Detection of Hybrid Fusion Transcripts, Aberrant Transcript Expression, and Specific Single Nucleotide Variants in Acute Leukemia and Myeloid Disorders with Recurrent Gene Rearrangements

Chromosomal rearrangements producing hybrid fusion transcripts occur in a subset of myeloid disorders including acute myeloid leukemia (AML), myeloid disorders with eosinophilia, as well as in acute lymphoblastic leukemia (ALL). In addition, some chromosomal rearrangements alter gene expression (e.g., EVI1 [MECOM]) without producing hybrid fusion transcripts. The detection of other genetic abnormalities (e.g., IKZF1 partial gene deletions in ALL) as well as specific point mutations (e.g., KIT p.Asp816Val mutation in core binding factor [CBF] + AML, ABL1 p.Thr315Ile mutation in BCR::ABL1 + ALL, NPM1 p.Trp288–p.Trp290 frameshift mutations in cytogenetically normal AML) is important in the pathobiology and clinical management of these disorders. The presence/absence of these genetic abnormalities is increasingly used for subclassification in updated International Consensus (ICC) [1, 2], World Health Organization (WHO) [3, 4], and European LeukemiaNet (ELN) [5] classifications. Some genomic alterations permit diagnosis of specific disorders, while others are seen in more than one entity (e.g., BCR::ABL1 in chronic myeloid leukemia [CML] or AML [6‒9]). Importantly, in the updated ICC [1] classification, specific gene fusions permit a diagnosis of AML when the percent blasts is 10% or more, which is less than the usual morphologic threshold of 20%. In addition, the presence of gene fusions can also guide therapy; for example, BCR::ABL1 may be targeted with tyrosine kinase inhibitors in AML [10], and PML::RARA + acute promyelocytic leukemia is treated differently than other types of AML [11, 12]. In addition to diagnostic and therapeutic implications, genomic alterations impact prognosis; for example, inv(3)(q21.3q26.2) or t(3; 3)(q21.3; q26.2) AML with overexpression of EVI1 (MECOM) has been associated with adverse outcome [13‒15], likewise with EVI1 (MECOM) overexpression in association with KMT2A translocation [16, 17]. IKZF1 alterations in ALL are reported to influence prognosis [18], and KIT mutations (especially exon 17) are reported to have adverse prognostic significance in t(8; 21)(q22; q22)(RUNX1::RUNXT1) AML and, to a lesser extent, in inv(16) AML [19‒24]. Lastly, NPM1 frameshift mutations define a distinct subtype of AML [1, 3] that play a role in the pathobiology of leukemia and also has prognostic significance.

Cytogenetic/fluorescence in situ hybridization (FISH) and reverse transcription PCR (RT-PCR) are most frequently used for gene fusion detection. However, they typically require a panel of several different FISH probes and/or primer sets. Furthermore, rearrangements with heterogenous breakpoints (and highly variable partners) complicate detection and identification of rare/novel fusion gene partners due to limited availability of FISH probes (and the requirement to know gene partner for RT-PCR primer design). These limitations can lead to false-negative results or require multiple rounds of testing. Next-generation sequencing (NGS) technology enables the evaluation of multiple gene targets and several patient samples in a single batch. In order to address the challenge of detecting the above-mentioned wide variety of clinically relevant genetic and molecular alterations, we have validated an RNA-based NGS assay (which leverages anchored multiplex sequencing approach [25‒27]) to detect: (i) hybrid fusion transcripts (independent of the fusion partner), (ii) increased EVI1/MECOM transcript expression, (iii) aberrant IKZF1 (Del exons 4–7) transcripts, (iv) mutations in KIT (in the context of CBF + AML), (v) mutations of ABL1 (in the context of BCR::ABL1 + ALL/CML), and (vi) mutations of NPM1 in AML cases lacking gene fusions. We also report rare and/or novel fusions found in several illustrative samples. Taken together, our findings indicate this RNA-based NGS assay is a useful tool in the diagnostic workup and clinical research of patients with acute leukemia and myeloid disorders with recurrent gene rearrangements.

