目的：本實驗主要利用筆射束卷積演算法（Pencil Beam Convolution, PBC）和非均向解析演算法 （Analytical Anisotropic Algorithm, AAA）計算頭頸癌病人調強放射治療之皮膚劑量，加上假體 測量後評估頭頸癌病人在治療計畫中的皮膚劑量計算值的可信度。 材料與方法：本實驗採用薄片型熱發光劑量計以及輻射變色軟片來做為量測工具，在IMRT head/ neck phantom 和 anthropomorphic Rando phantom 測量模擬的頸部淋巴區的表面劑量，利用 Eclipse version10.0 內兩種不同的的治療計畫系統演算法（PBC 和AAA）製作處方劑量為 180 cGy 的 7 個照野治療計畫，進行此兩種量測工具絕對劑量之測量。 結果：本實驗之薄片型 TLD 變異係數在 3% 以內，計讀值與給予劑量間呈現良好線性關係； 輻射變色軟片校正區線上每個點之變異係數在 1% 以內；計算隔點間距設定為 2.5 mm 時， IMRT head/neck phantom 右側表面劑量的計算值與 EBT2 film 與薄片型 TLD 的量測平均值之誤 差，PBC 和 AAA 為 13.02%、10.73% 和 -4.54%、-6.48% 以及左側為 -12.52%、-10.06% 和 -4.58%、-1.9%；anthropomorphic Rando® phantom 右側表面劑量的計算值與 EBT2 film 與薄 片型 TLD 的量測平均值之誤差，PBC 和 AAA 為 -0.99%、 -2.26% 和 0.64%、-0.64% 以及左側為 -7.68%、-12.47% 和 2.9%、-2.44%；當計算隔點間距設定為 2.5 mm 時，AAA 有較穩定的計算 值。 結論與討論：由本實驗得知，兩者演算法雖然皆無法準確計算表面劑量，但是 AAA 仍然比較接近 量測值而且有較穩定的數值，未來本研究將找出更準確的演算法並對輪廓做修正後找出直接利用 治療計畫系統就可以推估皮膚劑量的方式，以確實找出造成病人皮膚反應的劑量評估值。

Purpose : Pencil Beam Convolution (PBC) and Analytical Anisotropic Algorithm (AAA) were used to calculate the surface dose of simulated phantoms for head-and-neck patients treated with intensitymodulated radiation therapy. These were also used to evaluate the reliability of calculated results for the skin dose in vivo head-and-neck patients. Materials and Methods : This experiment adopted 0.1 mm ultrathin thermoluminescent dosimeters (ultrathin TLD) and ISP Gafchromic EBT2 film as measurement tools. The surface doses at hypothetical neck lymphatic regions were measured. Seven-field IMRT plans with a prescribed dose of 180 cGy were designed using an Eclipse (version10.0) treatment planning system where both PBC and AAA were implemented. These two different measurement tools were used to measure the absolute dose when calculated algorithms used different grid sizes. Result : In this experiment the variance coefficient of ultrathin TLD was within 3%. The reading values and delivered doses were in a good linear relationship. The points on calibration curve of Gafchromic film were within 1%. Using grid size of 2.5 mm, the errors between calculated values (PBC and AAA) and the mean surface measured doses for EBT2 film and ultrathin TLD were 13.02%, 10.73% and -4.54%, -6.48% on right side and -12.52%, -10.06% and -4.58%, -1.9% on left side neck of IMRT head/neck phantom, respectively. The errors between calculated values (PBC and AAA) and the mean surface measured doses for EBT2 film and ultrathin TLD were -0.99%, -2.26% and 0.64%, -0.64% on right side and -7.68%, -12.47% and 2.9%, -2.44% on left side neck of anthropomorphic Rando® phantom, respectively. Thus, AAA had more stable calculated results and performance than PBC using a grid size of 2.5 mm. Conclusion and Discussion : From the experimental results, neither algorithm was precise in calculating the surface dose. However, AAA was closer to the measured dose and its performance was stable. Future research will evaluate the most precise algorithm and modify the contours of the body to evaluate the skin dose. The treatment planning system can then be used directly to calculate patients’ skin dose indeed affecting the skin reaction.