Automated Discrimination of Pathological Regions in Tissue Images: Unsupervised Clustering vs. Supervised SVM Classification

Recognizing and isolating cancerous cells from non pathological tissue areas (e.g. connective stroma) is crucial for fast and objective immunohistochemical analysis of tissue images. This operation allows the further application of fully-automated techniques for quantitative evaluation of protein activity, since it avoids the necessity of a preventive manual selection of the representative pathological areas in the image, as well as of taking pictures only in the pure-cancerous portions of the tissue. In this paper we present a fully-automated method based on unsupervised clustering that performs tissue segmentations highly comparable with those provided by a skilled operator, achieving on average an accuracy of 90%. Experimental results on a heterogeneous dataset of immunohistochemical lung cancer tissue images demonstrate that our proposed unsupervised approach overcomes the accuracy of a theoretically superior supervised method such as Support Vector Machine (SVM) by 8%.

[1]  Richard Dybowski,et al.  Neural Computation in Medicine: Perspectives and Prospects , 2000, ANNIMAB.

[2]  Denis Demandolx,et al.  Multiparameter image cytometry: From confocal micrographs to subcellular fluorograms , 1997 .

[3]  Davide Anguita,et al.  Theoretical and Practical Model Selection Methods for Support Vector Classifiers , 2004 .

[4]  Massimiliano Pontil,et al.  Support Vector Machines: Theory and Applications , 2001, Machine Learning and Its Applications.

[5]  Tim W. Nattkemper,et al.  Automatic segmentation of digital micrographs: A survey , 2004, MedInfo.

[6]  M. Larkin.  Software , 2003, The Lancet.

[7]  Federico Girosi,et al.  Training support vector machines: an application to face detection , 1997, Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[8]  Vladimir Vapnik,et al.  Statistical learning theory , 1998 .

[9]  Constantin F. Aliferis,et al.  A comprehensive evaluation of multicategory classification methods for microarray gene expression cancer diagnosis , 2004, Bioinform..

[11]  A. Ruifrok,et al.  Comparison of Quantification of Histochemical Staining By Hue-Saturation-Intensity (HSI) Transformation and Color-Deconvolution , 2003, Applied immunohistochemistry & molecular morphology : AIMM.

[12]  A. Ruifrok,et al.  Quantification of histochemical staining by color deconvolution. , 2001, Analytical and quantitative cytology and histology.

[13]  M. Masotti,et al.  TESTING THE PERFORMANCES OF DIFFERENT IMAGE REPRESENTATIONS FOR MASS CLASSIFICATION IN DIGITAL MAMMOGRAMS , 2006 .

[14]  J J Vaquero,et al.  Applying watershed algorithms to the segmentation of clustered nuclei. , 1998, Cytometry.

[15]  Larry V McIntire,et al.  Automated Selection of DAB-labeled Tissue for Immunohistochemical Quantification , 2003, The journal of histochemistry and cytochemistry : official journal of the Histochemistry Society.

[16]  Helge Ritter,et al.  Cell Detection in Micrographs of Tissue Sections Using Support Vector Machines , 2001 .

[17]  John C. Platt,et al.  Fast training of support vector machines using sequential minimal optimization, advances in kernel methods , 1999 .

[18]  Anil K. Jain,et al.  Algorithms for Clustering Data , 1988 .

[19]  Sergey Ablameyko,et al.  Morphological segmentation of histology cell images , 2000, Proceedings 15th International Conference on Pattern Recognition. ICPR-2000.

[20]  Tarvinder K Taneja,et al.  Markers of small cell lung cancer , 2004, World journal of surgical oncology.

[21]  Luca Benini,et al.  Computer-Aided Evaluation of Protein Expression in Pathological Tissue Images , 2006, 19th IEEE Symposium on Computer-Based Medical Systems (CBMS'06).

[22]  Yu Zong Chen,et al.  Support Vector Machine Classification Of Physical And Biological Datasets , 2003 .

[23]  Gunnar Rätsch,et al.  An introduction to kernel-based learning algorithms , 2001, IEEE Trans. Neural Networks.