使用錐形電腦斷層掃描於肺癌弧形調控放射治療之治療計畫與劑量分布再驗證

研究目的: 現代放射治療技術日新月異，在如肺癌等細胞特性接近晚期反應組織，為了更有效保護腫瘤周圍的早期反應組織並降低等待復氧的時間，現今提出立體定位放射治療(Stereotactic Body Radiotherapy, SBRT)，使用寡分次高劑量的手法促使腫瘤細胞走向凋亡。基於對病患副作用的在意，現代放射治療領域對於劑量分布的要求越來越精密，甚至提出適應性計劃(Adaptive Radiotherapy, ART)以符合每日患者的身體變化。若要同時達到劑量合理抑低及定位影像的即時性，錐形射束電腦斷層(Cone Beam Computed Tomography, CBCT)比起傳統模擬定位使用的多偵檢器電腦斷層(Multiple detector computer tomography, MDCT)更為合適。若要得到兩者的差異，現今常規的點或平面劑量驗證無法滿足其複雜的劑量給予，若要得到三維的劑量分布，凝膠劑量計便是一大利器。 目前凝膠劑量計在計讀上使用工具有：磁振造影、光學電腦斷層、X光電腦斷層等，前兩者在現今放射腫瘤科內存在甚少，取得不易，甚至可能在移動過程中增加凝膠劑量計融化等變因。若要建立一個常規的劑量驗證法，X光電腦斷層是最為合適的方式，不光是取得影像的速度優於其他方式，且在放射腫瘤科是容易取得的影像工具。而不同的參數可能導致電腦斷層影像在雜訊、CT值的轉換有一定的誤差影響。本研究目的為探討以MDCT與CBCT讀取劑量時適當的濾波處理是否為必要手段，並以使用不同的濾波器消除雜訊，應用於凝膠劑量計量測及比較其差異性。材料與方法: 使用MAGAT凝膠劑量計取得三維劑量分布，並且以多偵檢器電腦斷層(MDCT)與千伏錐形射束電腦斷層(kV-CBCT)為讀取工具，回歸出劑量反應曲線。但因X光電腦斷層之中的雜訊，容易導致影像誤判使劑量讀取出現差異，所以本研究將會利用清除背景雜訊、平均影像及濾波器等方式，降低或減少雜訊以加強影像辨識度。藉此試圖證明兩種電腦斷層經相關影像處理後是否可達到相當的影像判讀能力。 若MDCT與CBCT可經影像處理手法，達到相當的影像辨識度，即可合理推論:CBCT具有模擬劑量計算的潛力。實際由兩者的定位影像模擬肺癌立體定位放射治療計畫，並使用3%/3mm及5%/5mm的容許誤差在不同深度平面下進行加馬分析，得出MDCT與CBCT治療計畫的通過率，以及治療計劃與使用凝膠劑量計接受照射後於目標區劑量分布的通過率。加馬分析通過率可說明兩者之間的吻合程度，再者可利用標準化兩個計畫在凝膠劑量計的測量結果，得出實際照射時的劑量吻合度。研究結果: MDCT與CBCT在影像處理後，分別得出劑量反應曲線。MDCT 1.33 Gy/HU，R-square高達0.96。而CBCT在接收劑量超過5戈雷後，劑量反應為0.33Gy/Δ Gray scale，R-square約為0.96，兩者皆表現出良好線性度。濾波後劑量分辨率，以NLFM濾波器在MDCT中D_∆^(95%)=0.80戈雷，表現最為優異。CBCT的部分則是D_∆^(95%)=0.80戈雷的Adaptive Mean Filter最佳。 劑量計畫方面，以MDCT與CBCT為定位影像的治療計畫，在3%/3mm的容許誤差內，於不同深度可達80%的通過率，5%/5mm甚至高達90%已是臨床接受的通過率。實際照射結果分析，MDCT與CBCT的結果經標準化後，在3%/3mm與5%/5mm的容許誤差通過率皆達到90%，兩者表現具有一定的吻合程度。討論與結論: 凝膠劑量計雖然改善了化學劑量計容易擴散導致劑量分布產生誤差的缺點，但是這個穩定度是有時效性的，劑量讀取的變因可能會隨著時間、重力等因素而增加，再加上配置工具於一般醫院取得不易，進而無法成為常規劑量驗證法。然而，本研究中發現凝膠劑量計於三維劑量驗證具有一定的能力，加上其輻射靈敏的範圍恰巧為SBRT高劑量低分次的治療範圍，透過有效的降低雜訊方法(如：NLMF filter)將來可能可提供這些複雜的治療法一個更為實際劑量驗證方式。 也藉由凝膠劑量計驗證MDCT與CBCT皆可用於治療計畫模擬，但CBCT因成像原理導致的影像缺失，在模擬劑量時容易出現電子密度誤判，增加計畫製作的難度與所需時間。在相同的處方劑量之下，CBCT在低劑量區的表現，也會與MDCT計畫具有明顯出入。 縱然CBCT可於治療當下獲得影像完成適應性計劃，但其影像品質的低落及計畫完成所需的時間，仍是CBCT目前遇到最大的障礙。若能提升其影像品質降低物質誤判，搭配人工智慧縮短計畫時間，CBCT在治療計畫的潛力將會是患者的一大福音。

Research purposes: First, adaptive radiation therapy (ART) based on cone-beam computed tomography (CBCT) has emerged as a domain of dedicated cancer radiotherapy system. This study aims to evaluate the dose calculation using CBCT in ART treatment, as comparing to that using planning CT by polymer gel dosimeters. X-ray cone-beam computed tomography (CBCT) is most commonly used as a method of guidance of inter-fractional corrections before treatment in radiotherapy. Multi-detector CT (MDCT) combined with adaptive mean filtering has successfully contributed dose information from irradiated polymer gel dosimeters; however, CBCT has not yet been investigated. Therefore, this study aims to evaluate whether CBCT with different noise reduction filters is potential to achieve adequate dose resolutions in polymer gel dosimeters.Material and Methods: Each vial was filled with normoxic polymer gel and irradiated with uniform doses of 0-16Gy to generate dose response curves. After irradiation, CBCT was used to perform the dose measurement. In addition, image postprocessing using noise reduction filtering techniques have been examined. Moreover, a CIRS thorax phantom