as mentioned, for final positioning kV-imaging is (still) our gold standard. We do kV-imaging for every single fraction. So it is important to understand that we use AlighRT (as a substitute for skin marks) for initial positioning before kV-imaging. To find the proper position quickly it helps to have pictures of the patients from CT with the laser projected on the skin to indicate the location of the isocenter.
To convince ourselfs that AlignRT performs at least as good as skin marks (for initial positioning) I analysed performed delta shifts resulting from the kV-image match with refrence DRRs for about 700 fraction for patients we positioned initially on skin marks and 700 fractions for patients initially positioned with AlignRT only (to do so you can skill mark your patients, but don’t use the marks). The analysis of our sample shows comparable results, slightly better for AlingRT. As with skin marks, AlignRT performs not always in a perfect way and proper selection of ROIs is crucial. But we can conclude that AlignRT only based initial positioning before kV-imaging is as reliable as skin marks and outperforms skin marks often with difficult cases. Given the large benefit in confort for patients as well as for RTTs when omitting skin markings, we decided to go markerless. Once again, we don’t say, for the moment, that AlignRT can be used to find the final (treatment) position without kV-imaging. This will anyhow never be possible for internal moving targets like e.g. prostate. I guess it is understandable that one has to convince himself that going without skin marks is reliable and safe after we have used skin marks for decades. But it is worse to do so as the advantages, and the gain in time, when going markerless are really substantial.
In case you are using Aria, you can extract the delta shifts quite easily from Oflline Review. I will present some of the results at the upcoming SGRT users meeting in London.
I hope this helps a bit.