Dose-interpolation of immunoassay data: uncertainties associated with curve-fitting

Area of Science:

• Biochemistry
• Analytical Chemistry
• Medical Diagnostics

Background:

• Immunoassays are widely used for quantifying analyte concentrations in biological samples.
• Current methods often underestimate the error distributions associated with these concentration estimates.
• Accurate error estimation is crucial for reliable diagnostic interpretation.

Purpose of the Study:

• To develop improved methods for estimating analyte concentrations from immunoassay data.
• To investigate the impact of calibration curve fitting on error distribution accuracy.
• To evaluate common curve-fitting models used in automated immunoassay systems.

Main Methods:

• Utilized computer simulation with practical immunoassay data.
• Incorporated the distribution of the response metameter of the calibration curve.
• Accounted for uncertainties in the fitted calibration curve's location.
• Analyzed three common curve-fitting functions: logit-log, 4-parameter logistic, and the four-parameter Amersham model.

Main Results:

• Computer simulations provided more accurate estimates of analyte concentrations compared to standard methods.
• The study quantified underestimation of error distributions in typical immunoassay analyses.
• Performance variations were observed among the logit-log, 4-parameter logistic, and Amersham models.

Conclusions:

• The developed simulation model offers a valuable tool for improving immunoassay data analysis.
• Addressing calibration curve uncertainties leads to more reliable analyte concentration estimations.
• This approach enhances the accuracy and diagnostic utility of automated immunoassay systems.