Progress in inferring desmoplastic stromal tissue microstructure noninvasively with ultrasound nonlinear elasticity imaging

Abstract

Desmoplasia is the abnormal growth of fibrous tissue that is often associated with malignancy in breast and other solid tumors. The collagen comprising the extracellular tissue matrix tends to be more dense than in normal breast tissue, and to exhibit altered microstructure. The distinctive mechanical properties of solid tumors are often attributed to the abnormal density and microstructure of their collagen extracellular matrices. For example, a higher concentration of collagen implies a larger linear elastic modulus, while a decrease in fiber tortuosity would mean a smaller “toe” region in the stress-strain curve. This talk describes progress of inferring local tissue microstructural properties from noninvasive, in vivo measurement of nonlinear elastic properties of breast tissue. An ultrasound elasticity-imaging based approach for determining the average local micro-structural properties of tissue was developed and implemented. It utilizes multiscale homogenization methods to determine effective macroscopic properties from assumed microstructural arrangements of collagen fibers. The resulting nonlinear constitutive equation is then used in a model-based elastic inverse problem solver to obtain local microstructural tissue properties from macroscopic observations of tissue deformation. Its potential in diagnosing malignant breast tumors in a small set of patients was also examined. [NSF Grant 50201109; NIH NCI-R01CA140271.]