Abstract

Infrared (IR) imaging spectroscopy of human liver tissue slices has been used to identify and characterize liver metastasis of colorectal origin which was surgically removed from a consenting patient and frozen without formalin fixation or dehydration procedures, so that lipids and water remain in the tissues. First, a k-means clustering analysis, using metrics from the IR spectra, identified groups within the image. The groups were identified as tumor or nontumor regions by comparing to an H and E stain of the same sample after IR imaging. Then, calibrant IR spectra of protein, several fats, glycogen, and polyvinyl alcohol were isolated by differencing spectra from different regions or groups in the image space. Finally, inner products (or scores) of the IR spectra at each pixel in the image with each of the various calibrants were calculated showing how the calibrant molecules vary in tumor and nontumor regions. In this particular case, glycogen and protein changes enable separation of tumor and nontumor regions as shown with a contour plot of the glycogen scores versus the protein scores.

Full Text
Paper version not known

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call