RESUMO
@#<b>Objective</b> To compare the set up errors derived from different registration methods of the X-ray volume imaging (XVI) system for radiotherapy in the treatment of middle/lower-segment esophageal cancer, and to provide a reference for radiation treatment of esophageal cancer. <b>Methods</b> We randomly selected 63 patients with middle/lower-segment esophageal cancer, and obtained their reconstructed XVI images at the first therapy to perform automatic registration with gray-value and bone registration methods. We acquired and compared the three translation errors (along <i>x</i> [left to right], <i>y</i> [head to feet], and <i>z</i> [front to back] axes) and three rotation errors (around the <i>x</i>, <i>y</i>, and <i>z</i> axes) derived from the two registration methods. <b>Results</b> Gray-value registration had significantly smaller translation errors along the <i>x</i> and <i>z</i> axes than bone registration (<i>x</i> azes <i>t</i> = −2.78, <i>z</i> azes <i>t</i> = −2.15, <i>P</i> < 0.05), but there was no significant difference along the <i>y</i> axes (<i>P</i> > 0.05). The rotation errors around the three axes were all smaller than 1°, and were smaller with gray-value registration than with bone registration, but without significant differences (<i>P</i> > 0.05). <b>Conclusion</b> We recommend gray-value registration for radiotherapy in the treatment of middle/lower-segment esophageal cancer. Manual verification or fine-tuning is recommended after automatic registration in clinical practice. Besides translation errors, rotation errors should also be paid attention to.
RESUMO
Objective The literature study the setup errors of head and neck, thoracic, abdominal and pelvic tumors by megavoltage fan-beam CT based image guidance in TOMO-HD to provide the margin enlarging from clinic target volume (CTV) to planning target volume (PTV) in treatment planning system of TOMO-HD. Methods 103 patients with head and neck (30 patients), thoracic (42 patients), abdominal and pelvic (31 patients) carcinoma were enrolled. Megavoltage fan-beam CT based image guidance in tomotherapy-HD was used to acquire CT scan before every treatment. The left-right (X), superior-inferior (Y), anterior-posterior (Z) and rotation (Fy) setup errors of patients can be obtained from the tomography image automatically restructured by the system. Calculating the systematic error and the random error in the three dimensions and check whether the setup data accord with the normal distribution or not, then acquire the data expand in the three directions. Results According to 2593 fan-beam CT scans, the shift errors (µ ± s) in X, Y, Z and Fy (rotation) of three study group were [(−0.31 ± 2.16) mm、(1.09 ± 3.56) mm、(2.36 ± 2.27) mm, (0.29 ± 0.96)°] (head and neck tumor), [(−0.98 ± 2.95) mm、(0.45 ± 6.86) mm、(3.79 ± 2.47) mm, (0.18 ± 0.60)°] (thoracic cancer) and [(−0.86 ± 2.85) mm、(−1.59 ± 6.91) mm、(5.77 ± 2.40) mm, (0.20 ± 0.68)°](abdominal and pelvic carcinoma). The systematic errors (∑) and random errors (σ) in X, Y, Z dimensions of patients with head and neck, thoracic, abdominal and pelvic tumors were (1.06 mm and 1.84 mm), (1.93 mm and 3.43 mm), (2.41 mm and 2.71 mm), (1.10 mm and 2.56 mm), (3.79 mm and 5.46 mm), (1.38 mm and 1.99 mm) and (1.39 mm and 0.87 mm), (4.98 mm and 5.69 mm), (1.19 mm and 2.05 mm), respectively. Conclusion It is recommended as a reference for image guidance in TOMO-HD according to the frequency distribution of setup errors, for patients with head and neck, chest and abdominal and pelvic tumors, the maximum range of motion in three dimensions are (5.00, 5.00, 5.00) mm, (6.63, 17.25, 16.00) mm and (6.49, 16.24, 13.60) mm.