Abstract 1725

Children up to the age of 5 years with trisomy 21 (T21, Down Syndrome) are at a 400-fold excess risk of developing myeloid leukemia (ML-DS). ∼5% of newborns with T21 develop a transient leukemia (TL). The megakaryoblastic phenotype by morphology and immunophenotyping is similar in both leukemias. Mutations in hematopoietic transcription factor GATA1 gene leading to expression of N-terminal truncated protein (GATA1s) have been detected in almost all TL and ML-DS patients and is the diagnostic genetic hallmark of these diseases.

Aims: Fast and accurate identification is required to:

  1. confirm the diagnosis of TL or ML-DS

  2. confirm the diagnosis of a GATA1s positive leukemia in children with no or little stigmata of Down Syndrome (T21 mosaic)

  3. monitor minimal residual disease (MRD)

  4. determine the pattern of GATA1 mutation in TL and ML-DS.


Here we report the largest cohort of children (n=229) with TL (n=129) and ML-DS (n=100). The blast percentage of blasts were significant different (TL 43±3% vs. ML-DS 29 ±2%, p<0.03).


The GATA1 mutation screening have been performed in two laboratories, the central reference of the AML-BFM Study Group (Hannover, Germany; TL n=90, ML-DS n=63) and at the Weatherall Institute of Molecular Medicine (Oxford, UK; TL n=39, ML-DS n=37). The AML-BFM Lab conducted direct sequencing. If this failed, sequencing was repeated with sorted blasts. If the result remained negative, subcloning of the blasts was performed (21 out of 137 patients). The Oxford lab screened all samples by WAVE and direct sequencing. The lower limit of blasts which allowed for successful detection of a GATA1 mutation was 2%.


GATA1 mutations were identified in 125 out of 129 (96%) newborns with TL and in 97/100 (97%) children with ML-DS. In 99% of cases GATA1 mutations were detected in exon 2; only in 2 cases were exon 3 mutations identified. GATA1 mutation were identified in 13 children with Down mosaic and acute leukemia (TL n=8; ML-DS n=5). The detection of GATA1 prevents intensive chemotherapy in newborns with TL and allowed reduced intensity chemotherapy to be administered in infants with ML-DS.

The mutations are diverse: deletions (37%), point mutations (24%), duplications (23%) and insertions (16%). With exception of substitutions, which were uniquely detected in TL (n=2; 1.6%), no differences between TL and ML-DS have been observed. Mutations were predicted to result in a stop codon(66%), affect splicing (16%), produce a frameshift that produced a subsequent stop codon (7%), or alter the start codon (3%). No differences in these predicted outcomes was present between TL and DS-ML.


Rapid detection of GATA1 mutations is possible in almost all children with T1 and mosaic T21 who develop TL or ML-DS with these approaches, even in samples where the blast count is as low as 2%. Mutation detection and conformation of the correct diagnosis is critical to ensure appropriate therapy is administered and to allow patient specific MRD monitoring.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.