动态血糖波动幅度评估参数的自动计算算法研究与应用

发布时间：2018-06-22 07:36

本文选题：动态血糖监测 + 血糖波动幅度评估参数　；参考：《南方医科大学》2017年硕士论文

【摘要】：继心脑血管病变、肿瘤之后,糖尿病已成为第三大严重威胁人类健康的慢性疾病,给个人健康和社会经济带来了严重影响,由糖尿病引发的慢性并发症是威胁糖尿病患者生命和健康的主要因素。研究表明,血糖波动是糖尿病慢性并发症发生、发展的潜在独立危险因素。若血糖长期波动过大,会增加患者罹患糖尿病慢性并发症的危险性,因此对糖尿病患者进行血糖波动的监测与控制是减少糖尿病慢性并发症发生的有效手段。为了更好的对血糖波动进行监测与控制,必须要准确地量化评估血糖波动幅度。目前已有许多参数被提出用于血糖波动幅度评估,主要包括:血糖值的平均值(MBG)、标准差(SDBG)、四分位距(IQR)、M值(M-value)、变异系数(%CV)、J系数(J-index)、最大血糖波动幅度(LAGE)、低血糖系数(LBG1)、高血糖系数(HBG1)、血糖风险评估(GRADE)、日均风险值(ADRR)、平均血糖波动幅度(MAGEA、日间血糖平均绝对差(MODD)、连续24小时血糖净作用(CONGAn=24)等,但迄今为止还没有统一公认的最佳指标。随着动态血糖监测技术(CGM)的发展及可靠性的提高,且MAGE在反映血糖波动与氧化应激反应有其独特的优势,而氧化应激反应有可能是导致糖尿病慢性并发症发生的机制,目前该参数已被越来越多的临床工作者认可,被认为可能是反映日内血糖波动的“金标准”。CGM为血糖波动评估带来了大量的血糖数据,然而CGM系统自身却并未提供目前所有血糖波动幅度评估参数的计算功能,当使用CGM数据进行以上评估参数的计算时,工作量将十分巨大,不利于参数的快速获得。尤其是进行MAGE参数计算,因为在临床上一般只能通过人工比较的方式筛选数据进行计算。若同时对多名患者CGM数据进行计算,参数计算的工作量将大大的增加,十分耗费时间。此外,临床研究人员必须要进行MAGE参数计算的相关专业培训方可确保数据筛选的准确性,否则容易导致计算结果存在一定的误差。目前国内外已经针对MAGE参数提出了几种自动计算方法及相关计算软件,但并未获得临床的广泛认可,主要是因为临床上缺乏计算结果准确性的检验标准,并且这些方法都缺乏数学理论的支撑使得计算结果的准确性易受到临床的质疑;同时这些自动计算软件并未囊括其他评估参数不利于多角度综合评估血糖波动情况。因此,为解决所存在的问题,本研究提出了动态血糖波动幅度评估参数的自动计算算法,很好地解决了现有评估参数的自动计算问题。尤其针对MAGE的自动计算,算法在MAGE定义的基础上,结合完善的有效血糖波动判断标准,构建了应用差分进化算法求解的基于非线性整数规划的M4GE计算数学模型。为使算法进一步应用于临床,利用C#语言在VS2010的编程环境中对算法进行开发,开发了相应的可以在Windows下安装使用的自动计算软件。该自动计算软件实现了现有评估参数的快速计算,同时该软件方便易操作、灵活的界面使得没有经过相关专业培训的人亦可以导入需要分析的CGM数据计算出参数值。为克服MAGE参数计算没有检验标准的难题,本研究通过统计学分析比较分别使用传统人工筛选数据的计算方法和自动计算算法对大量的不同类型的糖尿病患者的CGM临床数据的计算结果以验证自动计算算法的准确性,分析显示自动计算与手动计算结果之间具有高度相关性和一致性,从而验证了自动算法计算MAGE值的准确性。研究算法使血糖波动幅度评估参数值的获取更加简单方便与客观准确,大大地缩短了计算时间,提高了临床的效率,也为临床工作者同时综合利用多个参数评估血糖波动提供了可能,进一步推动血糖波动评估研究的开展。
[Abstract]:Following the cardiovascular and cerebrovascular diseases, after the tumor, diabetes has become the third major chronic disease that seriously threatens human health. It has a serious impact on personal health and social economy. The chronic complications caused by diabetes are the main factors that threaten the life and health of diabetics. The risk of chronic complications of diabetes can be increased if the long-term fluctuation of blood sugar is too large. Therefore, monitoring and control of blood glucose fluctuation in diabetic patients is an effective means to reduce the occurrence of diabetic chronic complications. It is necessary to accurately quantify the amplitude of blood glucose fluctuations. Many parameters have been proposed to assess the amplitude of blood glucose fluctuations, including the average blood sugar value (MBG), standard deviation (SDBG), four division distance (IQR), M value (M-value), variation coefficient (%CV), J number (J-index), maximum blood glucose fluctuation amplitude (LAGE), hypoglycemia coefficient (LBG1), hyperglycemia system Number (HBG1), blood glucose risk assessment (GRADE), average daily risk value (ADRR), average blood glucose fluctuation range (MAGEA, mean absolute difference of blood glucose (MODD), 24 hours of blood glucose net action (CONGAn=24), but so far there is no unified best indicator. With the development and reliability of dynamic glucose monitoring (CGM), MAGE is in the opposite direction. Hyperglycemic fluctuations and oxidative stress reactions have their unique advantages, and oxidative stress may be the mechanism that causes chronic diabetic complications. At present, this parameter has been recognized by more and more clinical workers. It is believed that the "golden standard".CGM, which may reflect the fluctuation of blood sugar in the day, has brought a lot of blood glucose fluctuations. Blood glucose data, however, the CGM system itself does not provide the computing function of all the parameters of the current