Evaluation of the image guided set-up for multiple metastases treated with tomotherapy

Objective: To compare the setup errors between single-site and two-site image guidance in treating multiple metastases using Tomotherapy. Methods: A total of 1220 sets of megavoltage CT (MVCT) images from 50 multiple metastases patients were collected. The setup errors of two anatomic sites were determined by registration of MVCT images with planning images. Bland-Altman plot analysis was used to assess the coincidence of these two methods. Pearson correlation analysis was performed to evaluate the correlation of the setup errors determined by two sets of data and to analyze the deviation values of setup errors. Results: The deviation values of setup errors more than 3 mm between two sites were 34%, 46% and 28% in lateral (x), longitudinal (y), vertical (z) directions, respectively. The deviation values of setup errors more than 5 mm were 10%, 16% and 8%,respectively. The Bland-Altman plot analysis showed that the 95% agreement limits of agreement were (9.3, -10.6), (10.5, -11.7), (7.3, -6.9) mm in x, y, z directions, respectively, which were all out of 5 mm tolerance. The Pearson coefficient of correlation along all three directions was less than 0.05, and R² was 0.074, 0.475, and 0.178 in x, y, z directions, respectively. Conclusions: To determine the setup errors for patients with multiple metastases, single-site image guidance method is not consistent, and the two site image guidance method would be recommended. (authors)