Proteomic Prediction of Disease Outcome in Cancer

Better than gene sequencing or quantitative amplification, proteomics tools allow the study of tumor phenotype. Indeed, most current prognostic tests in cancer (carcinoembryonary antigen [CEA], prostate-specific antigen [PSA], CA 19-1, CA 125, alpha-fetoprotein [AFP], etc.) are based on the detection and quantification of single proteins in body fluids. However, a common characteristic of these tests is their relatively low predictive value, so that they are usually complemented with other procedures such as biopsy and/or endoscopy. Recently, improved analytical and bioinformatics tools have driven the attention on pattern recognition approaches rather then single-marker tests for prognostic forecasting. It is expected that predicting metastasization on the basis of tumoral protein patterns will soon be a reality. However, currently available technologies either limit the number of proteins that can be analyzed simultaneously or they are expensive, difficult, and time-consuming. Moreover, the tools adapted for expression proteomics might not be the same as those for prognostic studies that require investigation of protein function over time. We believe that clinical proteomics research designed within a precise clinical and pathology framework should be strongly supported, since many prognostic factors are determined not by the tumor itself, but by the patient, the treatment and the environment.