數值及最佳化方法於生理模型之應用

Abstract

正電子發射斷層掃描（PET），已被廣泛用於分子成像。為了獲取有用的信息從測得的PET圖像診斷，起源於藥代動力學領域的模型已被廣泛應用的數據分析。模型中的未知參數，通常使用曲線擬合方法解決。然而擬合方法的準確性在存在大量的未知參數時，測得的結果通常為區域性的最小值。最計算過程中往往忽視一些小值或認為不重要的參數，以減少未知數。但這種減少可能會導致一些重要的信息遺失。在本論文中，我們提出了粒子群優化（PSO）的方法，具有較好的療效分析建模問題解決的18架F氟-1,4 - 二羥基苯基-L-丙氨酸（FDOPA）在帕金森氏動力學模型和測試我們的方法疾病的診斷。

癌症是現代社會中的死亡的首要原因。常規治療包括抗癌治療劑注射到腫瘤組織。治療效果取決於藥物於組之間的傳遞能力。大分子或納米粒子，可以針對腫瘤組織產生療效，且對正常組織的毒性有限，特別適合做為腫瘤藥物的載體。這些藥物濃度的強度分析可以結合螢光及使用光學顯微鏡，提供時間和空間分佈的藥物。在這項工作中，我們調查分子量不等Dextran 的反應。如血管壁通透度，擴散的影響因素進行了定性研究，通過計算機模擬。針對奈米載體在特定時間濃度和累計濃度分佈進行了研究，以評估不同條件下的治療效果。

Positron emission tomography (PET), with many kinds of radioactive tracers, have been used widely for molecular imaging. In order to retrieve useful information and render a diagnosis from measured PET images, the compartment models that originated from the area of pharmacokinetics have been employed extensively for data analysis. The unknown parameters in the models are usually solved by use of curve-fitting approaches. However, the accuracy of the fitting methods is usually below satisfactory level when there exist a large number of unknown parameters. As a result, some small-valued parameters are often neglected to reduce the number of unknowns. But such a reduction can lead a loss of some important information. In this work, we propose the particle swarm optimization (PSO) method to analyze the modeling problems with a better efficacy and tested our method by solving the 18 F Fluoro-3,4-dihydroxyphenyl-l-alanine (FDOPA) kinetic model in Parkinson’s disease diagnosis.

Cancer is the leading cause of death in the modern society. The conventional treatment includes the injection of anticancer therapeutic agent to tumor tissues. The efficacy of the treatment depends on the delivery of the drug in the interstitium. The substances that could improve the delivery process can be employed as drug carriers. Macromolecules or nanoparticles that can target at tumor tissues and contain limited toxicity to normal tissues are particularly suitable for serving as carriers. These carriers can be conjugated to flurophores and tracked using optical microscopy to provide spatial and temporal distribution of the drug. In this work, we investigated bio-distribution responses of dextrans with molecular weight ranging from 10 k-Da to 2 M-Da. Factors such as carrier molecular weight, and diffusion of carriers were quantitatively studied through computer simulations. The temporal and accumulated distributions of dextrans were investigated to evaluate therapeutic effects under different conditions.