Over the last 2 years we have developed a set of algorithms (called AQUA) that allow the rapid, automated, continuous and quantitative analysis of tissue microarrays, including the separation of tumor from stromal elements and the sub-cellular localization of signals. We validated the technology and have used it to discover new biologically based disease sub-classifications. Validation studies using estrogen receptor in breast carcinoma show that automated analysis matches or exceeds the results of conventional pathologist-based scoring. We then used the automated analysis to show 2 examples of disease classifications not discernable by traditional pathologist based analysis. First, by measuring membrane levels of HER2 in 350 node positive breast cancers, we were able to detect two groups of patients with poor outcome, over expressers, as has been seen by pathologists, and a second group of very low level expressers, never seen by pathologists, but previously described in a work where breast tissues were quantitatively analyzed by enzyme-linked immunoassay. Thus the accuracy of this new technology appears to be as sensitive as """"""""grind and measure"""""""" type assays but with the added critical advantage of in situ subcellular localization. Subcellular localization is crucial in some situations. Our second study showed that quantification of beta-catenin expression within defined sub-cellular compartments allowed fractionation of colon cancer patients into two novel, prognostically significant groups that would be impossible to discover by traditional pathologist-based scoring. Although the small size of tissue microarray spots facilitated the development of AQUA, that technology could now be translated to examination of biopsy tissue to translate the advantages of digital, continuous pathology to patient specimens. Here we propose translation of the tissue microarray based AQUA algorithms to whole tissue sections, toward the goal of digitally defining expression levels to optimally match biospecific therapies to their targets. Our model system will be breast cancer since there are already 3 biospecific therapies and more are in the pipeline and since core needle biopsy is the most common current method for assessment of breast masses.
Our aims i nclude 1) Translation of the AQUA score to an absolute protein concentration which will require optimization of the image acquisition and construction of a standard curve; and 2) proof of the concept by comparing AQUA analyzed biopsies to current standards and patient outcomes.
Showing the most recent 10 out of 21 publications
