3′ Position of the FLT3-LM and Loss of Heterozygosity Are Highly Correlated to Unfavourable Outcome in AML.

The FLT3-length mutation (FLT3-LM) also referred to as FLT3-ITD (internal tandem duplication) belongs to the most frequent genetic changes in AML. LMs cause structural changes of the juxtamembrane domain that disrupt the autoinhibitory conformation of the receptor, promoting constitutive activation of the receptor and downstream effectors. Most studies have shown an unfavourable prognostic impact of this mutation especially in the intermediate karyotype group. In addition, some studies have suggested that differences within the total FLT3-LM-positive cohort exist: not only the LM ratio in comparison to the wild type allele but also the length of the inserted/duplicated segment were reported to have impact on outcome. As some previous studies especially those on size of the mutation were of limited numbers we wanted to further work out differences of FLT3-LM with respect to mutation status, ratio, length and position of the mutation in a cohort of 3365 AML pts that were all intensively treated (65% in AMLCG trials, 35% in comparable AML standard protocols). 689 case (20.5 %) were positive for the FLT3-LM. Ratio and length of the mutation were assessed by gene scan analysis. Using Cox regression analysis we show that increasing FLT3-LM/WT ratio has a significantly unfavourable influence on OS and EFS (p<0,0001, each). Three groups were established according to LM/WT ratio: 1) simple mutated up to a ratio of < 0.99 2) ratio > 1 reflecting LOH 3) unmutated. We found that OS and EFS were not significantly different between the unmutated and simple mutated cases. In contrast, LOH cases had a shorter survival than unmutated and simple mutated cases (p=0.002 and p=0.004, respectively for OS; p=0.002 and p=0.006, respectively for EFS). This was similar when the analysis was restricted to the intermediate karyotype group only (822 unmutated, 240 simple mutated, 87 LOH) (p=0.003 and 0.006, respectively for OS and p<0.0001 and p=0.005 for EFS). Subsequently, 611 individual FLT3-LMs were further characterized by nucleotide sequencing for the exact position and sequence of the inserted fragment. Of these 611 FLT-LMs 387 where simple duplications of 3–153 (median 51) nucleotides (67.6%). 223 mutations (32.4 %) had sole or additional insertion of 1–81 extranucleotides of unknown origin. Startposition of the mutation was between nt 1702 and 1848 (according to HSFLT3RTK) and endposition between nt 1702 and 1848. Unlike previously suggested, we did not found an impact of the LM-size. However, the detailed analysis of insertion position revealed a shorter OS in cases with more 3′ than those with 5′ positions for the total (p=0.040) as well as for the intermediate karyotype group (p=0.009). Detailed analysis of 6 different regions, containing 59–73 pts each, revealed worst prognosis for the region containing nt 1783–1848 (62 cases) near to or within the TK1 region, indicating that disruption of this region either may most effectively disrupt the autoinhibitory conformation of the receptor or/and is activating the TK1 region. However, in multivariate analysis start position did not come out as independent factor for OS in contrast to age (p<0.001), WBC (p=0.017), and FLT3-LM/WT ratio (p=0.005). These data further support the hypothesis that biologic and clinical differences exist between the different individual FLT3-LM types. In detail 1) FLT3-LM per se is not predictive for survival but only LOH of the WT-allele. 2) No effect of length of the mutation was found. 3) Mutations starting next to or within the TK1 seem to be prognostically unfavourable compared to more 5′ mutations.