Figure 1 (click to view)

In Fig. 1, the lateral (a), ventrodorsal (b) and craniocaudal (c) set-up errors of nine patients treated with the Bellinzona technique are shown. On average portal images were collected from 10 irradiation fractions for each patient, which resulted in 93 sets of portal images from the AP and lateral verification beams and the left and right anterior oblique radiation beams. The set-up errors determined from the left and right anterior oblique beams were plotted as a function of those determined from the AP and lateral verification beams (circles). The lateral set-up errors showed a wide spread around the correspondence line. However, the set-up errors in the ventrodorsal and the craniocaudal direction showed a good correlation between the results from the oblique and the orthogonal beams. An overview of the Pearson correlation coefficients and the mean difference between the set-up errors from the oblique and orthogonal beams is given in the upper part of Table 2.

Table 2 (click to view)

The gantry angles of the left and right anterior beams for the Bellinzona technique were about 70° and 290° for most patients. Therefore, it can be expected that the inaccuracy in the lateral set-up error determined from these oblique beams will be relatively large. Therefore, we investigated the effect of adding an AP verification beam to the left and right anterior oblique beams. The resulting set-up errors are represented by plus signs in Fig. 1. The addition of the AP beam had a large effect on the spread of the scatter plot for the lateral set-up error. In the ventrodorsal direction there was no influence at all and in the craniocaudal direction there was little influence. An overview of the Pearson correlation coefficient and the mean difference between the set-up errors from the combination of the oblique irradiation plus the AP beam and the verification beams is given in the lower part of Table 2.