ELN-DAVID recommendations for NGS-based FLT3-ITD MRD testing for patients with acute myeloid leukemia

Abstract

FLT3-internal tandem duplication measurable residual disease testing for patients in remission from acute myeloid leukemia is now recommended by the recently updated clinical standard-of-care guidelines. This companion technical note provides important laboratory and clinical recommendations regarding such testing.

Next-generation sequencing (NGS) FLT3-internal tandem duplication (ITD) measurable residual disease (MRD) testing for patients in remission from acute myeloid leukemia is now recommended by the recently updated clinical standard-of-care guidelines. Hourigan et al provide important laboratory and clinical recommendations regarding such testing, focusing on how and when to do it as well as how to interpret the results.

TO THE EDITOR:

There is now substantial evidence that detection of persistent FLT3-internal tandem duplication (ITD) mutation in patients who are in morphological complete remission after intensive induction therapy for acute myeloid leukemia (AML) is associated with an increased risk of subsequent relapse and death compared with those without FLT3-ITD measurable residual disease (MRD^{1, 2, 3, 4, 5, 6, 7, 8, 9, 10}; Table 1). Because of its superiority over flow cytometry,^3,6 DNA-based next-generation sequencing (NGS) FLT3-ITD MRD testing is now recommended by both national¹¹ and international¹ clinical guidelines for prognostic as well as predictive purposes. Assays designed for diagnostic FLT3-ITD testing do not typically have sufficient sensitivity for MRD monitoring. Although technology evolves rapidly and individual laboratories may choose to innovate and optimize based on their own needs, a consensus on current best practices for FLT3-ITD NGS-MRD testing was determined by the authors at a meeting held in 2025 (Box 1).

Table 1.

Recent evidence of prognostic association between FLT3-ITD detection in remission and clinical outcome of intensively treated patients with AML (with or without FLT3 inhibitors)

Cohort	Patients, n	Remission	Collection time	Sample	Test	DNA used (ng)
PreMEASURE2 (111 sites CIBMTR, 2013-2019)	608	CR1	CR1 pre-alloHCT (≤100 d)	PB	Custom/Archer	500
ReMEASURE6 (subset of PreMEASURE)	537	CR1	CR1 pre-alloHCT (≤100 d)	PB	Invivoscribe	700
UK-Aus Retrospective4 MRC AML17 (2009-2014) Alfred/Royal Melbourne Hospital/Peter Mac (2010-2020)	104	CR1 (n = 89),CR2 (n = 15)	CR pre-alloHCT (median, 27.5 d before transplantation; range, 1-87)	BM (n = 92),PB (n = 12)	Modified getITD	500
HOVON-SAKK3 (multicenter prospective phase 3 trials)	161	CR1	Diagnosis, CR1 (pre-alloHCT, if received transplantation)	BM (n = 156),PB (n = 5)	Exon 14 PCR-NGS	100 and 500
QuANTUM-First8,12 (phase 3 RCT quizartinib vs placebo)	308	Composite CR1	Time of response assessment during induction in composite CR1	BM	Exon 14-15 PCR-NGS	—
BMT-CTN 15065,13 (MORPHO, phase 3 RCT gilteritinib vs placebo after alloHCT)	356	CR1	CR1 pre-alloHCT, prerandomization, and at 3, 6, 12, 18, and 24 mo after randomization	BM	Invivoscribe	700
ALLG AMLM1610 (phase 2 RCT of sorafenib vs placebo added to intensive induction)	74	CR1	CR/CRi after induction	BM	Modified getITD	500
AMLSG16-109 (midostaurin)	142	CR1	After 2 cycles + follow-up	BM (n = 497),PB (n = 66)	getITD	50
PrECOG 090514 Open-label phase 2 RCT (gilteritinib vs midostaurin)	177	CR1	After induction (1-2 cycles), after first consolidation	BM	Invivoscribe	700

alloHCT, allogeneic hematopoietic cell transplant; BM, bone marrow; CR, complete remission; CRi, complete remission with incomplete hematologic recovery; PB, peripheral blood; RCT, randomized controlled trial.

Box 1. Key recommendations for reliable FLT3-ITD NGS-based MRD assay (2025).

