Interpretable computational SHG microscopy enables quantitative tissue phenotyping of cardiac remodeling in nonischemic heart failure
Heart failure is characterized by extracellular matrix remodeling and collagen deposition that impair myocardial function. Second harmonic generation (SHG) microscopy provides label-free visualization of fibrillar collagen and myosin but lacks standardized, interpretable, and quantitative frameworks with clinical-level performance. We present a feature extraction pipeline integrating 87 biologically interpretable SHG-derived metrics for automated classification of nonischemic heart failure in murine specimen. The framework distinguished normal, heart failure, and natural recovery states with area-under-the-curve values of 0.84-0.99, outperforming established open-source SHG feature extraction tools. These findings support interpretable, SHG-based quantification of cardiac remodeling and establish a methodology for future multimodal biophotonics applications.