One hundred nineteen samples were collected from patients being worked up for AML, CML, myeloid neoplasms with eosinophilia, or ALL. Peripheral blood and bone marrow aspirates (collected in EDTA tubes) or fresh/frozen mononuclear cells (>5 million cells) were processed for total RNA extraction. RNA from cell lines (K-562 [ATCC #CCL-243], HL-60 [ATCC #CCL-240], or MV4-11 [DSMZ Leibniz Institute cell line, #ACC 102]) as well as a commercial Seraseq^® Myeloid Fusion RNA Mix (SeraCare; #0710-0407) were used as controls. Total RNA was extracted from samples using QIAamp^® RNA Blood Mini Kit (Qiagen, #52304) according to standard protocol. A positive control was prepared by combining 50% Seraseq^® Myeloid Fusion RNA Mix with 48.5% HL-60 RNA and 1.5% MV4-11 RNA. Data from karyotyping and FISH, relevant RT-PCR or quantitative RT-PCR (qRT-PCR) assays were used for comparison to NGS assay results.

NGS library preparation was performed using 200–250 ng RNA input into the 199-gene FusionPlex^®Pan-Heme Panel (ArcherDX/Invitae, #SK0089) for Illumina^® according to manufacturer’s instructions. RNA integrity was assessed with the PreSeq RNA QC assay (Ct >28.3 cycles indicates suboptimal RNA, and such samples were not further processed). Briefly, the overall workflow includes: cDNA synthesis, end Repair, adapter ligation (sample-specific index, molecular barcode adapters, and universal primer binding site), and PCR (multiplex gene-specific primers and universal primer). Individual libraries are pooled and quantified (Kapa Biosystems, #KK4824), and 1.8 pM of a denatured library pool containing 27% PhiX control (Illumina) was sequenced on NextSeq 550Dx (Illumina) with the NextSeq Mid Output Kit (Illumina). Over 5 million reads per sample (2 × 151 cycles) were obtained. Sequencing data were analyzed with Suite_Analysis_v6.2.7 (ArcherDX/Invitae). The read depth normalization was set to 5 million; otherwise, default settings were used for fusion calls (“strong fusions and oncogenic isoforms filter”; at least 3 unique start sites and at least 5 unique reads), for single nucleotide variant (SNV)/InDel variant calls, and for normalized RNA expression levels. Unique start sites are defined by the software as the number of unique start sites supporting the fusion event. For normalized MECOM RNA expression, the Suite Analysis software uses the unique RNA reads for each of 6 different MECOM gene-specific primer sites, which are each normalized by the geometric mean of the median unique RNA reads from 4 different housekeeping genes, to provide the normalized MECOM expression value for each of the 6 MECOM gene primer sites; these were then added together for each sample and compared between samples.

cDNA was synthesized with SuperScript™ IV VILO™ Master Mix (ThermoFisher Scientific, #11756050). 1.0 µg total RNA was used in 20 µL reaction. PCR was carried out in 25 µL with 1 µL of cDNA template and 0.25 µL of Platinum Taq DNA Polymerase (ThermoFisher Scientific, #10966018) with 40 cycles of PCR. The PCR products were visualized using High Sensitivity D1000 ScreenTape assay (Agilent). Sanger sequencing of the NUP214::ABL1 RT-PCR product was performed. The following oligonucleotide primers were used: NUP214::ABL1_F2 (exon 34): CGG​ATC​ACT​GTC​CCA​ACA​GA; NUP214::ABL1_R1 (exon 2): GAG​CGG​CTT​CAC​TCA​GAC​C. The expected size of the amplified fragment is approximately 140 bp.

RT-PCR was performed, and PCR products were visualized on ScreenTape assay (Agilent). Primers were IKZF1_F (exon 2): TGC​TGA​TGA​GGG​TCA​AGA​CAT​GTC​C and IKZF1_R1 (exon 8): TTC​GTT​CTC​CTT​CTC​GTA​GCT​GGC. The expected size of the IKZF1 RT-PCR fragment without deletion is approximately 1 kb and with deletion of exon 4–exon 7 is approximately 218 bp.

