Andre Khalil
Professor of Biomedical Engineering,
Director, CompuMAINE Lab,
University of Maine, Orono, ME, USA
https://umaine.edu/compumaine/
» Loss of mammographic tissue homeostasis in breast carcinomas «
The current breast cancer projects in my lab investigate individual breast tumors, as well as their microenvironment, through mammography.
Several studies show that malignant tumor growth may develop in fractal patterns. We used the 2D Wavelet-Transform Modulus Maxima (WTMM) method to estimate the 3D fractal structure of breast lesions—in this case, microcalcification clusters—by pairing the information from two separate 2D-projected mammographic views of the same breast. 92% of malignant breast lesions studied were fractal, while 88% of the benign lesions were Euclidean (non-fractal). These results support the notion that the fractal structure of malignant tumors is more likely to be associated with an invasive behavior into the surrounding tissue compared to the less invasive, Euclidean structure of benign tumors.
The multicellular architecture and organization of the ductal tree of the mammary gland has coherent cellular movement within an extracellular matrix cocoon that guides formation of structural units of tissues important for quiescence and homeostasis. This motivated us to “think outside the tumor”. We verified the hypothesis that this disruption of coherent angular motion, and the consequential adoption of randomized motility associated to malignant transformation, is a physical phenomenon that we can characterize quantitatively via the roughness fluctuation analysis of mammographic microenvironment tissue. When compared to normal tissue environment, the tissue in the microenvironment of tumors is disrupted, as quantified via the 2D WTMM method. The density fluctuations in healthy mammographic breast tissue, characterized by their surface roughness by the Hurst exponent, H, is either H~1/3 (high roughness) for fatty tissue or H~2/3 (low roughness) for dense tissue. However, increasing numbers of tissue regions with H~1/2 were found surrounding tumors. We infer that the underlying physical processes associated with breast tumors have a H~1/2 signature, which indicates randomness, lack of spatial correlation, and free diffusion.