Date of Award

Summer 8-19-2022

Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science in Biomedical Engineering


Biological Engineering


Andre Khalil

Second Committee Member

Karissa Tilbury

Third Committee Member

Peter Brooks


Despite recent advancements in biomedicine, cancer is still the second leading cause of death in the United States. Early detection of cancer is critical to improving patient care, but there are risks of over screening caused by the need for surgical biopsies in many cancers for final diagnostics. Recent advancements in computer aided diagnostics for breast cancer screening has reduced the need for biopsies and resulted in earlier diagnoses which has lowered the mortality rate from breast cancer within the past two decades. Developing new computer aided diagnostic tools that can be applied to a vast majority of cancers would serve to improve quality of life worldwide. These tools could also help researchers target and understand biological markers that lead to more malignant cancers improving both our treatment and understanding of cancer progression. The novel combination of the label-free, collagen-specific microscopy technique known as second harmonic generation (SHG) and the 2D Wavelet Transform Modulus Maxima (2D WTMM) Anisotropy Method is a prime candidate to serve this role. The 2D WTMM Anisotropy Method, originally developed for galactic astronomy and then used in multiple biological studies, was further adapted for SHG imaging of cancer in this work by improving both the binning and normalization techniques. This improved method was then applied to forty slides from pancreatic ductal adenocarcinoma (PDAC) patients and eight images were captured per each tissue category on each slide. Cancer and fibrosis had greater anisotropy factors (Fa) at small wavelet scales than normal and normal adjacent tissue. At scales larger than 21 μm this relationship changed with normal tissue having a higher Fa than all other tissue groups. This demonstrated that our developed method is sensitive to changes induced by PDAC. Our method was also compared to other open source SHG image analysis tools currently used by researchers in the field by generating 100 simulated fiber images at four different angle distributions of 0-180°, 30-150°, 60-120°, and 85-95°. The 2D WTMM Anisotropy method could differentiate the 0-180° and 30-150° groups at multiple scales whereas off-the-shelf tools could not. Four different levels of white noise were also added to the 60-120° angle distributions images to test each methods sensitivity to noise by comparing each noise convolved fiber image to pure white noise. The 2D WTMM Anisotropy Method was the only method capable of differentiating all added noise levels from white noise demonstrating its superior resistance to noise. This method will soon be applied to a larger breast cancer study and a breast cancer spheroid study. In both cases further developments to the method are planned, such as developing a version capable of analyzing 3D images and coupling the method with a machine learning technique.