Detection of hybrid fusion transcripts by the NGS assay was assessed for 11 of the most common recurrent hybrid gene fusions: BCR::ABL1 (both p210 and p190 protein isoforms), CBFB::MYH11, RUNX1::RUNX1T1, PML::RARA, ETV6, KMT2A, DEK::NUP214, PDGFRA, PDGFRB, FGFR1 (online suppl. Table 1; for all online suppl. material, see https://doi.org/10.1159/000532085) using 70 samples (56 known positive and 14 known negative) with a variety of blood, bone marrow aspirate, and frozen cell isolate specimens. Accuracy was established by comparing the results to orthogonal testing (i.e., FISH/Karyotype, RT-PCR, or qRT-PCR). The concordance was 55/56 (98%) for positive samples and 14/14 (100%) for negative samples (online suppl. Table 1). The NGS assay did not detect a BCR::ABL1 fusion (for p210 protein isoform) (raw data examined) in 1 patient sample (#5) where the BCR::ABL1 fusion was detected by qRT-PCR at a very low level (0.126% IS [International Scale]). A mixed positive control and cell lines (i.e., K-562, HL-60, or MV4-11) were used to further assess accuracy, and all of the expected fusion transcripts were detected (Table 1; online suppl. Table 2). Overall, these findings demonstrate the accuracy of the NGS assay, which showed a 98% concordance for common, recurrent fusions when compared to orthogonal testing.

The ability of this NGS panel to provide information of clinical relevance beyond that available by cytogenetics/FISH was noted. For example, sample #73 (online suppl. Table 1) demonstrated a CBFB::MYH11 rearrangement detected by the NGS assay and FISH; however, the additional detail provided by NGS regarding the exons involved (exon 5 [CBFB] and exon 28 [MYH11]) helped to explain the low-level fusion transcript signal on CBFB::MYH11 qRT-PCR (which is designed for the most common CBFB::MYH11 fusion transcript (exon 5 [CBFB] and exon 33 [MYH11]); this information helped inform the appropriate interpretation of qRT-PCR results and guided the selection of appropriate follow-up testing. In a separate case (sample #30) (online suppl. Table 1), the patient presented with ALL and eosinophilia and cytogenetics/FISH revealed a 46,XY,t(5;12)(q33;p13.2) that involved ETV6 by FISH; however, the partner gene could not be determined by cytogenetics/FISH. The NGS panel identified the fusion partners as ETV6 and ACSL6, which is important since the differential diagnosis for t(5;12)(q33;p13) rearrangements includes PDGFRB as potential partner, and when present, has therapeutic implications for the patient due to the availability of targeted therapy with tyrosine kinase inhibitors in cases with PDGFRB rearrangements [28, 29].

The analytical sensitivity (i.e., limit of detection) for hybrid fusion transcript detection was assessed by several approaches. First, patient samples positive for BCR::ABL fusion transcripts at different % IS levels (by qRT-PCR) were assessed by NGS and revealed that the NGS limit of detection was approximately 0.17% reads, corresponding to 0.65% IS BCR::ABL p210 (online suppl. Table 1; Fig. 1). Likewise, dilutions of patient samples with various other hybrid fusion transcripts showed a similar overall average limit of detection of 0.6% reads (online suppl. Table 3). Separately, a cell line (MV4-11) positive for KMT2A translocation was serially diluted and revealed an NGS limit of detection of approximately 0.8% cells (0.78% reads) (Table 2; Fig. 1). Lastly, dilutions of the Seraseq^® Myeloid Fusion RNA Mix with HL-60 RNA were performed in order to better understand the relationship between fusion copies and number of reads detected, which showed that at least 80 fusion copies per reaction were required for detection (see online suppl. Table 4). Overall, these findings demonstrate an adequate analytical sensitivity (i.e., limit of detection) for hybrid fusion transcript detection by NGS in diagnostic (i.e., pre-treatment) samples, which is the intended sample type for this NGS assay.

We evaluated the reproducibility of hybrid fusion transcript detection by NGS. For intra-run reproducibility, 4 different samples (each sample with 3 replicates) were tested, and for inter-run reproducibility, samples were tested in 3 different runs (Table 3). Excellent concordance (24/24 total replicates) with low standard error (0.7–1.6%) was seen among all samples. In addition, we analyzed 5 separate runs of the mixed positive control which includes additional fusion transcripts across a range of % reads (Table 1); the standard error was low (0.1–1.4%) for all replicates across the range of fusions/% reads (Table 1). Taken together, these findings demonstrate the excellent reproducibility of the NGS assay.

