血清胆红素的无创检测在疾病的预防、早期诊断与后期治疗等阶段都具有极其重要的作用。本研究提出了一种基于近红外光谱技术的无创血清胆红素新方法。通过采集舌尖的近红外反射光谱并运用偏最小二乘法对采集到的光谱数据进行建模,从而实现对血清胆红素的快速无创检测。将采集到的全部57例样本按照4:1的比例分训练集和预测集,分别建立总胆红素(TBIL)、直接胆红素(DBIL)和间接胆红素(IBIL)的偏最小二乘回归模型。3个模型的相关系数分别为0.9922,0.9947和0.9486,预测均方根误差(RMSEP)分别为6.13,4.61和4.05!mol/L。结果表明:舌尖近红外反射光谱结合偏最小二乘法可用于血清胆红素含量的无创检测,并且为其它血液成分的无创检测开辟了新路径。
The noninvasive measurement of serum bilirubin has important significance in early diagnosis, estimation, and prognosis diagnosis. A novel method was proposed to measure the serum bilirubin (BIL) noninvasively by the NIR spectroscopy. The experiment was made by collecting at the tip of tongue from 57 volunteers. All the 57 samples were divided into two parts: training set and prediction set by the proportion of 4 : 1. The prediction models of total BIL, direct BIL and indirect BIL were established by the PLS algorithm and satisfactory results were obtained. Correlation coefficients (R) were 0. 9922, 0. 9947 and 0. 9486; root mean square errors of prediction (RMSEP) were 6.13 μmol/L, 4.61 μmol/L and 4.05 μmol/L. All above have revealed that the method using NIR spectroscopy combined with PLS has been realized the noninvasive measurement of serum bilirubin, and could be a new breakthrough in noninvasive measurement of other human blood components.