Room: Foyer
Purpose: To investigate the setup errors recorded with multi-times ‘interrupt-scan-registration’ MVCT image guided registration scheme, and evaluate the feasibility of this method.
Methods: 21 patients performed TMI treatment from 2014-03 to 2016-05 were enrolled in our study. The image guided registration during treatment was divided into three sections for upper body treatment. First, a MVCT scan was conducted in the central region of Head&Neck (HN) part using normal mode, then register the MVCT images with original planning CT image using bone anatomy alignment method automatically then manually. After registration and couch shift we started the irradiation. When approaching to about 1/3 of the treatment, we interrupted and did a second scan in the region of Thorax&Abdomen (TA) region. To avoid the dose mismatch at the interrupt slice, the Y axis shift must be reset to zero. Reprocess this operation at 2/3 of the treatment in the region of Pelvic(p) region. For lower extremities plan, one MVCT image registration was performed only.
Results: The shift average of (X,Z) axis are (-0.45±2.24mm, 0.24±4.39mm) for HN, (0.12±2.68mm, -0.83±2.44mm) for TA, (0.25±3.5mm, -0.24±1.90mm) for P, and (-0.64±2.44mm, 0.51±2.56mm) for lower leg. The percentage that X shift falls into the interval of (-3, 3)mm, (-5, 5)mm and (-5, 5)mm is 84.13%, 92.06%, 85.71% for upper body respectively, for Z shift are 63.4%, 93.6%, 98.4% respectively. The backward original setup error shows that the setup error of (X,Z) are (0.62±3.05mm, -0.57±4.22mm) for T&A region, and (0.38±3.94mm, -0.82±4.23mm) for P region.
Conclusion: The original setup error before treatment is significantly large, especially in the adjacent tail end. 'interrupt-scan-registration' MVCT registration scheme would substantially reduce the variance of setup error feasibly, while compared with conventional cancer radiotherapy the shift is still vast. Results also imply that the PTV margin from bone should be defined in a cautious way.