Background: FLT3 internal tandem duplication (ITD) is frequently detected in AML patients and is an independent predictor of unfavourable outcome, while secondary point mutations in the FLT3 tyrosine kinase domain (KD) are common causes of acquired clinical resistance to FLT3 inhibitors, such as AC220 and Sorafenib.

Technologies allowing massively parallel, ultra-deep sequencing (UDS) are currently being evaluated in diagnostic settings since they may conjugate throughputness, sensitivity and accurate quantification of mutated clones.

Aim: Since recent whole genome sequencing studies have suggested that FLT3 ITD may evolve from small subclones undetectable at diagnosis by routine PCR, we tested the ability of an UDS strategy for FLT3 mutation screening to highlight small clones harbouring ITD mutations. Furthermore we evaluated if an UDS strategy could highlight in AML patients treated with FLT3 inhibitors emerging clones harbouring critical mutations, anticipating the development of drug-resistance.

Methods: 886 AML patients were analyzed in Seràgnoli Institute of Bologna between 2002 and 2013 for a panel of genetic alterations, including FLT3. For our purpose, we retrospectively analyzed five AML (four CN-AML and one t(3;3) AML) who were found negative for FLT3 ITD- at diagnosis by conventional PCR and Sanger Sequencing, but were then found FLT3 ITD+ during follow-up at relapse or disease progression and ten AML (five FLT3+ and five FLT3-) treated with the FLT3 inhibitor AC220.

In order to reconstruct the dynamic of mutation emergence, we performed a longitudinal re-analysis of RNA samples with UDS on a Roche GS Junior. UDS achieved a lower detection limit between 0,1% and 1%, depending on the relative number of sequence reads per sample obtained in each run.


886 AML patients were analyzed for a panel of genetic alterations including FLT3 ITD and TKD and 239 of 886 (27%) were FLT3+. In particular 256 were cytogenetically normal (CN) AML and of these 66 (25,8%) were FLT3+ and 192 were FLT3-.UDS strategy revealed that all the CN-AML analyzed already carried at diagnosis a small clone FLT3 ITD+ (allelic ratio 0.2-2%), that increased over time during follow-up, while in the t(3;3) AML the ITD clone emerged only at disease progression. In the three CN patients treated with chemotherapy the ITD+ was a minor clone until complete remission (CR), while after relapse the ITD+ clone expanded; one of these patients carried two rare ITD clones at diagnosis and only one became dominant at relapse, likely through loss of heterozygosity (LOH) of the mutated allele. In one CN patient who was treated only with FLT3 inhibitor AC220 the allelic load of the mutated clone increased over time before treatment and then followed the dynamic of the disease (regression at complete remission). For the five AML FLT3- treated with AC220 we didn’t find any novel FLT3 mutation after treatment, while for the five AML FLT3-ITD+ analyzed, we were able to follow the allelic load of the FLT3-ITD+ clone during treatment and the appearance in two patients of novel TKD mutations after treatment that are able to confer drug resistance (allelic ratio 2-55%).

Conclusions: The high sensitivity of UDS technology allows detection of emergence of mutated clones earlier than conventional methods: this is a precious tool to find small clones FLT3 ITD+ that may evolve over time and worsen the prognosis of otherwise good prognosis CN-AML patients and to better calibrate therapy for these patients. Furthermore the prognostic value of determining the presence of FLT3-ITD by UDS is stronger than conventional methods because of the possibility to determine the ratio of mutated versus wild-type allele, the length and the size of insertion within a single analysis. In the setting of FLT3 inhibitors, UDS gives advantage in monitoring MRD by determining exactly the allelic load of mutated FLT3 ITD clones before and after treatment and high sensitivity in highlighting emerging mutated clones during treatment that may confer resistance, giving possibility to switch eventually to other inhibitors before coming out of overt clinical resistance to therapy. For monitoring patients treated with FLT3 inhibitors with UDS we will go on by screening AML treated with Sorafenib, to follow the emergence of any novel critical mutation during treatment.

Acknowledgments: ELN, AIL, AIRC, PRIN, progetto Regione-Università 2010-12 (L. Bolondi), FP7 NGS-PTL project.


Martinelli:Novartis: Consultancy, Speakers Bureau; BMS: Consultancy, Speakers Bureau; Pfizer: Consultancy; ARIAD: Consultancy.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution