電腦斷層影像在不同形狀假體的重組體積干擾因素之分析

Analysis of Confounding Factors in Volume Reconstruction of Different Shaped Phantoms with Multi-slices Spiral CT Simulation

10.6717/JTMRT.201609_4(2).0004

宋國強(Kuo-Chiang Sung)；李財福(Tsair-Fwu Lee)；陳怡任(I-Jen Chen)；阮國榮(Kuo-Jung Juan)；吳鴻源(Horng-Yuan Wu)

多層螺旋電腦斷層掃描儀 ； 切片重疊因子 ； 切面厚度 ； Hounsfield unit ； Multi-slices computed tomography ； pitch: slice thickness ； Hounsfield unit

臺灣醫事放射期刊

4卷2期（2016 / 09 / 01）

27 - 34

繁體中文

這篇研究主要是利用多層螺旋電腦斷層掃描儀(Multi-slice spiral computed tomography)來評估不同假體間造成重組體積誤差的因素，我們將不同形狀假體在電腦斷層掃瞄條件下逐一進行掃描，並將得到的影像資訊傳輸至治療計劃系統，利用治療計劃系統的影像重組功能重建假體之體積，將重組織體積與假體實際之體積比較其誤差值，並探討哪些因素會影響體積重組的誤差。在實驗中我們使用的電腦斷層掃描儀為GE LIGHTSPEED RT16，使用的治療計劃系統為 Philip 公司所生產的Pinnacle3® (version 8.0m, Philips, Fitchburg, WI).。我們選擇了球體、圓柱體、正方體和四角錐等四種不同形狀大小的假體，用以模擬人體不同器官的掃描情況，假體的材質為壓克力內含甲基炳烯酸甲酯單體 99.765%、硬化劑(ABIN)0.1%、顏色染料0.135%，比重為1.19，而掃描條件的設定上有四種不同的切面厚度：1.25 mm、2.5 mm、3.75 mm、5 mm，而三種不同的切片重疊因子分別為0.562、0.938、1.75，重組影像設定的Hounsfield unit(H.U)為-400、-300、-200、-100、0。並利用統計方法討探誤差是否有意義。從研究結果我們可以發現會影響重組體積的掃描條件為切面厚度、掃描物體的體積大小、形狀和重組體積的H.U 值，切片重疊因子的不同對於重組體積的影響並不明顯，我們發現在治療計劃系統裡，圓柱體的假體在重組的誤差明顯比其他形狀的假體來得小，而體積較大的假體其重組體積的誤差皆小於體積小的假體，如果在小的假體，使用較大的切面厚度為5mm，和不當的H.U 值，體積誤差甚至可以高達30%。本研究找出了影響重組體積的因素，在定義腫瘤的大小和評估危及器官的劑量上都提供了很大的幫助，以利日後在臨床上使用。

Delineating precisely during the treatment planning procedure can improve the tumor control probability and reduce the normal tissue complication probability. Important issues have been raised concerning for the accuracy of volume reconstruction in computed tomography (CT) image for treatment planning. Different shaped phantoms, multi-slices spiral CT, and treatment planning systems (TPS) are used to analyze the confounding factors in volume reconstruction. In this study, four different shapes of phantoms were used: sphere, cube, cylinder, and pyramid. Scanning thickness was set at 1.25mm, 2.5mm, 3.75mm, 5mm; the scanning pitch was set at 0.562, 0.938,1.75, and H.U was selected at -400, -300, -200, -100, 0 H.U for this experimental. Pinnacle3® TPS (version 8.0m, Philips, Fitchburg, WI) was used. We observed that the confounding factors to affect the volume reconstruction are consisted by the slice thickness, volume size, and H.U setting. There was no statistics significance in the pitch and effect of different TPSs. We found that the reconstruction error was kept within 5% for the larger phantoms. The reconstruction error became bigger for the smaller phantoms in all experiments. The accuracy of small phantom setting scanning thickness at 5mm can reach up to 30%. This study found out the confounding factors in volume reconstruction. It can be provided for the target volume and the organs at risk definition in the medical practice in the future.