DNA input: ideally at least 700 ng (≥500 ng, ∼75 000 cell equivalents) Source: BM or PB (BM may have 5-10× greater FLT3-ITD burden than PB) Method: PCR-NGS (or another well-validated test with limit of detection [LOD] comparable to PCR-NGS) Error-correction/duplex: not required PCR cycles: 8 + 20 (or 25) Sequencing: 2 × 250 or 2 × 300 cycles (or another NGS method to reliably detect FLT3-ITDs >200 bp) LOD: at least 1 × 10−4 (0.01%, MRD4)∗ Quality control: controls and methods to identify contamination Bioinformatics: predefined validated pipeline, which should be able to call de novo FLT3-ITDs Report: report all identified FLT3-ITDs including variant allele fraction at each time point Current examples: commercial (Invivoscribe5,6,13,15) and noncommercial (modified Get-ITD4,9,10,16, 17, 18)

^∗Increased relapse risk has been reported for those testing positive even beneath this threshold.^5,6

Alt-text: Box 1

Several approaches are available for ultrasensitive detection of FLT3-ITD in patients with AML in remission.¹⁹ However, most clinical data to date are based on a similar testing strategy (Table 1) involving polymerase chain reaction (PCR) amplification of DNA using primers^16,17 that target the juxtamembrane region (flanking exon 14 and/or 15) of the FLT3 gene, followed by deep DNA sequencing.^{7,16, 17, 18,20} The limit of detection, defined as the lowest concentration that can be reliably detected above assay background, of an NGS-based AML MRD assay, is influenced by multiple factors, including the size and location of the FLT3-ITD.²¹ However, the amount of sample DNA used for library preparation is a critical limiting constraint. Given that ∼660 ng of genomic DNA is present within 100 000 human cells, together with imperfect assay efficiency and Poisson distribution sampling considerations, we recommend a genomic DNA input of ≥500 ng and, ideally, 700 ng for high-sensitivity FLT3-ITD NGS-MRD assays. In addition, laboratories should communicate with the ordering physician if the specimen submitted for testing is inadequate.

FLT3-ITDs vary widely in length, ranging from 3 to >200 base pairs (bp). Accordingly, FLT3-ITD NGS-MRD assays should be validated to detect the full spectrum of FLT3-ITD lengths and insertion sites. Recent studies of FLT3-ITD NGS-MRD have reported ITDs of up to 228 bp (median, 51 bp),⁶ consistent with the FLT3-ITD length distribution observed at diagnosis by capillary electrophoresis in adult patients with AML.²² Laboratories should be aware of the risk of false-negative MRD results associated with very long FLT3-ITDs and should select an appropriate sequencing strategy to mitigate this risk (Box 1).

Two general PCR approaches are commonly used: either 25 cycles of PCR using primers coupled with sequencing adapters and indexes,⁷ or 8 cycles of PCR with FLT3-specific primers followed by additional 20 to 25 cycles to add sequencing adapters and indexes.¹⁶ Because of index-hopping risks, unique dual indexes should be used to prevent reads from being assigned to the wrong sample. Given the low background false-positive error rate of FLT3-ITD detection using NGS, error-correction approaches are not required but unique molecular index use could be considered to improve low-level molecular quantification. PCR is a powerful technique but is highly susceptible to contamination. Diagnostic or overt relapse samples should not be prepared or sequenced alongside remission samples for MRD testing. Negative-control samples should be included with each batch, and low-level positive samples may be included either with each batch or as part of test runs. Capillary electrophoresis may be used as an initial screen to exclude strongly positive samples which have a higher risk of causing contamination.

The bioinformatic strategy should be predefined and capable of de novo detection of FLT3-ITDs rather than relying solely on tracking FLT3-ITDs identified at initial diagnosis.¹⁷ The analysis report should also include information on the DNA input amount, sample source, and the history of detected FLT3-ITD sequence(s) and allele fractions at initial diagnosis and across all MRD time points tested. Typically, at least 3 identical mutant reads are required to make a positive FLT3-ITD call. All FLT3-ITD sequences identified should be reported at each time point, ideally also with a comment on any FLT3-ITD sequences previously detected, including those present at initial diagnosis in that patient.

ELN-DAVID does not make commercial recommendations but acknowledges that at least 1 network of internationally harmonized and accredited commercial reference laboratories (United States, Europe, China, and Japan) offers a commercial FLT3-ITD NGS-MRD testing service that meets all previously described criteria.¹⁵ Laboratories should ensure that implementation of FLT3-ITD MRD testing complies with applicable local economic, regulatory, and legal requirements.

