-exhale reference image. Cranial-caudal motion was extracted from an ROI centered at the diaphragm and used as position input for a speedometer-style user interface. Volunteers were asked to follow a series of breathing instructions: (fixed) operator-driven 12 s breath-holds followed by 8 s recovery periods; (rest) 1 min free-breathing for recovery between breath-hold series; (flexible) subject-driven self-determined breath-holds and recovery periods. The fixed and flexible breath-hold series contained a total breath-hold time of 5 mins each. The 5 mm gating window was centered on end exhale. The heart rate was monitored continuously during imaging using an MRI-compatible pulse oximeter with finger wrap sensor (NONIN Medical Inc, Plymouth, MN). Subject’s heart rate and residual motion in the gating window were compared between the two breath-hold modes. Gating efficiency was defined as time in the breath-hold window versus total time per series.
Results
The heart rate (Fig. 2) typically showed a modest increase (10 bpm) at the beginning of each breath-hold series which then subsided within 2 mins, indicating that the subjects became more comfortable as each series progressed. The fixed breath-holds resulted in gating efficiencies of 59%, 60%, and 59%, close to the maximum of 60% due to the fixed 8 second recovery period, and a residual motion in the gating window of –0.6±0.7mm, -0.5±0.7mm, and 0.5±0.8mm depending on volunteer. The flexible breath-holds resulted in efficiencies of 76%, 81%, and 97%, with residual motion of -0.5±0.8mm, -0.8±0.8 mm, and –0.4±1.3mm. Volunteer 3 chose to stay within the breath-hold window continuously by breathing shallow.
Conclusion
Healthy volunteers were easily able to comply with breath-hold instructions provided via visual biofeedback on the Unity MR-linac. Any physiological discomfort (increase in heart rate) was minor and fleeting. The flexible, subject-driven breath-hold mode resulted in a higher gating efficiency in our small cohort while minimally increasing residual motion.