Fig. 1

Overview of scanning sessions. For each volunteer, three consecutive 4DMRI scans were obtained during a session, 4DMR1, 4DMR2 and 4DMR3 (sessions always on separate days). All volunteers had at least two sessions, while Volunteer 1 and Volunteer 2 had ten sessions. Respiratory induced motion estimates were generated from the 1D navigator signal of the right hemidiaphragm of a reference MRI for the outlier rejection strategies NoOR, MeanIE and Min95 (see text for details). a For the planning MRI strategy the first obtained scan was used as a reference (mimicking a planning MRI), indicated by the orange box. The motion estimates were evaluated for all subsequent scans (indicated by the grey boxes, simulating treatment fractions). b For the daily MRI strategy the first scan of each session (indicated by the purple box) was used as a reference and the motion estimates were evaluated on the two following scans of the same day (indicated by the grey boxes, simulating treatment fractions)