Terahertz Imaging for Margin Assessment of Excised Breast Cancer Tumors Slides

Breast cancer is a serious disease with almost one out of eight women in the U.S. expecting a diagnosis of breast cancer in her lifetime. The American Cancer Society (ACS) is the leader in the fight to end breast cancer with investing more in breast cancer research than any other cancer type to detect, prevent, treat, and cure the disease. This research is focused on investigating the capability of terahertz (THz) imaging and spectroscopy to differentiate between different types of tumor tissues in lumpectomy surgery. Breast Conserving Surgery (BCS), lumpectomy, reduces breast disfigurement compared with radical mastectomy. However, the margins of excised tumor tissue need to be analyzed by pathologists to judge if there remain any traces of cancerous tissue. This can take several days and in the case of positive margin detection a second surgery is needed to remove more tissue. Unfortunately, clinical studies indicated that in around 20-40% of cases excised malignant breast tumors contain positive margins. THz technology offers several favorable features, including high sub-millimeter resolution, minimal scattering, and sensitivity to water content. Experimental data in the literature demonstrated image contrast enhancement when using THz waves compared with near-infrared and optical radiation. The novelty of this research lies in imaging three dimensional (3D) tumors instead of processed two dimensional (2D) sections. By using the whole excised tumor tissues, we demonstrate that THz imaging has a potential to be incorporated directly in the operating room for real time margin assessment. This research demonstrates THz transmission and reflection imaging of freshly excised and formalin-fixed, paraffin-embedded breast cancer tumors. THz imaging is shown a capability in defining the three-dimensional boundaries of tumors embedded in paraffin blocks. This procedure produces cross-section images of the tumor borders at any depth and enables direct correlation with histopathology sections. Tumors obtained from mice models and human breast are investigated. Unsupervised classifier is implemented to evaluate the THz image correlation with pathology. This research is a collaboration between electrical engineering, biomedical engineering and pathology at the University of Arkansas and Oklahoma State University.