目的：影像導引已是放射治療中舉足輕重的技術，要確保影像導引結果準確，詳實的品質保證 （Quality Assurance, QA）作業不可缺少，故參考美國醫學物理學會（American Association of Physicists in Medicine, AAPM）142 號工作報告建議，進行本科每日品質保證作業。本研究分析 相關結果及效益。

材料與方法：直線加速器與影像導引系統為Varian Trilogy On Board Imager（OBI）系統，品 保使用假體為Isocenter Cube Assy（Varian TM 55150000 isocenter Cube tool）。由同一位醫 事放射師執行40 次影像導引系統每日品保作業，項目有碰撞連鎖（Collision interlocks）、影 像與治療中心點重合（Imaging and treatment coordinate coincidence）與治療床位移準確性 （Positioning）品保作業。針對治療床位移準確性，本研究執行方法一為判讀方格紙上位移的距 離（方格紙目視檢測）、方法二為影像導引系統拍攝正交片判讀小鐵球（Ball Bullet, BB）中心點差 異（BB 影像分析檢測）、方法三為觀察BB 與雷射交會情形（BB 目視檢測）的三種品保作業方式 探討其差異，並紀錄花費的時間。

結果：一、碰撞連鎖測試全數通過。二、影像與治療中心點重合品保結果皆為容許誤差之內。 三、治療床位移準確性三種方法結果皆為容許誤差之內。比較治療床位移準確性品保作業之方格 紙目視與 BB 影像分析檢測的X、Y、Z 軸與三軸總位移誤差沒有統計上顯著差異，P 值分別是 0.229、1、0.867 與0.196。比較品保治療床位移準確性三種方法之時間花費平均值，BB 影像分 析比方格紙目視檢測多0.4 秒，方格紙目視比 BB 目視檢測多出17.6 秒。方格紙目視與BB 影像 分析檢測沒有顯著差異（P ＞ 0.05）；方格紙目視與 BB 目視檢測有顯著差異（P = 0.00012）； BB 影像分析與 BB 目視檢測有顯著差異（P = 0.00006）。

結論：本研究試作之品質保證作業可用於OBI 與電子影像驗證裝備（Electronic Portal Imaging Device, EPID），品保流程不繁複，建議放射腫瘤科部門應建立影像導引系統品質保證作業，確 保病人治療的準確性與安全性。

Purpose : Image-guided radiation therapy is an important irradiating technique. To ensure the accuracy of image guidance, it is essential to conduct daily quality assurance (QA). According to Report No. 142 of the American Association of Physicists in Medicine, we developed daily quality assurance for a two-dimensional image-guided system. This study reports our results and benefit analysis.

Materials and Methods : Our image-guided linear accelerator was Varian Trilogy with On-Board Imager (OBI) system. The QA phantom was Isocenter Cube Assay (Varian TM55150000 isocenter Cube tool).

A qualified radiation technologist executed 40 daily QAs for our two-dimensional image-guided system. QA items included collision interlocks, imaging and treatment coordinate coincidence and positioning tests. To ensure the correct position of the treatment couch, we performed three QA methods. First, we interpreted the displacement on a graph paper visually. Second, we identified the displacement of Ball Bullet (BB) on the center of the image-guided system by using orthogonal films (BB image analysis and detection). Third, we calculated the difference between BB and the intersection of laser lines (visual inspection of BB). We compared these three QA methods, including the time cost.

Result : All collision interlocks tests had positive results. All imaging and treatment coordinates and positioning test results were within the range of allowable errors. Comparing the first two methods of position testing, no statistically significant difference was detected between them (P values in X-Y-Z three axes and total error, 0.229, 1, 0.867 and 0.196, respectively).

cComparing the mean time taken for couch positioning among the three QA methods, the BB image analysis and detection took 0.4 seconds longer than the visual inspection of graph paper. The visual inspection of graph paper took 17.6 second longer than the visual inspection of BB. There was no statistically significant difference between the visual inspection of graph paper and BB image analysis and detection (P> 0.05). However, statistically significant differences were detected both (1) between the visual inspection of graph paper and the visual inspection of BB (P = 0.00012), and (2) between the BB image analysis and the visual inspection of BB (P = 0.00006).

Conclusion : Our QA methods ensure the quality of two-dimensional image-guided radiation therapy in both Electronic Portal Imaging Device (EPID) and OBI systems. The QA procedures are easy to implement. To ensure the accuracy of radiotherapy, establishing QA procedures for image-guided systems is strongly recommended.