Gastrointestinal stromal tumors (GIST) are currently regarded as a heterogenous group of tumors sharing common histological appearance, KIT immunopositivity and supposed origin from tissue progenitor cells capable of differentiation into the phenotype of Cajal interstitial cells. GISTs can be divided according to immunoexpression of the beta subunit of mitochondrial enzyme succinate dehydrogenase (SDHB) to SDHB-positive (encompassing KIT, PDGFRA and NF1 mutated GISTs), and SDHB-deficient GISTs (including Carney-Stratakis syndrome, Carney triad, sporadic pediatric GISTs, and a small subset of sporadic adult GISTs). The individual molecular subtypes differ in biological behavior and in their response to systemic targeted therapy, which is indicated in metastatic GISTs or in tumors with high risk of recurrence. Although several risk-stratification classifications have been developed, strictly defined criteria to identify patients at risk are still lacking. Pharmacogenomics have been successful in designing drugs to overcome not only the primary resistance of GISTs to the action of imatinib (e.g. GISTs with a substitution of Asp842Val in exon 18 PDGFRA or SDHB-deficient GISTs), but also the secondary resistance caused by secondary mutation of a gene encoding either the receptor tyrosine kinase or other molecules involved in the respective signalling cascade. Future directions concentrate on rational molecular targeting for systemic therapy based on complex genetic investigation of the tumor. Peripheral blood is planned to be used as a source of information for genetic events responsible for the secondary resistance of metastatic tumors.