Prediction of chemosensitivity for patients with acute myeloid leukemia, according to expression levels of 28 genes selected by genome-wide complementary DNA microarray analysis.

To identify genes involved in the sensitivity of acute myeloid leukemia (AML) cells to chemotherapy, we monitored gene-expression profiles of cancer cells from 76 AML patients using a cDNA microarray consisting of 23,040 genes. We identified 63 genes that were commonly overexpressed and 372 genes suppressed in AML. Because these genes represent key molecules for disclosing the molecular mechanisms of AML, they may be potential targets for drug development. We also found 28 that revealed different expression levels between good and poor responders to chemotherapy and appeared to be associated with chemosensitivity. On that basis, we developed a "Drug Response Scoring" system that was correlated well with individual sensitivity to an anticancer drug regimen. Among the 44 cases with positive drug-response scores by our definition, 40 achieved complete remission after treatment, whereas the only 3 of the 20 cases with negative scores responded well to the treatment. An ability to predict chemosensitivity should eventually lead to achievement of our goal of "personalized therapy."