|Title||Knowledge discovery in medical imaging data using image analysis and machine learning methods and application to drug development|
|Abstract||High-throughput medical imaging and high-content screening methods have been widely adopted by pharmaceutical and biotechnology companies as well as by many academic labs over the past 20 years (1). With the goal of rapidly identifying potential drugs candidates that affect specific molecular targets high dimensional and increasingly larger data sets are generated (2). More precisely, in the field of digital pathology, the process of assessing digital images of histological slides, is gaining momentum in today's laboratory environment (3). Manual interpretation being a tedious, time consuming, subjective process, allows only a limited statistical confidence due to inherent intra and inter-observer variability. Automated image analysis approaches are an interesting and attractive alternative (4).
We have developed a novel non-supervised cell-nuclei pattern recognition approach validated in breast and ovarian cancer tissue sections and extendable to several other histological classes. We have also introduced a novel supervised approach for the quantification of different T-Cell populations, tumour islet and tumour stroma specific infiltration. Currently, we are extending the concept of single biomarker, semi-continuous data analysis, to the simultaneous analysis of continuous quantitative data from a panel of biomarkers, to provide greater insight into the true prognostic and predictive values of new biomarkers.
Indeed, high throughput digital image acquisition systems are becoming commonplace, and associated image analysis and machine learning solutions are viewed by most as the next critical step for knowledge discovery and automated annotation.
1. Weissleder, R., Pittet, M.J. Imaging in the era of molecular oncology. Nature. 2008 Apr 3;452(7187):580-9.
2. Bleicher, K.H., Bohm, H.J., Muller, K. & Alanine, A.I. Nat. Rev. Drug Discov. 2, 369–378 (2003).
3. Ryan, D., Mulrane, L., Rexhepaj, E., Gallagher, W.M. Tissue microarrays and digital image analysis. Methods Mol Biol. 2011;691:97-112.
4. Rexhepaj, E., Mulrane, L., Penny, S., Callanan, J., Gallagher, W.M. Automated image analysis in histopathology: a valuable tool in medical diagnostics. Expert Rev Mol Diagn. 2008 Nov;8(6):707-25.