Abstract
BACKGROUND AND OBJECTIVES: To refine cytogenetically based risk-stratification in acute myeloid leukemia (AML). DESIGN AND METHODS: Stratification was improved by combining cytogenetics and day 16 bone marrow blasts and by subdividing unfavorable cytogenetics. The new score identifying five prognostically different groups was developed in 321 patients (AMLCG 1992 trial) and subsequently validated in 680 patients (AMLCG 1999 trial). RESULTS: Subgroups defined were: 1) favorable cytogenetics (t(8;21), inv(16)); 2) intermediate cytogenetics (normal karyotype, other abnormalities not rated favorable or unfavorable) and day 16 blasts <10%; 3) intermediate cytogenetics and day 16 blasts >or=10%; 4) unfavorable cytogenetics (-5/5q-, -7/7q-, 3q21q26 aberrations, 11q23 aberrations, 12p-, 17p-) excluding complex aberrations; 5) complex aberrant karyotypes (>or=3 aberrations). In AMLCG 1992 patients significant differences were observed with regard to complete remission (CR) rate (82%, 83%, 58%, 76%, 53%), persistent leukemia (PL) rate (7%, 8%, 33%, 14%, 31%), median event-free survival (EFS; 25, 14, 5, 6, 2 months), median overall survival (OS; not reached, 26, 12, 14, 6 months), and median relapse-free survival (RFS; 26, 19, 13, 8, 4 months). The prospective validation of the score proved its significant power (AMLCG 1999 cohort) with regard to CR (63%, 65%, 51%, 45%, 35%), PL (17%, 18%, 40%, 35%, 48%), median EFS (14, 7, 3, 2, 2 months), median OS (25, 15, 12, 6, 4 months), and median RFS (not reached, 15, 10, 8, 5 months). INTERPRETATION AND CONCLUSIONS: This new prognostic score provides a highly valuable tool for future clinical trials in AML focusing on distinct and subgroup-specific treatment effects.