was used for the measurement of simulated treatment plans using both planning CT and CBCT. This phantom also was simulated a left-side lung cancer and heart, spine cord, and right-side lung, respectively, as organ-at-risk (OAR). Dose difference criterion /dose-to-agreement criteria of 3%/3 mm and 5%/5 mm were used for comparing the dose volumes calculated using the planning CT with those calculated using the CBCT.Results: Results indicated that dose response of slope was 1.33 Gy/ΔHU and 0.34 Gy/ΔGray scale for MDCT and CBCT in irradiated dose of 6-16Gy, respectively. Moreover, normalized least-mean fourth (NLMF) filter was performed the best dose resolution with 95% confidence of about 1 Gy, demonstrating a highly effective tool for noise reduction in CBCT gel dosimetry. Normoxic polymer gel dosimetry combined with CBCT and NLMF filtering provides a useful method for dose measurement in radiotherapy, especially in case of stereotactic body radiation therapy. In treatment plan system , the pass rate of the gamma evaluations with 3%/3mm is 80% for the treatment , which simulated by MDCT and CBCT at SAD =99.5 ,100 ,100.5 cm ,and the pass rate of the gamma evaluations with 5%/5mm even up to 90% , and it is clinically accepted.. Then, we analysis of actual irradiation to get the result. the results have to normalize to get the pass rate of MDCT and CBCT treatmeant. After the results are normalized, the pass rate of the gamma evaluations with f 3%/3mm and 5%/5mm both are over 90%, and the performance of the them has a certain degree of agreement.4. Discussion and conclusion Gel dosimeters cannot serve as a conventional dose-verification method because of the low availability of the required instruments in radiation oncology departments, and because of variables that affect the readout process, which are as follows. (1) Interval between radiation exposure and readout (2) Response dose and time after radiation exposure(3) Gravity (4) Beam dose rate during radiation exposure. However, in this study, it was found that the gel dosimeter has ability in three-dimensional dose verification, and its radiation sensitive range happens in SBRT. if can find an effective filter of reducing noise (eg NLMF filter). The gel dosimeter will become more useful in dose verification for these complex treatments in the future. It is also verified that both MDCT and CBCT can be used for treatment planning simulation by gel dosimeter. However, CBCT often miss image information, and it is easy to misjudge electron density when simulating dose, which increases the difficulty and time required for planning. Under the same dose, the performance of CBCT in the low-dose area will also be significantly different from the MDCT program. Even though CBCT can obtain an image to complete the adaptive plan for treatment, its low image quality and the time required for the planning is still the biggest problem in CBCT. If we can improve its image quality to reduce material misjudgment and advance the artificial intelligence (AI) to shorten the planning time. CBCT is a nice tool to popularize the adaptive radiotherapy.(ART)