Part I | Chapter 3 56 Results The mean (standard deviation [SD]) time between MR scans was 5.7 (± 1.4) min and 20.0 (± 5.2) min, for short interval (n = 10) and long interval (n = 50) respectively. Stratified by none/few minor (score 1-2) adaptations or multiple minor/major (score 3-4) adaptations needed, agreement between Observer 1 and observer 2 was 98% (59/60), 95% (57/60), and 85% (51/60) for CTV, bladder, and rectum contour, respectively (Table 1). For the long interval, it was estimated that 46/50 (92%) and 47/50 (94%) cases (Observer 1 and 2, respectively) could be edited within 3 min. Both observers estimated that this would be possible for all (100%) short interval cases. The remaining long interval cases showed larger intrafraction rectal deformations, in addition to a need for (minor) adjustment of the CTV and/or bladder contour. Adaptions needed Number of fractions (%) CTV Bladder Rectum Short (n = 10) Obs 1 Obs 2 Obs 1 Obs 2 Obs 1 Obs 2 None (1) 8 (80) 8 (80) 10 (100) 9 (90) 4 (40) 8 (80) Few minor (2) 2 (20) 2 (20) 0 (0) 1 (10) 5 (50) 1 (10) Multiple minor/few major (3) 0 (0) 0 (0) 0 (0) 0 (0) 1 (10) 1 (10) Multiple major (4) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) Long (n = 50) None (1) 28 (56) 41 (82) 30 (60) 39 (78) 13 (26) 26 (52) Few minor (2) 21 (42) 9 (18) 18 (34) 9 (18) 24 (48) 18 (36) Multiple minor/few major (3) 1 (2) 0 (0) 1 (3) 2 (4) 8 (16) 4 (8) Multiple major (4) 0 (0) 0 (0) 1 (3) 0 (0) 5 (10) 2 (4) Discussion We have explored and demonstrated the clinical usability of intrafraction propagated contours provided by a DIR algorithm for MRI-guided prostate cancer radiotherapy treatment. Contours should be generated in a quick and accurate manner, to minimise operator interaction and to maximise the potential benefits adaptive workflows can offer when delivering large fractional doses. Our results suggest that intrafraction contours provided by EVolution were in general directly acceptable (CTV and bladder) or mostly needed only minor manual editing (rectum). Although Table 1 – Need for adaptations of propagated contours as scored per observer, stratified by ‘short’ and ‘long’ interval between MRI scans. Legend: CTV = Clinical Target Volume. Obs = Observer. The corresponding scores as provided by the observers (see Figure S2) are presented within brackets.
RkJQdWJsaXNoZXIy MTk4NDMw