動態對比增強磁共振影像參數之估測 : 位移效應之修正

Abstract

動態對比增強磁共振影像(DCE-MRI)為一應用快速MRI掃瞄序列觀測身體內注射的對比劑進入血液中微灌流(perfusion)的情形。此應用對於觀察服用抗血管新增藥物病患的治療與評估有很顯著的成果。而對比劑進入人體血液循環系統到達腫瘤部位到代謝出來的行為模式往往可由不同的數學模型來描述。 目前有相當多的應用在於藉由量測這些數學模型的參數進而推得對比劑在循環系統或腫瘤組織附近的流動情形。由這些流動的難易程度亦可以非侵入方式估測腫瘤特性。 本研究的病人以肺癌病患為主，並施以Avastin抗癌藥物作為抗血管新增治療藥物。希望可以透過動態對比增強磁共振影像(DCE-MRI)的特性，早期預測並評估此抗癌藥物的化療效果。而本研究所採用的分析軟體Mistar可供我們選擇不同的數學模型來進一步評估治療的效果。 不過此軟體受限於本身功能的限制，對於肺部切面影像在掃瞄中的呼吸等自然位移與誤差未能進一步調整與修正。故本研究除了獲取動態對比增強磁共振影像(DCE-MRI)的數學模型參數外，更進一步提出影像前處理的方式來修正因為移動所造成的誤差影像。經過動態比對，修正後的影像有很明顯的改進。 除了影像位移的改進，另外我們也發現透過對於位移的校正，對這些動態對比增強磁共振影像(DCE-MRI)的數學模型參數之分佈曲線有很顯著且明顯的影響。 因此位移效應之修正對於影像品質和參數之數值分佈有很重要的影響，我們希望這些影響對於此抗癌藥物的化療效果有指標性的評估作用。

Dynamic-contrast-enhanced MRI (DCE-MRI) is the usage of fast pulse sequence MRI for monitoring the perfusion of contrast materials in the blood stream. This application is useful in evaluating patients taking angiogenesis inhibitors. The behavior of the contrast material in the blood stream can then be modeled using a variety of different mathematical models. Currently, by measuring select data and employing the different mathematical models, it is possible to estimate the flow characteristics of the contrast materials in the blood stream as well as around the tumor. Subsequently, by using the different flow characteristics, it is possible to evaluate the tumor in a non-invasive way. The patients in this study were lung cancer patients, and had been given Avastin. By employing DCE-MRI, it may be possible to predict and evaluate the effect of the chemotherapy early in the treatment course. This study employs Mistar, which is the software that provides the multitude of mathematical models for the evaluation treatment response. Due to limitations of the software, however, inconsistencies resulting from image translation between tomography slices—due to spontaneous movements such as breathing—cannot be adjusted or corrected. Therefore, besides acquiring the DCE-MRI data for mathematical models, this study further employs image pre-processing for the correction of imaging errors due to subject movement. After dynamic comparison and correction, the image is vastly improved. Besides improvements in image quality, translational correction also drastically improves the data used in the mathematical models. The usage of translational correction, therefore, greatly affects the final image quality, as well as the statistical distribution of the data being measured. We hope these changes will have a landmark impact on the final evaluation of the effectiveness of the angiogenesis inhibitors as well.