Jared Westreich 1, Mohammadali Khorasani 2, Adam Gribble 1, I. Alex Vitkin 1,3,4
1 Department of Medical Biophysics, University of Toronto
2 Division of Surgical Oncology, Department of Surgery, University of Toronto
3 Department of Radiation Oncology, University of Toronto
4 Division of Biophysics and Bioimaging, Princess Margaret Cancer Centre
Polarimetry is a promising optical method to noninvasively assess biophysical characteristics of tissues. As polarized light propagates through tissue, its polarization state is altered by heterogeneous scattering bio-structures. A potential application of polarimetry is in intra-operative margin assessment to identify regions suspected to contain tumor and then using mass spectrometry (MS) to obtain a definitive classification. Thus, our proposed methodology is using polarimetry, which is rapid and widefield, to guide MS, which is accurate but very slow. If successful, this could reduce the number of breast cancer patients requiring reoperation due to positive margins (currently around 20%) [1]. The information to characterize tissue is contained within the Mueller matrix (MM), a complete mathematical description of tissue’s interaction with polarized light. Biophysical quantities derived from the MM, such as depolarization (a measure of tissue heterogeneity) and birefringence (a measure of anisotropy) can help differentiate tumor from surrounding healthy tissue. The goal of my project is to use intensity and texture features measured from various polarimetry images to classify tissue type. Machine learning classifiers, such as decision trees and support vector machines are used to segment out the tumorous region within a microscopy slide, containing human breast cancer. If this classifier can avoid false negatives (tumor predicted to be healthy tissue), it can guide MS, making it more feasible for intraoperative margin assessment.
[1]
E. R. St John, R. Al-Khudairi, H. Ashrafian, T. Athanasiou, Z. Takats, D. J. Hadjiminas, A. Darzi and D. R. Leff. “Diagnostinc Accuracy of Intraoperative Techniques for Margin Assessment in Breast Cancer Surgery.” Annals of Surgery. 265 300-310 (2017).