In addition to detection of hybrid fusion transcripts, we determined that this NGS assay can identify increased MECOM expression in the presence of MECOM (EVI1) rearrangement. We assessed MECOM RNA expression in 43 patient samples (online suppl. Table 5; Fig. 2) across three groups (group 1: 11 patient samples: positive MECOM translocation; group 2: 3 patient samples: negative MECOM translocation/positive KMT2A translocation; group 3: 29 patient samples: negative MECOM/negative KMT2A translocation). Samples positive for MECOM rearrangement demonstrated a mean normalized MECOM expression of 2.2 +/− 0.67 (mean +/− SEM) (range: 0.25–8.6), while cases negative for MECOM rearrangement showed a mean normalized MECOM expression of 0.018 +/− 0.0039 (mean +/− SEM) (range: 1 × 10⁻⁶ to 0.074). Samples with KMT2A rearrangement with increased MECOM expression showed a mean normalized MECOM expression of 10.99 +/− 7.958 (mean +/− SEM) (range: 1.7–34.8). These findings suggest a threshold of approximately 0.1 indicates increased normalized MECOM expression above background. We also assessed the reproducibility of MECOM expression: All repeated samples with increased MECOM expression remained higher than 0.1, while all repeated samples without increased MECOM expression remained lower than 0.1 (online suppl. Table 6; Fig. 2). Taken together, these findings indicate that this NGS assay can identify samples with increased normalized MECOM expression in the context of MECOM rearrangement and KMT2A rearrangement.

We observed aberrant IKZF1 transcripts with deletion of exons 4–7 (Del Ex4–Ex7) in 4 ALL samples (Table 4). To confirm the presence of aberrant IKZF1 transcripts with Del Ex4–Ex7, we performed RT-PCR (Fig. 3; Table 4). The RT-PCR product (approximately 218 bp) corresponding to IKZF1 transcript with Del Ex4–Ex7 was observed in expected cases (which were positive by NGS). Furthermore, the aberrant RT-PCR product was absent in samples which did not show IKZF1 Del Ex4–Ex7 transcripts by NGS. Taken together, these findings indicate that the NGS assay can detect aberrant IKZF1 transcripts with Del Ex4–Ex7. However, since a variety of IKZF1 alterations may occur in acute leukemias, correlation with other molecular testing methods (e.g., multiplex ligation-dependent probe amplification and DNA-based NGS) is required for comprehensive IKZF1 analysis when clinically indicated.

We assessed the ability of this NGS assay to detect selective SNVs known to be significant in the context of hybrid fusion translocation status. These include KIT hotspot mutations (exon 8 and exon 17) in CBF + AML, BCR::ABL1 p.Thr315Ile resistance mutation in BCR::ABL1 + CML/ALL, and NPM1 p.Trp288fs* mutation in normal karyotype AML. This NGS panel demonstrated results concordant with orthogonal myeloid DNA-based NGS panel for all 13 patient samples (online suppl. Table 7) for detection of KIT p.Thr417-D419, p.Asp816, p.Asn822 mutations in the context of CBFB::MYH11 and RUNX1::RUNX1T1 fusions (CBF + AML), with a sensitivity of approximately 10% variant allele frequency (VAF). For ABL1, the NGS panel demonstrated results concordant with orthogonal ABL1 kinase domain mutation sequencing for all 8 patient samples (online suppl. Table 8) with ABL1 p.Thr315Ile and p.Glu255Lys mutations in the context of BCR::ABL1 fusions, with a sensitivity of approximately 15% VAF. Lastly, mutated NPM1 is a distinct category of AML where various NPM1 frameshift variants may occur, most often, in the context of a normal karyotype. We tested 17 patient samples, and the expected NPM1 mutations were detected in all known positive samples and were absent from known negative samples (online suppl. Table 9); a sensitivity of approximately 10% VAF was observed. Taken together, these findings indicate that this RNA NGS assay provides information regarding specific SNV and InDels in KIT, BCR::ABL1 and NPM1, which are important in the context of hybrid fusion translocation status.

During testing, several additional samples were encountered which highlight the clinical utility of this assay; a brief summary of the findings and their clinical relevance is described below.

