< Back to topics
> Cover Page, Authors, Institutions
> View tables/images/videos
To date, online correction data from five patients has been acquired. The mean magnitudes of residuals after correction were 0.5, 1.3, and 0.6 mm in RL, AP, and SI directions. The greater error in AP was not present in the phantom study and thus appears patient-related. System reproducibility in the absence of a shift was good, so remaining errors are attributable to accuracy of couch motion. Individual patient setup errors (mean 7 mm) tended to be systematic, being of similar size and direction in each session. Fluoroscopy showed clip excursion due to breathing to be <3 mm. The retrospective part of the study showed the difference between clip-based correction and the "true" shift to be ~3 mm. This discrepancy is correlated to the amount of post-planning cavity shrinkage: on average ~4 mm radial, mostly within 4 weeks of surgery. Rotation analysis showed a mean change of 4 degrees in any direction. In only 10 of 23 cases did rotation correction improve target coverage. The likely reason is that a rotation correction must, in practice, be made around the clips' center of mass, which, as noted above, averages 3 mm off the true cavity center.
Note: an AVI file of clips moving under fluoroscopy is available under "View tables/images/videos."
Click on images to enlarge (a new pop-up window will open)