A calibration curve model based on sine function / 中华检验医学杂志

ABSTRACT

Objective:This paper is to propose a calibration model based on sine function which enables more choices to determine the functional relationship between the absorbance and the concentration of the tested substance.Methods:This paper uses Taylor series expansion and Levenberg-Marquardt to obtain the optimal parameters for the Sine model and then summarizes the characters of the Sine model. On the basis of these characters, this paper compares and evaluates the experimental data processed by the Sine model from four aspects: correctness, precision, linearity and correlation.Results:The generated sine function calibration model achieved deviations within ±3% of the national standard substance, precision ( CV%) less than 2%, and a linear correlation coefficient greater than 0.990 within the measurement range of 32-710 mg/L. The correlation coefficients between the sine model and other well-performing linear calibration models for 104 clinical samples were all greater than 0.990. Conclusions:The performance evaluation of the prealbumin assay kit using the sine function calibration model meets industry standards and shows good correlation with the results of clinical sample measurements. This indicates that the sine function calibration model can serve as a new calibration model for in vitro diagnostic research and clinical applications.