Abstract
OBJECTIVE: The study purpose was to evaluate the associations between patient characteristics and the histologic remodeling scores of acellular dermal matrices (ADMs) biopsied from breast reconstruction sites in the first attempt to generate a multivariable risk prediction model of nonconstructive remodeling. It was hypothesized that host characteristics and surgical site assessments predict the degree of graft remodeling for ADMs used during breast reconstruction. METHODS: The ADMs were biopsied from the breast reconstruction sites of n = 62 patients during a subsequent breast procedure, stained with hematoxylin-eosin, and evaluated according to a semi-quantitative scoring system for remodeling characteristics (cell types, cell infiltration, extracellular matrix deposition, scaffold degradation, fibrous encapsulation, and neovascularization) and a mean composite score. Biopsies were stained with Sirius Red and Fast Green, and analyzed to determine the collagen I:III ratio. Based on univariate analyses between subject clinical characteristics and the histologic remodeling scores, cohort variables were selected for multivariable regression models using a P value of 0.20 or less. RESULTS: The composite score model yielded 3 variables: pack-year history, corticosteroid use, and radiation timing (r pseudo = 0.81). The model for collagen I yielded 2 variables: corticosteroid use and reason for reoperation (r pseudo = 0.78). The model for collagen III yielded 1 variable: reason for reoperation (r pseudo = 0.35). CONCLUSIONS: These preliminary results constitute the first steps in generating a risk prediction model that predicts the patients and clinical circumstances most likely to experience nonconstructive remodeling of biologic grafts used to reconstruct the breast.