Room: Exhibit Hall
Purpose: To present an analytical dosimetric metric for evaluation of setup error in external beam radiation therapy that accounts for mobile deformable structures.
Methods: A stereotactic body radiation therapy spine case was originally planned on a diagnostic CT(A). At the time of treatment a Cone Beam CT was obtained that revealed significant changes in bowel filling and another diagnostic CT(B) was obtained. Conventional setup error computed with auto matching on rigid bony anatomy between data sets CT(A)-CBCT was comparable to CT(B)-CBCT. The treatment plan for data set CT(A) was recalculated on data sets CBCT and CT(B). The plan on data set CT(B) was also calculated with a 1mm shift in the three principal axis. For each plan the 90,100,110 and 120% isodoses were converted to structures and the Dice Coefficient DK computed as DK=2*I/S where I is the intersection of each isodose for plan sets CT(A)-CBCT and CT(B)-CBCT and S is the sum of the respective volumes.
Results: (1)Anatomical evaluation of setup error is insensitive to soft tissue changes. (2)Dosimetric evaluation of setup error using Dice Coefficient is sensitive to 1mm shifts with typical 5% change in DK. (3)DK for the 90,100,110 and 120% isodoses were 0.986, 0.978, 0.871, and 0.750 for sets CT(A)-CBCT and 0.995, 0.994, 0.954, and 0.921 for sets CT(B)-CBCT. The DK metric clearly distinguishes the improved setup and subsequent dosimetric correspondence between the CBCT and the repeat CT(B).
Conclusion: In clinical practice the setup is a combination of anatomical â€œauto-matchâ€? followed by user dependent shifts based on numerous clinical factors such as fiducials, proximity to mobile deformable critical structures, and contour changes due to bowel filling or weight loss. Since multiple conditions can override geometric considerations, setup error especially in the era of six degree of freedom can best be evaluated using dosimetric parameters.