High-throughput Image-based Matrix Analysis Platform Software (HIMAPS): Development of an algorithm for standardized characterization of high-resolution fibrin clot images

Background Scanning Electron Microscopy (SEM) has been gaining traction to provide high-resolution images of biological specimens such as blood clots to associate their properties with clinical outcomes. However, there lacks a "gold-standard" method of quantifying these images. We developed and validated a novel high-throughput, semi-automated computational pipeline for quantifying high-resolution images of biological samples called HIMAPS (High-throughput Image-based Matrix Analysis Platform Software). Methods From an imported image, region of interest (ROI) is defined and individual pixels are calibrated to physical dimensions (μm). The brightness intensity values of each pixel are then normalized to generate relative intensity value (RIV) and are subjected to the stratification algorithm to partition the image into layers. The fiber structures identified are enhanced and visualization is optimized as binary masks undergo morphological refinement to reduce noise, which can then be quantified for: a) fiber thickness from a skeletonized representation of the network, and b) network porosity and density from identification of pores. Results Fiber thickness was used as a validation method by comparing against the published standardization manual protocol. Our algorithm was within the 95% confidence interval of the manual measurements (N = 8 images, > 80 fibers per image) when the top 60% of the RIV was included. The top 55%-60% intensity-stratified reconstruction also led to the most robust pore size quantitation under the imaging conditions tested. Conclusion This algorithm can accurately quantify physical properties of clots such as fiber thickness, fiber matrix density, and porosity in a semi-automated, rapid, and unbiased manner.