Role of gene-expression profiling in chronic myeloid leukemia.

Gene-expression profiling enables disease classification and risk stratification, and provides important insights into possible pathogenetic mechanisms. The clinical management and prognosis of chronic myeloid leukemia (CML) has substantially changed after the introduction of targeted therapies, such as imatinib and the second-generation tyrosine kinase inhibitors nilotinib and dasatinib. Although exact characterization of CML pathogenetics has been performed by showing the causal pathogenetic relevance of the reciprocal translocation between chromosomes 9 and 22 for CML development, the disease still exhibits a marked clinical and biological heterogenicity. Thus, prognostic scores for a more exact disease classification, as well as for prediction of response to tyrosine kinase inhibitor therapy are warranted, especially because scores established within the interferon era have lost their prognostic value when applied to patients treated with imatinib. Gene-expression profiling has been proven to represent a powerful tool for early identification of nonresponders to cancer therapy. Several profiling studies in CML have been reported thus far. However, the available data are inconsistent, which is mainly due to technical reasons, such as the use of various different chips, different biostatistical algorithms for data analysis and, most importantly, the use of various different cellular sources (i.e., mononuclear cells from peripheral blood, whole bone marrow cells or selected bone marrow-derived stem/progenitor cells). This review will summarize the recent literature on gene-expression profiling for CML classification and response prediction.