In the present study, 98 cases of acute leukemias (AL) were diagnosed and classified based on morphologic, immunologic and cytogenetic (MIC) features to assess their diagnostic value in AI.. The results showed that: the conformity rate of cytomorphologic/cytochemical classification with M1C classification was 90.8%. For ALL, the conformity rate of immunologic classification with MIC classification was 95.6% while it was only 70.8% for AML. Of the 48 AML, 10 expressed lymphoidlineage- associated antigens and 8 of 43 ALL. expressed myeloid-lineage-associated antigens. Seven cases were diagnosed as hybrid acute leukemia according to Catovsky's scoring criterion. The clonal chromosomal aberrations were found in 70 cases, of them 46 cases showed characteristic changes including t(9; 22), t(4; 11), t(11; t4), t(8; 12), t(8; 14), 6q-, 9p- and t(15; 17), t(8; 21), inv(16), etc. These data suggested that MIC classification of acute leukemias could provide more diagnostic and biologic information than traditional FAB classification.