Abstract
Structural and mechanical differences between cancerous and healthy tissuegive rise to variations in macroscopic properties such as visual appearance andelastic modulus that show promise as signatures for early cancer detection.Atomic force microscopy (AFM) has been used to measure significant differencesin stiffness between cancerous and healthy cells owing to its high forcesensitivity and spatial resolution, however due to absorption and scattering oflight, it is often challenging to accurately locate where AFM measurements havebeen made on a bulk tissue sample. In this paper we describe an imageregistration method that localizes AFM elastic stiffness measurements withhigh-resolution images of haematoxylin and eosin (H\&E)-stained tissue towithin 1.5 microns. Color RGB images are segmented into three structure types(lumen, cells and stroma) by a neural network classifier trained onground-truth pixel data obtained through k-means clustering in HSV color space.Using the localized stiffness maps and corresponding structural information, awhole-sample stiffness map is generated with a region matching andinterpolation algorithm that associates similar structures with measuredstiffness values. We present results showing significant differences instiffness between healthy and cancerous liver tissue and discuss potentialapplications of this technique.