68-4

62/136

times for the FIF plans are presented in Figure 6f. In manual FIF and semiautomatic FIF plans, V105% decreased by almost 100% compared with the orig-inal plans, as shown in Figure 6a. This indicates that the hotspot region above Dth almost disap-peared when using the semiautomatic FIF tech-nique. V95% values on average (± one standard deviation) for the original, manual FIF, and semiau-tomatic FIF plans were 99.4 ± 0.3%, 99.3 ± 0.3%, and 99.4 ± 0.3%, respectively (Figure 6b). Similarly, D95% values on average for the original, manual FIF, and semiautomatic FIF plans were 99.6 ± 0.5%, 99.2 ± 0.6%, and 99.4 ± 0.5%, respectively (Figure 6c). Dmax values on average for the original, manual FIF, and semiautomatic FIF plans were 107.3 ± 1.1%, 105.0 ± 0.1%, and 105.1 ± 0.1%, respectively (Figure 6d). The HIs values on average for the original, manual FIF, and semiautomatic FIF plans were 8.2 ± 0.8%, 6.6 ± 0.8%, and 6.6 ± 0.7%, respec-tively (Figure 6e). The paired Wilcoxon signed-rank test indicated 382Figure 7　Dose distributions of the original, manual field-in-field (FIF), and semiautomatic FIF plans for case no. 22. (a) Field shape and dose distributions of the original plan. The upper, middle, and lower panels show the field shape, 3D dose distribution in the beam’s eye view, and 2D dose distribution in a slice, respectively. The slice position is shown in the middle panel with a white dotted line. (b) Same as (a) but for the manual FIF plan. (c) Same as (a) but for the FIF plan in Step 1 (FIF). In the 2D dose distributions, the yellow lines are the 100% isodose lines, and the pink lines are the 105% isodose lines. The doses are relative to the prescribed dose.that there were significant differences in V105%, V95%, D95%, and Dmax in the original and manual FIF plans compared with those in the semiautomatic FIF plans. There was also significant difference in HI in the original plans compared with that in the semiautomatic FIF plans.The semiautomatic FIF planning times, which were defined as the processing times by the auto-matic FIF script to generate the sub-beams and dose calculations, were 25-29 s for FIF (Step 1) and 41 s for FIF-4SF (Step 2). The average time including FIF and FIF-4SF was 28 ± 4 s. The processing time for the semiautomatic FIF plan technique significantly decreased by 207 ± 84 s compared with that for the manual FIF plans (Figure 6f).Comparison of dose distributionFigure 7 shows a typical example of the dose distributions in the (a) original, (b) manual FIF, and (c) semiautomatic FIF plans for the case in