Quantitative estimates of mechanical properties from in vivo tumors of rodent animals

Images of tumor mechanical properties provide important insights into malignant processes characterized by extracellular matrix remodeling and stiffening. This thesis focuses on using ultrasonic tools to image shear waves generated in spontaneous rodent mammary tumors to estimate the complex shear modulus under realistic conditions, and to assess the sensitivity for relating matrix viscoelasticity to tumor type and stage. Cylindrical, harmonic shear waves generated by a vibrating needle (50-450 Hz) were tracked by Doppler-ultrasound in vivo. Estimates of spatial-phase gradients yielded shear-speed dispersion curves that were fit to values predicted from a Kelvin-Voigt rheological model to find the elastic and viscous coefficients of the complex modulus. Subsequently, tumors were excised for rheometry testing to independently validate coefficient estimates and later histologically processed to evaluate tumor type. Four benign mammary fibroadenomas (2-4 cm diameter), one benign lobular hyperplasia (3 cm diameter) and four 4T1 cell line induced carcinomas (1-2 cm diameter) were studied. In rat tumors the average measured elastic (1.18 kPa) and viscous (1.79 Pa•s) coefficients compared well with standard rheometry. Viscoelastic properties correlated with fibrosis grade; both coefficients were seen to increase monotonically with fibrotic levels. In modeled mice carcinomas, diagnostic-specific changes of viscoelastic behavior were seen. Factors which might lead to variation in tissue mechanical property estimation are discussed, including ECM composition, rheological model, nonlinear wave propagation as well as boundary reflection. Addressing these issues is of great importance in enhancing modulus measurement accuracy and lesion image contrast by making the right decisions about mechanical excitation methods and constitutive models. Dynamic shear-wave estimation of complex modulus has demonstrated an ability to describe mammary tumors. The measurable differences between lesion types are greater than inter-animal or technique variabilities. Nevertheless, to reliably classify tumors using elasticity imaging, more knowledge on the biological contrast mechanism in tissue mechanical properties is needed. Future work will be on discovering how mechanical properties change with cancer progression and the way to monitor this change by assessing mechanoenvironment