Last week there was a patient case which fits well to this topic. FB breast treatment was the case. kV/kV imaging was berformed, no couch shifts were needed with DICOM surface setup, all the bony structures were perfectly aligned in the images. Then some errors appeared to machine software at the time treatment was supposed to begin. Patient was at the setup position during the delay time. It took about 7 minutes and we were ready to continue. New images were acquired before treatment. AlignRT DICOM RTDs were shifted during that delay time such that VRT showed -0.4cm, LNG was +0.4cm and PITCH was now +2. In the kV/kV images sternum was shifted 0.4cm towards vertebra and 0.4 cm caudally in comparison to vertebra, lower part of the sternum was ventrally than the upper vertebra. All the AlignRT values were then seen in the images in this case. Addition to those shoulder joint was shifted 1 cm caudally during that time. If we think it that way that this new position would be a position in DRR (acquired at the CT), one could not entirely copy that kind of patient position at the treatment setup no matter what. The good thing was that surface guidance device showed everything right what was happening for the patient position during that time and gave support for the knowledge that intra-fractional errors can be related to time and this may have effect to the inter-fractional error if this delay has happened at the CT.