In addition to technical laboratory factors (Box 1), several important clinical considerations are covered elsewhere,¹ including the clinical time point of testing, planned posttest therapeutic interventions (if any), sample type (sample source and hematopoietic recovery), and the intended clinical use of positive or negative test results. For example, detection of FLT3-ITD in DNA from either blood or bone marrow aspirate is associated with an increased risk of relapse and death in patients with AML in first complete remission in multiple studies (Table 1). Although direct comparative studies using paired sample sources from the same patient at the same time point are limited, 2 recent studies of large cohorts of adult patients with AML showed FLT3-ITD MRD rates of 33% in blood and 46% in bone marrow samples using the same highly sensitive test before transplantation.^5,6 Bone marrow samples are likely to have higher FLT3-ITD burden on average,¹² thereby providing greater confidence in a negative result. However, as has been seen with mutant NPM1 MRD, testing marrow during treatment will likely identify more patients as test positive compared with blood but at the expense of a higher clinical false-positive rate, particularly when performed before 2 cycles of intensive therapy.¹² Therefore, FLT3-ITD NGS-MRD testing should be validated for the specific clinical context in which it is used.

Beyond such landmark assessments, evidence remains limited to support the selection of optimal timing or sample source for surveillance monitoring.¹ However, FLT3-ITD testing should not be used as the sole MRD surveillance modality. In patients comutated for FLT3-ITD and NPM1 at diagnosis, mutant NPM1 MRD monitoring may be considered as a guide to disease status.^{23, 24, 25} Serial flow cytometry may be informative during the period of highest relapse risk.¹

The optimal approach to FLT3-ITD NGS-MRD testing in patients with AML treated with less-intensive or targeted maintenance therapy remains to be determined.

The fast kinetics of FLT3-ITD AML relapse, together with the need for MRD testing results to inform clinical decision-making, require that turnaround time from sample receipt to result reporting be as short as possible and not exceed 3 weeks. Considerations related to cost, batch size, and test reimbursement are beyond the scope of this guidance; however, it should be acknowledged that FLT3-ITD NGS-MRD testing should be accessible for repeated use throughout a patient’s disease course to identify patients with impending relapses and inform management decisions.

Conflict-of-interest disclosure: A.H.W. serves on advisory boards and reports consulting roles for Johnson & Johnson, AbbVie, Servier, Gilead, Molecular Partners, BeOne, Astellas, Bristol Myers Squibb (BMS), Pfizer, Geron, and Syndax; has received research funding to the institution from Novartis, Storm Therapeutics, AbbVie, Servier, Johnson & Johnson, Syndax, Mendus, Dimericon, and AstraZeneca; and is an employee of Walter and Eliza Hall Institute (WEHI) and receives financial benefits. The WEHI received milestone and royalty payments related to the development of Venetoclax. Current and past employees of WEHI may be eligible for financial benefits related to these payments. C.S.H. has received research support from the Foundation of the National Institutes of Health Biomarkers Consortium and Illumina; travel support from Invivoscribe; and served on advisory boards for Astellas, Bio-Rad, Janssen, and Syndax. C.T. is co-owner, chief medical officer, and chief executive officer of AgenDix GmbH; and has received honoraria from Novartis, Astellas, and Jazz Pharmaceuticals. L.W.D. reports honoraria from Bio-Rad and travel support from Invivoscribe. I.S.T. reports honoraria from Pfizer, Jazz Pharmaceuticals, Novartis, BMS, Praxhub, and Neovation. J.O. reports honoraria from Astellas, Jazz Pharmaceuticals, and Pfizer. M.H. reports institutional research funding from AbbVie, Bayer Pharma AG, Jazz Pharmaceuticals, Glycostem, Pinotbio, Servier, and Toray; honoraria from Astellas, Daiichi Sankyo, Otsuka, Pfizer, Servier; and consultancy for AbbVie, Ascentage Pharma, Astellas, Daiichi Sankyo, Janssen, Miltenyi Biotec, and Servier. P.J.M.V. received honoraria from Astellas, Daiichi Sankyo, and Servier. The remaining authors declare no competing financial interests.