Hello from Tampere
With the limbs we use vagbacks indexed on the couch. We use markings on the patient skin to see initial straightness of the patient. After that we use TrueBeam automatic button to go to acquired couch values and use AlignRT send to couch button to shift the patient to near correct isocenter, we do not send rotations in this phase. Then we check the postural video to see the posture of the limb. If it is ok based on 3 cameras, we are happy and take images. If the posture is not ok based on video, we try to setup the limb mostly based on video. To be honest, if we have to choose, we rely more on video than RTD rotations (rot, pitch, roll) values, of course in optimal cases we take both into account. As well, in these limb cases correct posture of the target can be finally achieved with CBCT and with 6D couch shifts, so small inaccuracies at the setup can be now mostly corrected with daily CBCT.
I know that several units are happy with the markerless limbs, using both ROI and video for every patients, and after a while I think this is the case here as well. ROI should be large enough, I think. Too small ROI leads to inaccuracies. Already in our practice only using mostly postural video (+ROI) could be in most cases as accurate as markings + video (+ROI)? However, limbs are not that easy to setup with ROI.
The main reason based on my experience is the anatomy of the limbs. We do not necessarily want to take knee into ROI if we are treating thigh for example, also pelvis may be too far away from isocenter. There are situations that we cannot avoid creating ROIs including mostly cylindrical shapes, if we want to create ROIs near isocenter. Cylindrical shapes are the one we should avoid, since it is difficult for the SGRT to detect the correct location and posture. As well, there may be remarkable edema and deformation visible on the surface with these patients making SGRT setup difficult. On the other hand, as an advantage for the SGRT; SGRT may give valuable information in the deformation cases and new reference surfaces for the setup can be taken as well.
For the intrafractional motion control AlignRT is of course useful as well.
Here some references:
https://www.sciencedirect.com/science/a ... 9424000897
j.prro.2013.09.00120230626-1-9a4p50-libre.pdf
https://link.springer.com/article/10.10 ... 24-00806-w
https://sgrt.org/video-library/using-sg ... sitioning/
Finally, the video made here in TAUH, with ROI+postural video based setup in extremity with optimal model (time point 18 min.).
https://sgrt.org/video-library/benefit- ... tremities/
Hope there was something useful.
Best regards Marko
-
