68-6

30/104

1234567812345678594Detected Jaw PositionsVerification of detection accuracy of jaw position in the static-jaw BPF planTable 1 summarizes the jaw positions in the static jaw position BPF plan. The planned X1 and X2 jaw positions were −90.0 mm and +90.0 mm, respectively; the detected jaw positions at each segment (average ± 1 SD) of X1 and X2 were −90.06 ± 0.06 mm and +90.02 ± 0.13 mm, respec-tively. The detected jaw positions were in good agreement with the planned positions. This finding indicates that our method can accurately detect jaw positions under the MLC leaves (MLC trans-mission region) in the BPF plan.Detected jaw positions in the jaw-tracking BPF plansThe detected jaw positions in the jaw-tracking BPF plans with gantry angles of 0, 90, and 270 degrees, and 360-degree rotation are shown in SegmentPlanned−90.00−90.00−90.00−90.00−90.00−90.00−90.00−90.00Diff: Planned-Detected; SD: Standard deviation.Gantry 0 degreesSegmentPlannedDetected−74.00−73.78−54.00−54.06−34.00−33.81−14.00−13.81−16.00−16.06−36.00−35.93−56.00−56.13−76.00−76.02Table 1　Summary of detected and planned jaw positions in static jaw plansTable 2　Planned and detected jaw positions of X1 jaw (mm)Gantry 270 degreesX1 jaw position (mm)Detected−90.06−90.13−89.93−90.07−90.06−90.14−90.05−90.03SD0.050.030.090.040.220.340.140.05DiffSD0.24−0.220.120.060.21−0.190.27−0.190.050.060.17−0.070.190.130.020.12Detected−73.92−54.10−33.69−13.99−15.73−36.18−55.70−76.18Diff0.060.13−0.070.070.060.140.050.03Planned90.0090.0090.0090.0090.0090.0090.0090.00Gantry 90 degreesDiffSD0.10−0.080.130.100.24−0.310.19−0.010.29−0.270.230.180.28−0.300.180.03Diff: Planned-Detected; SD: Standard deviation; X1: left-sided jaw.Detected−73.99−54.13−34.11−13.96−16.14−36.29−55.99−76.05Table 2 (X1) and Table 3 (X2). For the X1 (X2) jaw, the absolute differences between detected and expected jaw positions (average ± 1 SD) were 0.12 ± 0.17 mm (0.12 ± 0.15), 0.18 ± 0.19 mm (0.12 ± 0.17), 0.10 ± 0.21 mm (0.06 ± 0.11), and 0.16 ± 0.21 mm (0.14 ± 0.22) with the gantry angles of 0, 90, and 270 degrees, and 360-degree rotation, respectively.Detected jaw positions in the jaw-tracking BPF plans with artificial errorsTable 4 shows the detected jaw positions in the BPF plans with artificial errors of 0.5 mm or 1.0 mm. The absolute differences between detected and expected jaw positions including artificial errors (average ± 1 SD) were 0.14 ± 0.10 mm for the X1 jaw and 0.13 ± 0.15 mm for the X2 jaw. This result shows that our method can detect a jaw positioning error of 0.5 mm with an accuracy of 0.3 mm.Picket Fence Test Using Accumulated Cine ImagesFigure 4 shows the results of analysis for the X2 jaw position (mm)Detected89.9789.9590.0390.1190.2690.0889.8989.89DiffSD0.24−0.010.130.170.170.110.22−0.040.140.160.290.190.3−0.010.050.19Detected−73.71−54.10−34.27−13.83−15.91−36.19−56.04−75.87SD0.040.040.040.050.040.050.050.06Diff0.030.05−0.03−0.11−0.26−0.080.110.11Gantry rotationDiffSD0.33−0.290.100.10.130.270.13−0.170.13−0.090.370.190.140.040.36−0.13Results