Cherenkov Imaging to Compare Positional Accuracy of Right Breast Irradiation Setup Using Optical Surface Imaging vs. Traditional Laser Alignment

Abstract

<h3>Purpose/Objective(s)</h3> Cherenkov imaging can uniquely record beam shapes directly on the patient, offering a brand new means of assessing patient position and treatment accuracy. We used Cherenkov imaging to assess accuracy of two patient positioning techniques, optical surfacing imaging versus traditional skin marks and room laser alignment (skin marks/lasers), during right breast irradiation. <h3>Materials/Methods</h3> All imaging data was obtained with IRB approval from patients receiving free-breathing, right breast radiation therapy with either 6X or 6X and 10X beams. Data was obtained from two separate clinics that utilize different positioning methods. Patients treated at one clinic were positioned using optical surface imaging and patients treated at a second clinic were positioned using skin marks/lasers. For each fraction imaged, DICE coefficient (% similarity) and MDC (mean distance to conformity) were measured by comparing Cherenkov imaging obtained for that fraction to the Cherenkov imaging from the initial fraction. A two-tailed t-test of unequal variance was performed to obtain <b>P</b>-values comparing MDC and DICE from the two positioning techniques. <h3>Results</h3> Cherenkov images taken of 127 fractions using optical surface imaging for breast setup and 47 fractions using skin marks/lasers setup showed that MDCs were found to average 2.02 mm (std dev of 0.83mm) and 2.47mm (std dev of 0.68mm), respectively, resulting in a <i>P</i> value of 0.07. The mean DICE coefficients averaged 97.33% (std dev of 1.13%) and 97.19% (std dev of 1.90%) for optical surface imaging and skin marks/lasers, respectively, with a <i>P</i> value of 0.65. The data confirms the null hypothesis in both cases. <h3>Conclusion</h3> Comparison of our Cherenkov image analytics from two centers, each using a separate positioning method (optical surface imaging and skin marks/lasers), revealed that there is no statistical difference in terms of MDC and DICE coefficient. This analysis confirms that both setup methods can be used confidently for patient positioning without compromising quality of care. This novel method for assessing alignment techniques is easy to use on a daily basis, adds no extra time or imaging dose to the patient, and analysis is automated, making this tool ideal for evaluating patient positioning for large numbers of treatments. Future work will examine different methods for alignment of more complex treatments such as deep-inspiration breath hold for left breast radiation.