| Objective: To study the setup errors in three-demensional conformal radiotherapy (3D-CRT) of thoracic esophageal carcinoma using an electronic portal imaging device (EPID) and calculate the setup margins.Methods: The study includes three stages:(1) The forty-one patients of thoracic esophageal carcinoma, who were treated with 3D-CRT in our hospital from Jun 2007 to February 2008, entered into this study. The anterior and lateral isocentric electronic portal images (EPIs) of every patient were recorded by EPID once a week. The 227 pieces of EPIs were matched with digitally reconstructed radiograph (DRR) using a software called Vision Review. With the software, the traslational setup errors, in X-axis (left-right direction), Y-axis (superior-inferior direction) and Z-axis (anterior- posterior direction), and the rotational setup errors around X-axis or Z-axis could be measured. Then the individual or global systematic errors and random errors were calculated. (2) According to the formula for setup margin, the setup margins of three axises in 3D-CRT of thoracic esophageal carcinoma were produced. (3) Simulative therapy plans A were produced with deleting the multileaf used to protect the spinal cord in the 22 therapy plans of radical cure, which had been designed with the setup margin of 10mm and executed on the 22 patients of thoracic esophageal carcinoma. On basis of plans A, the simulative plans B were produced with adjusting the planning target volumes and sizes of irradiation fields according to the margins calculated in this study. The two series of plans were evaluated by the dose volume histogram. And the Paired Samples T-Test or Wilcoxon Signed Ranks Test was applied for comparing the differences. Results: (1) The medians of individual systematic translational errors in X-axis, Y-axis and Z-axis were 3.08mm (-7.20～5.86mm), 4.30mm (-7.03～7.95mm) and 2.10mm ( -3.38～3.65mm), and the medians of random errors for each axis were 1.96mm (0.35～8.75mm), 1.81mm (0.76～8.17mm)and 1.00mm (0.14～6.46mm), respectively. The global systematic translational errors of X-axis, Y-axis and Z-axis were 0.13mm, 0.86mm and -0.43mm, and the random errors were 2.53mm, 3.76mm and 1.95mm, respectively. And the medians of individual systematic rotational errors around Z-axis and X-axis were 0.78°(-1.93°～1.47°) and 0.32°( -0.39°～0.53°), and the median of individual random errors was 0.42°(0.01°～3.35°)around Z-axis and 0.32°(0.03°to 0.83°) around X-axis. The global systematic rotational error was 0.32°around Z-axis and 0.10°around X-axis, and the random error was 0.82°around Z-axis and 0.42°around X-axis. (2) The margin of planning target volumes in 3D-CRT of thoracic esophageal carcinoma was 8.72mm for X-axis, 10.50mm for Y-axis, and 5.62mm for Z-axis. (3) Between the two serieses of simulative plans , the tumor control probability (TCP), normal tissue complication probability (NTCP) for lungs, the minimum dose in planning target volume of gross tumor volume (PTV-gtv), the mean dose in PTV-gtv and the minimum dose for 95% of PTV-gtv had no difference(P =0.630,0.056,0.106,0.697 and 0.611). And the differences of maximum dose and NTCP of spinal cord between group A (4638.7±1449.6cGy, 4.82%±5.99%)and group B (4310.2±1528.7 cGy, 3.64%±4.70% ) were significant (P =0.000 and 0.007).Conclusions: In 3D-CRT of thoracic esophageal carcinoma of our department, the systematic translational errors for X-axis, Y-axis and Z-axis were 0.13mm, 0.86mm and -0.43mm, and the random errors were 2.53mm, 3.76mm and 1.95mm. And the systematic rotational error was 0.32°around Z-axis and 0.10°around X-axis, the random error was 0.82°around Z-axis and 0.42°around X-axis. And the setup margin was 8.72mm for X-axis, 10.50mm for Y-axis, and 5.62mm for Z-axis. Compared with the setup margin of 10mm in each axis, the result of this study can protect spinal cord more effectively.
|
PDF Full Text Request