Authors: Caroline Torgersen, Tyler O'Loughlin, Ellie Evans, Vern Hart, Clint Flinders
Mentors: Vern Hart
Insitution: Utah Valley University
Stage-IV cancers are commonly identified by tumors having metastasized to other parts of the body. However, studies have shown that cancerous tissues often release “seeds” of circulating tumor clusters (CTCs) into the cardiovascular and lymphatic systems long before metastasized sections of the tumor are identifiable. These CTCs can circulate or remain dormant for long periods of time, even after the lesion is excised. In addition, these structures exist on scales that are not currently identifiable using conventional imaging modalities and are only detectable after being isolated. To address this issue, we demonstrate a simple optical diffraction system utilizing visible laser light and a beam profiler to collect speckle images from polystyrene microbeads (mimicking CTCs) flowing through an IV tube (mimicking a vein or artery). These scattering images were used to train a convolutional neural network, which was able to distinguish bead sizes ranging from 30 to 120 microns (comparable to the diameter of CTCs). A Softmax classifier was included with multiple target categories corresponding to different clusters sizes. As blood cells are significantly smaller (5-20 microns), this system could be used to non-invasively identify the presence of larger scatterers in a blood stream in situ, indicating the presence of CTCs, and providing a potential diagnostic biomarker for early-stage cancer.
If a oral presentation is not available we would still love to present a poster.