Intensity Histogram-Based Reliable Image Analysis Method for Bead-Based Fluorescence Immunoassay

Abstract

Bead-based fluorescence immunoassay is drawing attention as a next-generation technology in disease diagnosis owing to its high sensitivity and multiplexing capability. Fluorescence imaging of beads is typically used to determine their mean fluorescence intensity. However, the mean intensity can be evaluated differently depending on the analysis methods [such as the shape and size of the region of interest (ROI)]. To address these problems, this study proposes a highly reliable and reproducible image analysis method utilizing a fluorescence intensity-based effective pixel extraction technique. Various potential sources of defective signals (e.g., fluorescence aggregation, non-specific antigen–antibody reactions, and bead defects) can be prevented from contributing to the average value by selectively extracting pixels representing the specific reactions of antigens and antibodies in the ROI. In this study, we fabricated a microfluidic chip composed of multiple bead-based detection lines, performed fluorescence immunoassay, and then compared the mean fluorescence intensity calculated from the fluorescence images with that of a conventional analysis method. Using the conventional method, the evaluated average mean intensity value of beads varied significantly based on the size of the ROI with the coefficients of variation ranging from approximately 29–95%. In contrast, the effective pixel extraction method resulted in a coefficient of variation of approximately 3–7% under varying ROI size. Furthermore, the coefficients of variation for four detection lines containing various types of defective signals significantly decreased from approximately 7.1% to 2.6%. The proposed technique will help in minimizing the analysis deviation caused by different ROI selections or defective signals in fluorescent image-based immunoassays.