Evaluation of setup accuracy for prone breast radiotherapy with a three-dimensional surface imaging system

It has been shown that the detection of flat body surfaces (thorax, pelvis) using the three-dimensional optical surface imaging system (3D–OSIS) is less accurate than curved surfaces (head, neck). Here, we evaluated the applicability and technical performance of the 3D–OSIS for prone breast radiotherapy using silicon cup markers. Three silicon cups were attached on the dorsal surface of a female atom phantom. We evaluated the data reproducibility with these marker cups and compared the accuracy with or without their use. For the reproducibility test, we scanned an object 11 times every 15 s with the 3D–OSIS and registered the differences. For the accuracy test, we positioned the phantom on the table and manually moved it in the X, Y, and Z directions to predetermined values. These values were compared to those measured with the 3D–OSIS. This process was repeated 21 times and the differences of the measured values were compared according to the presence or absence of markers. In reproducibility test, the mean variations for the shifts in the lateral, longitudinal, and vertical directions were 0.15 ± 0.22 mm, 0.47 ± 0.13 mm, and 0.34 ± 0.19 mm, respectively. And the mean deviations elicited based on the accuracy test for the manually applied and the measured shifts in the lateral, longitudinal, and vertical directions were (a) 0.23 ± 0.1 mm, 0.044 ± 0.39 mm, and 0.18 ± 0.1 mm with the use of markers, and (b) 0.44 ± 0.33 mm, 0.97 ± 0.55 mm, and 0.63 ± 0.55 mm without the markers, respectively. The use of silicon cup markers yielded highly accurate results in the 3D–OSIS. The use of 3D–OSIS and silicon cups has potential in achieving precise patient setup and in reducing the CBCT frequency in prone breast radiotherapy.