Nice talking with you. I understand and agree your point completely in the problematic cases as you mentioned and may be at one fraction or so. Substantial variation in the heart position is reported in the DIBH (Heart position variability during voluntary moderate deep inspiration breath-hold radiotherapy for breast cancer determined by repeated CBCT scans, 2017) and partial DIBH is better than FB (How important is a reproducible breath hold for deep inspiration breath hold? 2015). In commonly to this forum: does these knowledges give us a permission not to take care of correct BHL? There are studies where statistically significant heart dose changes is noticed if BHL remains systematically partial (Cardiac dosimetric evaluation of deep inspiration breat-hold level variances using computed tomography scans from deformable image registration displacement vectors, 2015). These systematic changes in BHL is what we want to pay attention to. I do not know many hospitals that are systematically verifying the BHL at the beginning of the treatment (with measuring the errors of the distances between vertebra and sternum in VRT and LNG) from the LAT- kV images and Ellen mentioned that they do, and I was excited and pleased.
Based on RTTs daily work with SGRT, I am not for sure that is it in all the problematic situations the patient which can not cope with what we want them to do in DIBH. I think there is still something for us users to learn about how patient should repeat DIBH daily similarly with the SGRT-quided, and not at all the situations trust blindly to the workflow we have created with operating with RTDs, ROIs etc.
I would also like to go towards easy and commonly used IGRT protocol in DIBH where we acquire mostly tangential images only (which are indeed in good accuracy already) or even weekly IGRT at some point, but I think it is not possible before we are completely sure and can trust that based on SGRT- setup all the bony structures, and thereby OARs, inside or near PTV are systematically inside margins used for whole PTV, in the individual- and population basis. Is one workflow with SGRT better than the other, we do not know yet, but the goal with this system is the systematic reproducibility of the bony structures, breast and isocenter in the setup images (CBCT, orthogonal and tangential) during the whole treatment session, hopefully with as small effort as possible. It needs offline investigation of the images at the beginning of the treatment of individual patient and finally investigation of the accuracy in larger population to estimate that is there something to improve in the system or workflow.