【摘要】 目的研究基于近红外(NIR)光谱的无试剂快速同时测定高血脂症指标总胆固醇(TC)、甘油三酯(TG)的方法。方法利用NIR光谱结合偏最小二乘(PLS)回归建立人血清样品TC、TG的定量分析模型,并基于定标、预测集的多次划分讨论模型稳定性。结果经过比较、检验,组合波段(780~1 878 nm,2 100~2 320 nm)达到了良好的预测精度和稳定性。对于TC、TG指标,最优PLS因子个数分别为13、15,检验预测均方根偏差分别为0.394、0.228(mmol·L-1),检验预测相关系数分别为0.937、0.982。光谱预测值与常规方法实测值的吻合性、相关性均良好,定量识别高脂血症的特异性、灵敏度分别达到91.5%、92.1%。结论 NIR光谱可以提供一种适合于高脂血症人群检测的有潜力的简便方法。更多还原

【Abstract】 Objective To study the no-reagent rapid simultaneous determination method for the total cholesterol(TC) and triglyceride(TG) of hyperlipidemia indicators based on near-infrared(NIR) spectroscopy. Methods NIR spectroscopy combined with partial least squares(PLS) regression was used to establish the quantitative analysis model of TC and TG for human serum samples, and the modeling stability was discussed on the basis of varied partitioning of the calibration and prediction sets. Results By comparison and validation, the composite waveband(780 ~ 1 878 nm, 2 100 ~ 2 320 nm) achieved good prediction accuracy and stability. As for TC, TG indicators, the optimal numbers of PLS factors were 13, 15, and the validation root mean square error of prediction were 0.394, 0.228(mmol·L-1), and the validation correlation coefficients of prediction were 0.937, 0.982, respectively. The spectral predicted values showed good consistent and correlation with the conventional measured values. The specificity and sensitivity for quantitative identification of hyperlipidemia achieved91.5% and 92.1%, respectively. Conclusions The results show that NIR spectroscopy can provide a potential simple method suitable for detecting hyperlipidemia in large population.更多还原