Section-level spatial autocorrelation enables cross-study evaluation of myocardial organization after dECM deployment

Abstract

Decellularized extracellular matrix (dECM) is the tissue-derived scaffold that remains after cellular material is removed. After delivery, it encounters heterogeneous myocardium, creating a need for a tissue-scale readout that can be applied across formulations and delivery settings. We developed a section-level spatial-autocorrelation readout and applied it to 21 publicly separable capture/section units from two rat Visium studies of intramyocardial dECM injection and simulated intracoronary infusion. A histology-based, sample-balanced atlas aligned the broad cardiac fields represented in both studies. For each section, we averaged separately calculated Moran's I values for cardiomyocyte, fibroblast, myeloid and endothelial fields, then estimated the association with dECM exposure in a study–time-stratified model. The dECM coefficient was +0.0832 (95% CI -0.0240 to +0.1903; exact two-sided P=0.0933). The estimate was positive, while the interval included zero and retained substantial uncertainty in direction and magnitude. All four field coefficients and all 21 leave-one-unit-out estimates were positive, and the direction persisted after field omission, broader spatial graphs and use of a separately constructed field representation. Orthology-corrected spatial ranks identified an overlapping ECM/collagen program family whose fixed-rule scores were spatially structured. The analysis provides a common section-level axis for heterogeneous dECM studies and defines an ECM/collagen hypothesis for prospective experiments that measure material properties and tissue organization together.