用CT值推导人体组织属性参数的方法及改进

蒙特卡罗方法是计算精度最高的人体组织剂量计算方法。常见的蒙特卡罗模拟程序,如M CNP,EG S4等都自带了各种介质的截面数据,而剂量计算的一个重要步骤就是得到人体组织的属性参数——质量密度和化学组成。本文主要介绍了两种根据医学影像CT值推导人体组织属性参数的方法——区间法和两分法,并经过分析比较认为两分法比区间法具有更高的精确度和可行性。同时,针对两分法存在的不足和缺陷,本文将CT值全范围划分为七段,根据两分法的假设条件假设每一段均由两种不同的人体组织混合而成。在每一段CT值范围内,人体组织的质量密度和化学组成通过对两种假定物质参数的拟合得到。结果说明该方法所得到的人体组织属性参数完全可以满足蒙特卡罗人体组织剂量计算的要求。

Method for Converting CT Numbers into Mass Density and Elemental Weights of Tissue and Its Improvement

The Monte Carlo simulation is available to calculate dose distributions in patients with the highest accuracy. To perform the dose calculation with Monte Carlo codes such as MCNP, EGS4 and so on, it is necessary to calculate the mass density and element weights of tissues. Because there is no direct unique relation between CT numbers and tissue parameters, this paper is to describe two methods-multisection and dichotomy methods-to convert CT numbers into the mass density and elemental weights of the tissue. It is known that the dichotomy method is more accurate and practicable compared with the multisection method. Furthermore, the dichotomy method can be improved by seven sections divided on the CT number scale and each confined to two selected tissues. With each section, the mass density and the elemental weight of the selected tissue are interpolated. For this purpose, functional relationships between the CT number and each of the tissue parameters composed of only two components in varying proportions are derived. Compared with conventional data fitness, no loss of accuracy in tissue parameters is accepted when using the interpolation function.