ECM profiling of patient-matched endometrial PDX-tumors reveals translational drift and TME remodeling

Background: Extracellular matrix (ECM) remodeling, particularly collagen deposition, have been linked with aggressive features of endometrial. It plays a critical role in tissue architecture, deep myometrial invasion, immune infiltration, and therapeutic response. Patient-derived xenograft (PDX) models are widely used in cancer research; however, their ability to preserve patient-specific ECM organization remains uncertain. This study quantified ECM composition and spatial topology to evaluate species-specific translational drift. Methods: Formalin-fixed, paraffin-embedded tissue from two primary endometrial tumors and matched PDX tumors underwent dual Picrosirius Red (collagen) and Alcian Blue (hyaluronic acid; HA) staining. High-resolution brightfield images were captured under standardized conditions using an Agilent Lionheart system. ECM profiling was performed using the CerFlux PEER AI/ML imaging pipeline, extracting 63 histomorphometric and spatial parameters, including ECM area fractions, HA:collagen ratios, gray-level co-occurrence matrix (GLCM) texture metrics, and Moran's I spatial autocorrelation. Analyses were cross validated in Fiji. Group comparisons used nonparametric statistics, and principal component analysis (PCA) summarized multivariate ECM signatures. Results: Patient tumors demonstrated substantially higher collagen area fraction than matched PDX tumors (0.20 ±0.01 vs 0.05 ±0.04; p<0.05), while HA levels were similar (0.41 ±0.03 vs 0.33 ±0.14). This yielded a marked shift in HA:collagen ratio, from 2.1 ±0.3 in patients to 7.9 ±2.7 in PDX tumors (p<0.05). Texture analysis revealed more heterogeneous collagen organization in patient tumors, with higher entropy (8.22 ±0.10 vs 8.07 ±0.02; p<0.05) and lower homogeneity (0.15 ±0.02 vs 0.19 ±0.01; p<0.05). HA exhibited a similar pattern of increased uniformity in PDX tumors (homogeneity 0.16 ±0.07 vs 0.10 ±0.01; p<0.05). Spatial statistics showed a nearly two-fold increase in collagen Moran's I in PDX tumors (0.51 ±0.02 vs 0.26 ±0.07; p<0.05), indicating stronger ECM clustering and reduced spatial heterogeneity relative to patient tissue. PCA cleanly separated patient and PDX samples along PC1 (58% variance), driven by HA:collagen balance, textural heterogeneity, and spatial autocorrelation. Conclusions: Quantitative ECM profiling demonstrates that patient-matched endometrial PDX tumors undergo pronounced translational drift, including collagen depletion, increased HA:collagen ratio, higher ECM homogeneity, and strengthened spatial clustering. These changes reflect species-specific TME remodeling following xenografting and may influence drug penetration, response, and biomarker interpretation. Assessing ECM fidelity is therefore vital when using PDX models to inform translational research and preclinical decision-making in endometrial cancer.