Molecular-based choice of cancer therapy: realities and expectations.

Current choice of cancer therapy is usually empirical and relies mainly on the statistical prediction of the treatment success. Molecular research provides some opportunities to personalize antitumor treatment. For example, life-threatening toxic reactions can be avoided by the identification of subjects, who carry susceptible genotypes of drug-metabolizing genes (e.g. TPMT, UGT1A1, MTHFR, DPYD). Tumor sensitivity can be predicted by molecular portraying of targets and other molecules associated with drug response. Tailoring of antiestrogen and trastuzumab therapy based on hormone and HER2 receptor status has already become a classical example of customized medicine. Other predictive markers have been identified both for cytotoxic and for targeted therapies, and include, for example, expression of TS, TP, DPD, OPRT, ERCC1, MGMT, TOP2A, class III beta-tubulin molecules as well as genomic alterations of EGFR, KIT, ABL oncogenes.