Prognostic Value for Quantification of AML1-ETO Fusion Transcripts in Acute Myeloid Leukemia.

Background and Objectives. In spite of the favorable prognosis of patients with acute myeloid leukemia (AML) with AML1-ETO gene rearrangement, relapse is the major cause of treatment failure. We aimed to determine the factors to predict the risk of relapse to quantitate the AML1-ETO fusion transcript levels by real-time quantitative reverse transcription polymerase chain reaction (RQ-PCR).

Design and Methods. Twenty-one patients (14 adults and 7 children) diagnosed as de novo AML with AML1-ETO gene rearrangement were included. The diagnoses were done with the bone marrow (BM) examination, conventional karyotyping, and nested RT-PCR at University of Ulsan College of Medicine and Asan Medical Center in a period from July 2001 to March 2004. The blood and BM specimens were collected at diagnosis or relapse and two or more specimens during and after therapy were included. AML1-ETO chimeric mRNA was quantified by RQ-PCR and represented as AML1-ETO/GAPDH ratio. Medical records were reviewed to evaluate epidemiologic and hematologic findings at diagnosis and clinical outcomes.

Results. From 21 patients, 100 BM and 13 PB samples were obtained. Most patients had CR after 1 or 2 cycles of induction chemotherapy, except one patient. Four adults who relapsed and four children in the first CR states received allogeneic BMT. Thirteen patients relapsed after chemotherapy or BMT. The median follow-up was 20.3 months (range, 14.7 - 37.1 months). The standard curve of RQ-PCR showed a good correlation between the AML1-ETO mRNA and the C_T value (r = 0.99).

At diagnosis AML1-ETO/GAPDH ranged from 2.74 x 10⁻⁴ to 9.79 and these levels were not correlated with PB white cell counts, percentages of blasts, CD56 expression and karyotypes. Higher transcript levels at diagnosis (above the level of 50^th percentile) or decrease of less than 3-logs at the CR achievement revealed increased risk of relapses by Kaplan-Meier analysis of event-free survival (EFS) (P = 0.038 and P = 0.035). The differences of EFS were more evident among adults; All 5 adults with high initial transcript levels relapsed (median CR duration 8.6 months), however, 4 of 9 adults with low transcript levels relapsed (median CR duration 18.9 months, P = 0.001); All 4 adults with less than 3-log reduction at the CR status relapsed but 4 of 9 adults showing more than 3-log reduction relapsed (P = 0.011).

In hypocellular marrow state after induction, 2 patients showed negative conversion with continuing CR but 5 out of 7 patients with persistence of transcripts relapsed (P = 0.343).

RQ-PCR detected the reappearance of AML1-ETO fusion transcripts from PB or BM in the CR states preceding the hematologic relapse by 1 to 3 months in 2 patients. Comparison of PB specimens and BM samples showed similar sensitivities for detecting AML1-ETO fusion transcripts at the same time point (R = 0.917, P = 0.01, Pearson’s correlation).

Interpretation and Conclusions. The initial level AML1-ETO fusion transcripts was valuable to predict the prognosis of AML patients and the regular monitoring of chimeric transcript levels by RQ-PCR from BM or PB samples might be a mandatory tool to predict the relapse risk