- 學位論文
利用百萬伏特電腦斷層準確估計高密度金屬植入物之放射治療計畫劑量：假體與病患研究
Using MVCT to Accurately Estimate Radiotherapy Dose for High-Density Metal Prosthesis: Phantom and Patient Studies
摘要
千伏特電腦斷層(kilo-voltage computed tomography, kVCT)是目前主要用在放射治療劑量計畫的影像工具，然而金屬植入物於kVCT影像上和散射假影有關並導致劑量計算的誤差。百萬伏特電腦斷層(mega-voltage computed tomography, MVCT)具有減少金屬散射假影和分辨不同金屬密度的優點，在金屬植入物中MVCT具有正確估算劑量的潛力取代kVCT。 本研究首先建立MVCT高密度金屬材料的密度對應表(image value-to-density table, IVDT)，其次，我們利用新的演算法(Acuros XB, AXB)和非均向解析演算法(Anisotropic Analytical Algorithm, AAA)做非均質修正的比較，所有的估測會在臨床使用上的不同金屬材料做假體實驗及病人實際劑量量測之驗證。假體研究包含簡單的幾何金屬置於固態水假體之上和cheese假體之內，以及金屬材料置於擬人假體表面和水箱內。劑量預測公式根據已知的金屬密度和射程距離產生而來，運用在假體劑量計算的研究，最後，於臨床患者中進行實際量測。 我們在治療計劃系統上建立了MVCT的線性密度對應表運用在劑量計算上。假體研究中，固態水假體在kVCT、MVCT_AAA和MVCT_AXB的平均劑量差異分別為36.7%、5.9%和0%；擬人假體研究在kVCT、MVCT_AAA和MVCT_AXB的平均劑量差異分別為0.9%、-0.5%和-0.3%；水箱假體在kVCT、MVCT_AAA和MVCT_AXB的平均劑量差異分別為3.5%、0.7%和-0.3%。在劑量預測公式上，量測劑量和預測之間的平均劑量差異為0.2%。病人量測中，有假牙的病人kVCT、MVCT_AAA和MVCT_AXB的平均劑量差異分別為4.4%、0.6%和0.1%；有人工髖關節的病人kVCT、MVCT_AAA和MVCT_AXB的平均劑量差異分別為分別為-19.4%、-7.5%和-6.7%。 在假體實驗和病人量測中，照野內若有大於2.835 g/cm3的不同較高密度的材料時，MVCT比kVCT能更準確預測劑量，此外，相較於只做密度修正，AXB對於組織有更好的非均質修正。MVCT可以替代kVCT，特別是結合AXB演算法，對金屬植入物患者在放射治療計劃中有更好的劑量估計。
並列摘要
Kilo-voltage computed tomography (kVCT) is the main imaging tool for radiotherapy planning work. However, metal prosthesis is associated with streak artifact and deviates dose calculation on kVCT. Mega-voltage CT (MVCT) has the advantages of reducing metal streak artifact and distinguishing the densities of different metals. MVCT has the potential to replace kVCT for accurate dose estimation of metal implants. In this study, we first established the image value-to-density table (IVDT) of MVCT for the high-density metal materials. Second, we used a novel algorithm, Acuros XB (AXB), compared with Analytical Anisotropic Algorithm (AAA) for heterogeneity correction. All these estimations were validated by phantom experiments and real patient measurements on different metallic materials in clinical use. The phantom studies included simple geometric metals on top of solid water phantom and in cheese phantom, as well as the metallic materials on the surface of Alderson radiation therapy phantom (ART phantom) and inside the water tank phantom. The dose prediction formulation based on the known metal density and distance of beam path was generated to calculate the dose of phantom studies. Lastly, real patient measurements were conducted in clinical patients. We established the linearity of IVDT for MVCT in our treatment planning system (TPS) for dose calculation. In solid water phantom study, the average dose difference was 36.7% by kVCT, 5.9% by MVCT_AAA, and 0% by MVCT_AXB, respectively. In ART phantom study, the average dose difference was 0.9% by kVCT, -0.5% by MVCT_AAA, and -0.3% by MVCT_AXB, respectively. In water tank phantom study, the average dose difference was 3.5% by kVCT, 0.7% by MVCT_AAA, and -0.3% by MVCT_AXB, respectively. A dose prediction formula was generated with the average dose difference between measured and predicted doses of 0.2%. In patient measurements, the average dose differences were 4.4% by kVCT, 0.6% by MVCT_AAA, and 0.1% by MVCT_AXB, respectively, for the patients with dental fillings, and -19.4% by kVCT, -7.5% by MVCT_AAA, and -6.7% by MVCT_AXB, respectively, for patients with hip prosthesis. MVCT estimated the doses more accurately than kVCT in both phantom study and patient measurement for different higher-density materials of density >2.835 g/cm3 within the radiation field. Besides, AXB is better for heterogeneous correction of tissue than metal density correction alone. MVCT may be able to replace kVCT, especially in combination of AXB algorithm, with better dose estimation in planning radiotherapy for patients with metal prosthesis.