Mid-Infrared Spectroscopic Imaging for Breast Tissue Histopathology: Towards 'Stainless Staining'.
Michael J Walsh, David M Mayerich, Elizabeth L Wiley, Andre Kajdacsy-Balla, Rohit Bhargava. University of Illinois at Urbana-Champaign; University of Illinois Medical Center, Chicago
Background: Mid-Infrared (IR) spectroscopic imaging is a novel approach to derive chemical images from tissues based on their inherent biochemistry. Histopathology is the gold standard for disease diagnosis. Current histopathological techniques use of panel of special stains and immunohistochemistry (IHC) to assess tissue architecture, determine cell types present and to classify cancers.
Design: Mid-IR images were obtained from over 200 individual patients using breast tissue microarrays. Serial sections were stained with a panel of 13 routinely used special stains and IHC stains. A modified Bayesian classifier was built to assign image pixels to the correct cell types and Artificial Neural Networks (ANN) to replicate staining.
Results: Using Mid-IR imaging coupled with the modified Bayesian classifier it was possible to segment breast tissue into the main 8-cell types of breast tissue from a single unstained tissue section. The sensitivity and specificity as measured by average Area Under the Curve (AUC) were very high (AUC=0.9). Mid-IR imaging coupled with ANN demonstrated that it was possible to accurately reproduce the staining of the panel of stains, all in a single unstained slide.
Conclusions: Mid-IR imaging coupled with Bayesian classification and ANN could potentially be a very valuable tool as an adjunct to current histopathological procedures, with the ability to take a single unstained tissue section and give a decision on the cell types present and also to replicate staining patterns. This approach could be particularly advantageous where limited histological and cytological specimen is available for analysis. Moreover, it is amenable to quantitative analysis of each component. This novel approach promises to revolutionize and expand the role of the pathologists in both research and tissue diagnosis.
Monday, February 28, 2011 1:00 PM
Poster Session II # 282, Monday Afternoon