New Challenges in the Identification of Gastrointestinal Stromal Tumors and Other Possible KIT-Driven Tumors

The successful treatment of patients with metastatic gastrointestinal stromal tumors (GISTs) with a new KIT tyrosine kinase inhibitor, imatinib (STI571, Gleevec, Novartis, Basel, Switzerland), has made it important to specifically identify GISTs,1-3 and it also has raised the hopes that other sarcomas could be treated similarly. On the other hand, patients whose sarcomas have other pathogenetic mechanisms would not benefit from such a treatment, and they should not be exposed to its potential side effects and cost. Imatinib, originally found to inhibit the cytoplasmic/nuclear tyrosine kinase of the abl oncogene activated in chronic myeloid leukemia, was subsequently also found to inhibit the receptor tyrosine kinases of KIT and the closely related platelet-derived growth factor receptor, PDGFR.4,5 All GISTs express the KIT protein, and a great majority of them are believed to be driven by oncogenic mutations of KIT, leading to its constitutional activation (phosphorylation), which in normal circumstances takes place upon binding of the ligand, the stem cell factor. The activating (gain-of-function) mutation perpetuates the KIT signal and the downstream phosphorylation cascade in the signal transduction pathway, ultimately leading to activation of the cellular proliferation.6 Such oncogenic conversions of growth factor receptors are a common theme in the molecular pathogenesis of many cancers, in which (tyrosine) kinases become activated by mutations, translocations, or gene amplifications. A majority of GISTs have been shown to have KIT-activating mutations,7-9 and all GISTs are immunohistochemically positive for KIT (CD117), which along with compatible histologic findings, ie, highly cellular spindle cells or epithelioid mesenchymal tumor of the gastrointestinal tract, are the defining features for these tumors.10-12 Immunohistochemical detection of KIT is the key in the diagnosis of a GIST. Most GISTs are strongly and uniformly positive for KIT, with a membrane Golgi zone–like or cytoplasmic staining pattern. Also other nonepithelial and mesenchymal tumors have been variably reported as positive for KIT, among them 50% of angiosarcomas, Ewing sarcomas, and clear cell sarcomas13,14 and melanoma. The KIT positivity of the latter seems to reflect KIT positivity normally present in melanocytes, but the positivity tends to be less consistent in metastatic tumors.15 One group reported common (75%) KIT positivity in desmoid tumors,16 but this was not reproducible with another antibody.17 In this issue of the Journal, Hornick and Fletcher18 present data on KIT immunoreactivity of 365 soft tissue tumors other than GISTs. This study used a polyclonal KIT antibody (DAKO, Carpinteria, CA) without an epitope retrieval modality; previous studies have mainly used another polyclonal KIT antibody (SC-168, Santa Cruz Biotechnology, Santa Cruz, CA) with heat-induced epitope retrieval. The authors conclude that KIT immunoreactivity is relatively limited in tumors other than GISTs. However, they found 25% of angiosarcomas, 20% of Ewing sarcomas, and 20% of metastatic melanomas positive, along with occasional extraskeletal myxoid chondrosarcomas (10%), perineuriomas (20%), and low-grade fibromyxoid sarcomas (10%). The frequencies of KIT positivity among tumors such as Ewing sarcoma and angiosarcoma were lower than previously reported, and the authors suspected that the higher percentages found by others could be a result of background staining.18 Alternatively, KIT detection without epitope retrieval could result in lower detection sensitivity and a lower percentage of positive tumors. Independent data