目的：本研究主要目的為前列腺癌接受放射線治療的病患以錐狀射束電腦斷層（CBCT）來量測 擺位的系統及隨機誤差，放射線治療包含有根治性及術後輔助性放射線治療。次要目的為評估 是否有臨床相關因子會影響擺位誤差的大小，並計算若沒有使用 CBCT 影像導引時的最適宜計 畫靶體積的範圍。

材料與方法：包含根治性及術後輔助性放射線治療，自西元 2011 至 2015 年本研究共納入 59位 前列腺癌症患者；每位患者至少接受五次以上的 CBCT 影像導引校正。病患在執行放射線治療 前先接受 CBCT 掃瞄，並由臨床醫師判讀擺位誤差之位移數據。我們將誤差位移軸向定義為： X 軸（左右側）、Y 軸（頭腳側）及 Z 軸（胸背側）。根據各軸向數據來計算擺位之平均誤差、 系統性誤差（Σ）及隨機誤差（σ），並以變異數分析（ANOVA）及 T 檢定評估各臨床相關因子 與擺位誤差之間的關聯性。無影像導引之最佳計畫靶體積的增加距離則藉由 Van Herk 公式計算。

結果：本研究共分析 1595 組影像導引之影像。三軸的擺位平均誤差、系統性誤差（Σ）及隨機 誤差（σ）分別為：X 軸0.37 mm、1.92 mm、2.50 mm；Y 軸0.29 mm、1.33 mm、2.00 mm；及 Z 軸 0.58 mm、1.59 mm、2.45 mm。臨床相關因子的各中值數為年齡 74 歲、身高 165 公分、體 重 69.1 公斤、身體質量指數（BMI）25.5 kg/m2 、前列腺體積 45.9 ml。其他因子有臨床或病理 分期、治療目的及直腸擴張等。經分析發現直腸擴張與否會顯著影響Z軸方向之隨機誤差（p = 0.013）。 X、Y、Z三軸之最佳計畫靶體積範圍則分別為 6.6 mm、4.7 mm及5.7 mm。

結論：本研究之三軸向的系統及隨機誤差與先前文獻比較，有類似之結果，且電腦斷層定位時 直腸的擴張與否會影響胸背側的擺位誤差。在未執行 CBCT 影像導引放射線治療的情況下，最 佳的計畫靶體積範圍至少為 4.7至6.6 mm。 [ 放射治療與腫瘤學2016; 23(3): 145-152 ]

Purpose : The primary goal of this study is to measure the systemic and random errors in prostatic cancer patients receiving definitive and adjuvant radiotherapy with cone beam computerized tomography (CBCT). The secondary aim is to evaluate the clinical factors that may affect these setup errors, and to determine the optimal planning target volume (PTV) margins for patients without using CBCT.

Materials & Methods : From 2011 to 2015, we retrospectively reviewed 59 patients who underwent at least five sessions of definitive or adjuvant radiotherapy with CBCT. All the images were obtained prior to treatment. All the setup corrections were determined by physicians. The definitions of axis was X in left-right, Y in superior-inferior and Z in anterior-posterior. The overall mean error, systemic error (Σ) and random error (σ) were calculated for each axis. We used ANOVA and Student t-test models to evaluate the significance among various clinical factors and the setup errors. The optimal PTV margins without image-guided radiotherapy (IGRT) were calculated based on Van Herk Formula.

Results : In total, 1595 sets of images were obtained from this study. The mean error (mm), Σ (mm) and σ (mm) were 0.37, 1.92, 2.50 in X; 0.29, 1.33, 2.00 in Y; and 0.58, 1.59, 2.45 in Z axis, respectively. The median value of various clinical factors were 74 years of age, 165 cm in height, 69.1 kg in weight, BMI of 25.5 kg/m2 and 45.9 ml of prostate volume. Other clinical factors such as clinical / pathological stage, definitive or adjuvant treatment and distension of rectum were also analyzed. The rectal distension affected the random error in direction Z significantly (p= 0.013). The optimal PTV margins without CBCT were 6.6, 4.7 and 5.7 mm in directions X, Y and Z, respectively.

Conclusion : The dimensions of systemic and random errors in various axes were similar to previous studies. Rectal distension can influence anterior-posterior setup errors. The optimal PTV margins should be at least 4.7 ~ 6.6 mm to overcome the daily setup uncertainties without CBCT corrections.