Inter-fraction heart displacement during voluntary deep inspiration breath hold radiation therapy without visual feedback measured by daily CBCT

PURPOSE/OBJECTIVE: Deep Inspiration Breath Hold (DIBH) is now considered as the standard of care for many breast cancer patients. However, there are still uncertainties about the dose given to the heart, and it is unknown if patients may improve voluntary DIBH depth by gaining experience during treatment. In this study, we will examine the interfractional three-dimensional (3D) heart displacement throughout voluntary DIBH (vDIBH) radiotherapy by means of daily cone-beam computed tomography (CBCT). MATERIAL AND METHODS: Two hundred twenty-five unique CBCTs from 15 patients treated in 15 fractions were analyzed. During CBCT, a vDIBH was conducted without any visual feedback. Patients performed their DIBH freely after receiving explanations and training. After daily CBCT matching to the chest wall (CW), surface-guided radiation therapy (SGRT) tracked DIBH depth to ensure that the CW position was the same as the daily acquired CBCT. The CBCTs were retrospectively registered to the DIBH planning-CT to calculate daily changes in heart displacement relative to the CW. RESULTS: The mean displacement of the heart during DIBH treatment relative to the DIBH planning-CT was as follows: 1.1 mm to the right, interquartile range (IQR) 8.0; 0.5 mm superiorly, IQR 4.8; and 0 mm posteriorly, IQR 6.4. The Spearman correlation coefficients (r(s)) were -0.15 (p=0.025), 0.04 (p=0.549), and 0.03 (p=0.612) for the X, Y, and Z directions, respectively. The differences in median heart displacement were significant: Friedmann rank sum test p=0.031 and pairwise comparison using the Wilcoxon rank-sum test were p=0.008 for X and Y; p=0.33 for X and Z; and p=0.07 for Y and Z. The total median heart motion was δ(tot median)= 7.26 mm, IQR= 6.86 mm. CONCLUSION: During DIBH, clinicians must be aware of the wide range of intra- and inter-individual heart position variations. The inter-individual heterogeneity shown in our study should be investigated further in order to avoid unexpected cardiac overexposure and to develop a more accurate heart dose-volume model.