Fig. 6
a: Distributions of paired comparisons of inter-versus intra-fractional anatomic deviation of the respective parameter from the same dose fraction over all 91 dose fractions. a, the first parameter is the length of the anatomic shift of the Dmax point in CBW of the subsequent CT as determined by elastic deformation from the subsequent to the source CT. The respective shift of Dmax in CBW on CBCT2 by the adaptive plan to CBCT1 was recorded as intra-fraction motion. b, the second parameter is||CBCTree||, the length of the rigid structure based coregistration vector of the central bronchial tree contoured on subsequent and source CT after online match. The medians of the inter-fractional deviations were larger than of the intra-fractional deviations for both parameters (p < 0.0001, signed rank test). b: The isocenter of the scheduled plans according to method S1 (green curve, open green circles), S2 (blue curve, closed blue diamonds), and S3 (red curve, red closed triangles) was systematically shifted in± y (a) or± x (b) direction of the Dicom coordinate system in CTplan. Data were fitted by 6th degree polynomials. The 95% confidence bands for the expected predicted values were filled with the respective semi opaque colour. Dmax increases in CBW according to the length of the shift were largest for plans according to method S1, followed by S2, and S3.c: Relation between Dmax on subsequent CT and the Hausdorff distance between the Dmax point in CBW on the subsequent CT backdeformed to the source CT by elastic deformation and the volume that shows sensitivity to water override in the trachea (V10CBALfrom the difference plan in CBAL of the source CT with or without water override) Blue circles indicate data points from the clinically used scheduled plans, red triangles from the adaptive plans