68-4

60/136

380Application of the automated FIF algorithmPatientsWe used the planning data of 22 patients with brain metastases from lung, breast, and rectal cancers treated with WBI from January 2016 to December 2017 at Showa University Yokohama Northern Hospital. The study protocol was approved by the institutional review board (##17HO92) of our hospital, and the need for informed consent was waived. This study was conducted in accor-dance with the Declaration of Helsinki and Japa-nese ethical guidelines for epidemiologic research.The clinical treatment plans were used as the original plan for the automated FIF algorithm. All plans used two laterally opposing fields with 10 MV of X-ray energy and prescribed a total dose of 30 Gy (2 Gy per fraction) or 31.2 Gy (1.2 Gy per frac-tion delivered twice daily) at the isocenter. The dose calculation algorithm used for all plans was an analytical anisotropic algorithm15) in the TPS (Eclipse, version 13.7.14; Varian Medical Systems, USA). The PTV included the cerebral parenchyma, with a margin of 2 or 3 mm. The original plans were created for a linac (TrueBeam STx; Varian Medical Systems, USA) with the MLC containing 14 pairs of leaves with a 2.5-mm width and 32 pairs of leaves with a 5-mm width. The dose grid size was 2.5×2.5×2.5 mm3 in the TPS.Figure 5　Beam’s eye views (BEVs) with the multileaf collimator shapes of the sub-beams in the four-sub-beam field-in-field plan. (a) The BEV of one of the first sub-beams. The high-hotspot region of 106.4 % of the prescribed dose in the original plan is represented by Hotspot 1. (b) The BEV of one of the second sub-beams. The low-hotspot region of 105 % of the prescribed dose after the first step is indicated as Hotspot 2.Creation of a manual FIF plan for comparison with automatic FIF We also manually created FIF plans from the original plans by a board-certified physicist . Two sub-beams were created by copying the two main beams. Hotspot regions in the original plan were blocked using MLC leaves in both sides of the sub-beams. The hotspot regions were visually identified in the BEV. The same threshold for hotspot was taken as in the automatic FIF case. The same weights were given to the sub-beams to remove the hotspot. The procedures of making subfields and adjusting beam weights are almost the same as for the automatic FIF case.Dose index evaluationV105%, V95%, D95%, Dmax, and the homogeneity index (HI) of the PTV in the semiautomatic FIF plans were compared with those in the original and manual FIF plans. V105% was used as an index to indicate whether the hotspot decreased using the FIF technique. V95% and D95% were used as indica-tors of dose coverage. Dmax was used as an indi-cator of excess dose. The dose and volume were expressed as the relative dose to the prescribed dose and relative volume to the PTV, respectively. DV was the dose to a specified fractional volume V in the PTV. VD was the volume that received at least the dose D in the PTV. The HI is defined as