Quantitation in toxicologic pathology: liver fibrosis scoring and Measuring glycogen in livers
Here are a few examples of ways image analysis can be used which have practical applications in clinical use. Check out Flagship Biociences LLC:
Quantitation in toxicologic pathology: liver fibrosis scoring
"Industry has accepted relative scoring of toxicity lesions for many years. However, the real demand from investigators in the area of discovery toxicology is for quantitative data. In the scenario below, there are four images of periportal fibrosis in the livers of diabetic rats. The upper left, the control, clearly shows an increase in fibrosis as compared to the other three, on fenofibrate, pioglitazone, and a proprietary compound X. I can see levels of fibrosis as a pathologist, and grade according to the common four level scores of minimal, mild, moderate or marked. Alternatively, with image analysis I can deliver quantitative data to my investigators, as the graph (insert) shows. With more data I can then look at whether there are any statistically significant differences in terms of percent area of fibrosis between the control and these three compounds. Quantitation allows others, especially non-pathologists, to see and confirm what I am seeing."
"Measurement of hepatocellular glycogen is an appropriate exercise in determining level of gluconeogenesis in rodent efficacy studies for evaluating potential antidiabetic compounds. However, accuracy in glycogen quantitation requires strict control of the fasting period in all study animals because intracellular levels can vary greatly depending on postprandial time. Technical challenges in visualizing glycogen also exist because it is highly soluble, and can quickly dissipate with use of common soluble fixatives such as 10% formalin.
The periodic acid Schiff (PAS) stain is commonly utilized for visualizing glycogen, but its use suffers from the fact that the blue hematoxylin component of the stain tends to obscure the red Schiff stain in the hepatocyte cytoplasm. More accurate quantitation is enabled by using a color deconvolution algorithm which “deconvolves” or separates the various stain components (ie. blue hematoxylin and red Schiff). Effectively, then the pathologist can actually look “under” the obscuring hematoxylin to quantitate intensity and area of the red-stained glycogen."
