Breast Cancer Imaging – Diagnostic Algorithms

With the increased volume of data that is collected by imaging entire margins of tumor specimens, the need for automated methods to analyze and classify potentially suspicious areas in real time is critical. By looking at the individual axial scans, new classification methods based on the frequency analysis can yield real time guidance to suspicious regions of interest. This same technology can also be used to guide surgical needle biopsy procedures by providing information about the tissue at the needle tip potentially improving the diagnostic sampling rates.

Axial OCT scan data and frequency analysis showing differences between breast tissue types: (a,d) adipose, (b,e) tumor, and (c,f) stroma.

Automated computer diagnostic algorithms are being developed to differentiate breast tissue types using OCT data. Classification is shown for (a) combined, (b) Fourier-domain, and (c) periodicity analysis techniques.

Zysk AM, Boppart SA.  Computational methods for analysis of human breast tumor tissue in optical coherence tomography images.  Journal of Biomedical Optics, 11(5):054015, 2006. PubMed Abstract PDF
Zysk AM, Adie SG, Armstrong JJ, Leigh MS, Paduch A, Sampson DD, Nguyen FT, Boppart SA. Needle-based refractive index measurement using low-coherence interferometry.  Optics Letters, 32(4):385-387, 2007. PubMed Abstract PDF
Zysk AM, Nguyen FT, Chaney EJ, Kotynek JG, Oliphant UJ, Bellafiore FJ, Rowland KM, Johnson PA, Boppart SA. Clinical feasibility of microscopically-guided breast needle biopsy using a fiber optic probe with computer-aided detection. Technology in Cancer Research and Technology, 8:315-322, 2009. PubMed Abstract n/a