Phenotyping Liver Cancer Registry

"The image processing operations required for local content-based image feature extraction consist of two main tasks: 1) tiling the images in smaller VOIs, typically a small cube, whose size depends on the modality, on the image resolution and on the purpose of the content-based query, and 2) performing feature extraction operations on the VOIs.~The Feature Extraction Engine performs totally unsupervised, automatic and asynchronous extractions of features from the images, organizes and indexes them in a no-SQL database based on unique similarity metric. The results of this phase are a series of clusters of phenotype signatures."

Since the clusters are self-organizing their pathophysiological meaning is not readily apparent and requires further analysis. The characterization of each cluster is performed by analyzing representative samples and their respective correlation with histopathology results. After a series of iterations, the clusters are organized to correlate with distinct tissue subtypes identified by their signature similarity. The final number of clusters is not known a priori and depends on the heterogeneity of the underlying imaging phenotypes.