An automated 3D polarization-resolved second harmonic generation imaging system for large scale cardiovascular tissue mapping

Modern AI techniques for biomedical applications, particularly predictive modeling, are limited by the size and quality of available datasets. Existing datasets often lack sub-cellular structural information over large tissue volumes, limiting their utility for cellular-level mechanical modeling. Here, we introduce an automated polarization-resolved second harmonic generation (P-SHG) imaging system that maximizes throughput of large scale, label-free, 3D images of myocardial tissues at sub-cellular resolution. Once a sample is loaded, the entire imaging process is fully automated, enabling unattended, continuous operation to generate large-scale datasets suitable for AI-driven modeling of tissue mechanics. We demonstrate the system by acquiring a stitched 160 μm × 160 μm × 31.25 μm volumetric SHG scan of porcine myocardium, clearly resolving collagen and myosin microstructures, and a 20 μm × 20 μm P-SHG image of murine myocardium, illustrating the system’s ability to capture molecular-orientation-dependent contrast.