blood glucose fluctuation assessment. When using the CGM data to calculate the above evaluation parameters, the workload will be very large, not conducive to the rapid acquisition of parameters, especially for the MAGE parameter calculation, because it is generally only by manual comparison in clinical. If the number of patients' CGM data is calculated at the same time, the workload of the parameter calculation will be greatly increased and time consuming. In addition, the relevant professional trainers who have to calculate the MAGE parameters to the clinical researchers can ensure the accuracy of the data screening, otherwise it will easily lead to the calculation results. At present, several automatic calculation methods and related computing software have been proposed for MAGE parameters at home and abroad, but it has not been widely accepted in clinical practice, mainly because of the lack of testing standards for the accuracy of the results of calculation, and these methods lack the support of mathematical theory so that the accuracy of the calculation results is easily subject to clinical practice. At the same time, the automatic calculation software does not include other evaluation parameters which are not conducive to multi angle comprehensive evaluation of blood glucose fluctuations. Therefore, in order to solve the problems, this study proposes an automatic calculation algorithm for the dynamic parameters of the fluctuation of blood glucose fluctuation, which is a good solution to the problem of the automatic calculation of the existing evaluation parameters. On the basis of the definition of MAGE and on the basis of the definition of MAGE, the algorithm constructs a mathematical model of M4GE computing based on nonlinear integer programming, which is solved by differential evolution algorithm. In order to make the algorithm further applied to the clinic, the algorithm is developed in the VS2010 programming environment by using the C# language. The automatic computing software which can be installed and used under the Windows is sent. The automatic calculation software realizes the rapid calculation of the existing evaluation parameters. At the same time, the software is convenient and easy to operate. The flexible interface makes the people without the related professional training can also import the CGM data that needs analysis to calculate the parameters. In order to overcome the MAGE parameters, the software can also be used to overcome the parameters. The calculation of CGM clinical data of a large number of different types of diabetic patients was compared with the results of the calculation method of traditional artificial screening data and automatic calculation algorithm to verify the accuracy of the automatic calculation algorithm, and the analysis showed that the automatic calculation and manual calculation were shown by the statistical analysis. There is a high correlation and consistency between the fruits, which verifies the accuracy of the MAGE value calculated by the automatic algorithm. The research algorithm makes the acquisition of the parameter value of the blood glucose fluctuation range more simple and more objective and accurate, greatly shortens the calculation time, improves the clinical efficiency, and uses multiple parameters for the clinical workers to evaluate the parameters simultaneously. It is possible to estimate the fluctuation of blood glucose and further promote the research of blood glucose fluctuation assessment.
【学位授予单位】：南方医科大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R587.2

【参考文献】