Association Between Amyloid Deposition and Collagen-to-Elastin Ratio in the Ligamentum Flavum: A Machine Learning Histologic Study in Matched Groups

BACKGROUND AND OBJECTIVES: Amyloid deposition in the ligamentum flavum (LF) occurs in 25% to 45% of patients with lumbar stenosis (LS). Amyloid-positive LF is associated with greater LF thickness and healthier intervertebral discs on MRI, suggesting an alternative pathophysiologic mechanism for LF hypertrophy may be present in these patients. Whether this increased thickness reflects changes in the LF collagen and elastin composition is unknown. METHODS: LF specimens from 64 patients with LS were analyzed in a case-control design, drawn from a database of 324 consecutive cases with amyloid status confirmed by Congo Red staining and mass spectrometry. Thirty-two amyloid-positive patients were age-matched and diagnosis-matched to 32 amyloid-negative controls. Two adjacent sections per specimen were stained with Masson Trichrome and Verhoeff-Van Gieson stain. Collagen and elastin content were quantified using a random forest pixel classifier developed in QuPath and validated against manual segmentation supervised by a board-certified pathologist. Primary analysis compared the collagen-to-elastin ratio between groups, and secondary analysis compared the individual collagen and elastin content. RESULTS: Both collagen and elastin classifiers agreed with manual segmentation in the Bland-Altman analysis and showed bias of <1% tissue area. Amyloid-positive LF demonstrated a greater collagen-to-elastin ratio of 1.36 compared with 1.05 in amyloid-negative patients (P = .014). This was primarily driven by a lower elastin content, which occupied 7.5% less tissue area in amyloid-positive LF per secondary analyses (P = .008). The average LF thickness at a lumbar level was 0.538 mm thicker on T2-weighted MRI for amyloid-positive patients. CONCLUSION: Amyloid-positive LF exhibits a higher collagen-to-elastin ratio, driven by reduced elastin, suggesting distinct extracellular matrix remodeling in these patients with LS. This ratio may explain the greater LF thickness seen on MRI in amyloid-positive patients. There may be an amyloid-associated pathway contributing to LF hypertrophy in some patients, and further studies of this are warranted.