Ehlers-Danlos syndromes (EDS) are inherited connective tissue disorders characterized by extracellular matrix abnormalities, mainly affecting collagen. Key clinical features include skin hyperextensibility, joint hypermobility, and tissue fragility. Common subtypes are classical EDS (cEDS) associated with specific gene mutations, and hypermobile EDS (hEDS), which lacks a confirmed genetic marker and is diagnosed based on the clinical criteria. Current diagnosis is largely subjective, often relying on assessments like the Beighton scoring, which can result in significant delays, specifically for hEDS.
The aim of this study was to assess whether Mueller matrix polarimetry (MMP), a label-free optical imaging technique, could detect quantitative polarization-based biomarkers capable of distinguishing healthy skin from EDS-affected skin and cEDS from hEDS. MMP characterizes tissue microstructure by measuring how polarized light is modified as it passes by tissue, allowing the extraction of parameters related to birefringence, depolarization, anisotropy, and symmetry features closely linked to collagen organization. The authors hypothesized that polarimetric parameters derived from unstained skin biopsy sections would reveal structural differences linked to EDS and potentially serve as objective diagnostic adjuncts.
The study included 19 participants (mean age = 33 ± 7 years): 3 healthy controls, 5 patients with genetically confirmed cEDS, and 11 patients with hEDS diagnosed as per the 2017 clinical criteria. A 3-mm punch biopsy was obtained from the upper arm of each participant. Samples were embedded in optimal cutting temperature compound, frozen, and sectioned at 7 μm thickness. One section was stained with hematoxylin and eosin (H&E) for conventional histology, and one was left unstained for polarimetric imaging for each biopsy. Whole-slide polarized light microscopy was performed without pathologist-selected regions of interest to minimize selection bias. A 4× objective lens and a 2048 × 2048 pixel CMOS camera were used, yielding a field of view of 3.33 × 3.33 mm² with below 2.2 μm resolution. For each pixel, a full 4 × 4 Mueller matrix was reconstructed from 24 polarization-resolved images to improve signal-to-noise ratio. The raw Mueller matrices were processed by using 4 complementary mathematical frameworks-Lu-Chipman decomposition, Mueller matrix transformation, rotation invariance, and linear identity, yielding 24 derived polarimetric parameters.
About 10⁶ pixel values per parameter were obtained per slide for quantitative analysis. Tissue regions were segmented by using intensity-based masking. Median values were calculated to generate one representative value per patient for each sample. Statistical comparisons were performed in 2 stages: poled EDS vs healthy controls and cEDS vs hEDS. Normality was assessed by using the Shapiro-Wilk test. Independent-samples t-tests or Mann–Whitney U tests were applied with significance defined as p < 0.05 based on distribution. No samples were excluded; however, results involving the small healthy group (n = 3) were interpreted cautiously.
Visual inspection of H&E and polarimetric maps (retardance R, diattenuation D, and depolarization power Δ) did not reveal consistent group-level differences, necessitating statistical evaluation. In the healthy versus EDS comparison, three parameters showed statistically significant differences. Linear polarizance (Pₗ) was significantly higher in healthy tissue (p = 0.0033), indicating preferential transmission of linearly polarized light consistent with more uniformly aligned collagen. The parameter β linked to circular birefringence was higher in healthy samples (p = 0.02), suggesting reduced anisotropic effects in EDS. Circular retardance (ψ) showed a borderline significant difference (p = 0.048), although most values clustered near zero, and interpretation was limited by small sample size. Total retardance (R) and linear retardance (δ) were not significantly different between groups, indicating that circular polarization-related metrics may provide greater discriminatory power than traditional linear birefringence measures.
In comparing EDS subtypes, five parameters showed significant differences. Pₗ remained significant (p = 0.03), although with reduced discriminatory strength compared to healthy vs EDS analysis. The parameter rₗ showed stronger statistical separation (p = 0.008). P₁ (p = 0.046), P₃ (p = 0.02), and Pₜₘₛ (p = 0.01) were higher in hEDS compared with cEDS. These parameters reflect aspects of linear polarization behavior and deviations from idealized optical symmetry. The consistently higher linear polarization–related values in hEDS suggest that its collagen microarchitecture may not be simply more disorganized than cEDS but may exhibit distinct anisotropic features. Some derived parameters remain complex to interpret physically due to their algebraic construction from Mueller matrix elements.
This exploratory study shows that label-free polarized light microscopy and Mueller matrix analysis can identify significant differences in dermal microarchitecture between healthy and EDS-affected skin, as well as between cEDS and hEDS subtypes. Parameter Pₗ emerged as a consistent discriminator, while β and ψ suggested the relevance of circular birefringence in EDS characterization. Despite limitations from small and imbalanced sample sizes, the findings indicate that polarimetric biomarkers could serve as objective tools for EDS classification. Further research with larger, balanced cohorts and advanced methodologies is required to confirm diagnostic utility and clinical relevance.
Reference: Tumanova K, Min E, Louie DC, et al. Label-free differentiation of classical and hypermobile Ehlers–Danlos syndromes using Mueller matrix polarimetry. Biophoton Discov. 2026;3(1):015002. doi:10.1117/1.BIOS.3.1.015002
