Patient stratification and determining clinical outcome for cancer patients

A method of generating a scheme allowing classification of a cancer of an individual patient for estimating a clinical outcome for said patient, wherein the method comprises:

• a) identifying subtype clusters of a cancer type based on comparison of a set of gene expression data derived from clinical cancer samples of multiple test subjects suffering from said cancer type;
• wherein clinical samples are divided into subtype clusters of said cancer type based on the similarity of their gene expression data;
• b) correlating the subtype clusters identified under a) to a specific clinical outcome of cancer;
• c) determining an epithelial-mesenchymal transition (EMT) score, comprising:
• i) determining the quantitative expression of at least two markers wherein each of these markers is either characteristic for a mesenchymal cell phenotype or for a epithelial cell phenotype, for determining said epithelial-mesenchymal transition (EMT) signature;
• ii) determining the quantitative expression of the at least two markers in said patient sample and comparing it with the epithelial-mesenchymal transition (EMT) signature to assign an epithelial-mesenchymal transition (EMT) score for said patient sample;
• d) estimating an outcome of a patient suffering from said cancer by assigning the set of gene expression data of said patient sample to any one of the subtype clusters according to a) and b), and by determining the patients epithelial-mesenchymal transition (EMT) score.