Prospective Randomized Trial Comparing 2 Devices for Deep Inspiration Breath Hold Management in Breast Radiation Therapy: Results of the BRAVEHeart Trial

Abstract

The Breast Radiotherapy Audio Visual Enhancement for sparing the Heart (BRAVEHeart) trial prospectively randomized patients with left-sided breast cancer to 1 of 2 deep inspiration breath hold biofeedback devices: a novel chest surface tracking system and an abdominal block tracking system. The primary hypothesis was that the accuracy of chest tracking would be higher than that of abdominal tracking as the chest is a more direct surrogate of the breast target. Patients with left-sided breast cancer were treated in deep inspiration breath hold with intensity modulated radiation therapy delivery. Patients were randomized to either the novel chest surface system or abdominal block system for active management of breath hold with visual feedback. On both trial arms, the unallocated system was monitored passively. A total of 239,296 cine electronic portal imaging device images were analyzed retrospectively to extract the chest wall position. Treatment accuracy was quantified as the deviation of the internal chest wall during treatment relative to the planned position from the digitally reconstructed radiograph. The correlation between motion of the external surrogate and internal chest wall was calculated per-breath hold. Ease of use was assessed with questionnaires for both radiation therapists and patients and appointment length recorded. Data from 26 participants were available for analysis. No difference was found in delivered treatment accuracy between arms. Across all patients and fractions, the median correlation between internal chest wall movement and external surrogate was 0.69 for the chest surface and 0.17 for the abdominal block. Patients found it easy to follow visual feedback from both systems. No difference was found in appointment length between arms. No statistical evidence was found for superior treatment accuracy, satisfaction, or appointment length for the novel chest surface tracking device compared with the abdominal block system. During deep inspiration breath hold, the median per-breath hold correlation of internal chest wall movement to the motion of the chest surface was higher than the median correlation of the abdominal block to the chest surface.