Automated Quantitative Assessment of Elastic Fibers in Verhoeff-Van Gieson-Stained Mouse Aorta Histological Images

Abstract

The mechanical resilience of the aortic wall hinges on the organisation of its concentric elastic laminae, yet histological evaluation of these fibers remains largely qualitative and observer-dependent. We present a fully automated, stain-aware pipeline that transforms Verhoeff-Van Gieson (VVG) whole-slide images of mouse aortae into reproducible, quantitative maps of elastic-fiber architecture. Leveraging optical-density deconvolution to disentangle elastin from collagen, the workflow couples multi-resolution processing with graph-based skeletonisation to preserve gigapixel detail while scaling efficiently. It returns pixel-level measurements of fiber thickness, tortuosity, lamina count and network complexity, together with validation snapshots for transparent quality control. By eliminating observer bias and delivering high-throughput morphometry, our framework enables powered genotype-phenotype screens in genetically diverse mouse populations and provides objective read-outs for interventions aimed at preserving matrix integrity. The modular codebase is open-source, readily extendable to other elastin-rich tissues or stains, and forms a bridge between qualitative microscopy and biomechanical phenotyping--setting the stage for large-scale, data-driven exploration of vascular structure-function relationships.