Virtual Microscopy and Grid-enabled Decision Support for Large Scale Analysis of Imaged Pathology Specimens.

Lin Yang(1,2), Wenjin Chen(2), Peter Meer(1), Gratian Salaru(3), Lauri A. Goodell(3), Victor Berstis(4) and David J. Foran(2)

(1)Department of Electrical and Computer Engineering
Rutgers University, Piscataway, NJ 08854, USA

(2)BioImaging Laboratory, Department of Pathology and Laboratory Medicine
UMDNJ-Robert Wood Johnson Medical School
Piscataway, NJ 08855, USA

(3) Cancer Institue of New Jersey, New Brunswick, NJ, 08903, USA

(4) IBM Research, Austin, TX, 73301, USA.

Breast cancer accounts for about 30% of all cancers and 15% of cancer deaths in women. Advances in computer assisted analysis hold promise for classifying subtypes of disease and improving prognostic accuracy. We introduce a Grid-enabled decision support system for performing automatic analysis of imaged breast tissue microarrays. To date, we have processed more than 100,000 digitized specimens (1200x1200 pixels each) on IBM's World Community Grid (WCG). As part of the Help Defeat Cancer (HDC) project, we have analyzed the data returned from WCG along with retrospective patient clinical profiles for a subset of 3744 breast tissue samples and the results are reported in this paper. Texture based features were extracted from the digitized images and isometric feature mapping (ISOMAP) was applied to achieve nonlinear dimension reduction. Iterative prototyping and testing were performed to classify several major subtypes of breast cancer. Overall the most reliable approach was gentle AdaBoost using an eight node classification and regression tree (CART) as the weak learner. Using the proposed algorithm, a binary classification accuracy of 89% and the multi-class accuracy of 80% were achieved. Throughout the course of the experiments only 30% of the dataset was used for training.

IEEE Transactions on Information Technology in BioMedicine, 13 , 636-644, 2009.

Return to Research: Content-based retrieval       Return to List of Publications
Download the paper