Abstract: In one embodiment, the invention provides methods for predicting a clinical outcome of a patient's neoplastic disease comprising: (a) determining a predictor value algorithmically using patient sample values for (1) at least one tumor marker or at least one immune marker, and (2) at least one marker that is (i) an extracellular matrix (ECM) marker (ii) a marker that is indicative of extracellular matrix synthesis (fibrogenesis), or (iii) a marker that is indicative of extracellular matrix degradation (fibrolysis); and (b) predicting the clinical outcome of the neoplastic disease by evaluating the predictor value.

Abstract: The invention relates to a method (i) for detecting a carcinoma, especially an adenocarcinoma, preferably a gastrointestinal carcinoma and more preferably a colorectal carcinoma, (ii) for predicting metastases, preferably liver metastases, depending on a primary colon carcinoma and/or (iii) for predicting the response of metastases to a 5-fluorouracil-containing chemotherapy. The inventive method comprises determining a gene expression profile of 120 marker genes or a selection thereof.