A follow-up sample for acute leukemia of mixed phenotype (myeloid/B), which at initial diagnosis showed a normal karyotype and FLT3 internal tandem duplication, and had been treated with azacitidine/decitabine, venetoclax, and gilteritinib (leading to complete remission with incomplete hematologic recovery and minimal residual disease [MRD]), demonstrated recurrent AML showing a complex karyotype (46,XY,der(4)t(4;9)(p15.2;q22), der(9)t(1;9)(q21;q22) [3]/46,XY,+1,der(1;15)(q10;q10) [2]/46,XY,+1,add(1)(p13),−18 [2]/46,XY [15]). The genes involved in the chromosomal rearrangements were cryptic by cytogenetic analysis. FLT3 ITD was negative (consistent with eradication of the FLT3 + clone with gilteritinib). Treatment continued over the subsequent months; however, the AML and complex cytogenetic findings persisted. The RNA NGS panel revealed a rearrangement involving NUP214::ABL1 (Table 5, sample #113). The NUP214::ABL1 finding was confirmed through RT-PCR and Sanger sequencing (Fig. 4). The NUP214::ABL1 fusion is a rare fusion, which is potentially targetable; however, medical complications precluded further therapy.

An abnormal karyotype was seen during workup of a chronic myeloid neoplasm that was difficult to further subclassify (molecular studies were negative for JAK2, MPL, and CALR mutations). The karyotypic abnormality (46,XX,t[9;15][q22; q22] [20]) involved PML by FISH, but the partner could not be identified. Cytogenetic analysis was negative for BCR::ABL1, PDGFRA/PDGFRB, and FGFR1 rearrangements. A matched unrelated donor allogeneic stem cell transplant lead to complete remission, but subsequent relapse occurred, with the same cytogenetic finding as at diagnosis. Over time, the disease progressed showing increased blasts and eventually AML, with the same isolated cytogenetic finding. Whole transcriptomic analysis via RNA seq revealed a PML::SYK rearrangement [30], which prompted investigational use of epichaperome inhibitor, in combination with other therapy. As shown in Table 5 (sample 112), the RNA NGS assay described herein also identified the PML::SYK fusion, demonstrating that this current RNA NGS assay can provide clinical molecular laboratories with the ability to identify clinically relevant, novel/rare fusion events, such as PML::SYK.

A sample from a workup for anemia, thrombocytopenia, and leukocytosis revealed B-lymphoblastic leukemia/lymphoma by flow cytometry. The RNA NGS assay revealed a marked increase in CRLF2 expression (sum of normalized CRLF2 expression: 1,080 vs. normal reference: 0.1–0.6), together with a CRLF2 p.Phe232Cys variant (98% VAF) (Table 5, sample 110). NRAS p.Gly12Ala (46% VAF) and SETD2 p.Ser1572TrpfsTer16 (53% VAF) were also seen. These findings permitted classification as B-ALL with Ph-like findings and prompted consideration of tyrosine kinase inhibitor therapy. The RNA NGS assay findings were consistent with separate FISH (which revealed CRLF2 rearrangement) and with separate DNA-based lymphoid NGS panel, which confirmed the CRLF2 p.Phe232Cys variant (and other variants in NRAS and SETD2). This sample demonstrates how this RNA NGS assay has the potential to simultaneously identify several clinically relevant findings in a single assay.

A follow-up sample for CML (on dasatinib) with leukocytosis (including basophilia), thrombocytosis, and anemia revealed 46,XY,t(9;22)(q34;q11.2) [20] and BCR::ABL1 gene fusion (92% of cells) by FISH. However, qRT-PCR studies for BCR::ABL1 revealed only minimal levels of e1a2 (<0.01%) and e13a2/e14a2 (0.002%) BCR::ABL1 fusion transcripts. The RNA NGS assay (Table 5, sample #115) revealed a rare type of BCR::ABL1 fusion transcript (e19a2) and also revealed ABL1 p.Thr315Ile and ABL1 p.Glu255Lys mutations, known to be associated with tyrosine kinase inhibitor resistance. Thus, the RNA NGS assay provided important information by: (i) explaining the discrepancy between qRT-PCR and cytogenetics, (ii) permitting identification of appropriate primers for qRT-PCR follow-up, (iii) identifying ABL1 resistance mutations (which were confirmed by separate ABL1 kinase domain sequencing).

A variety of molecular alterations are frequent in hematopoietic disorders and influence diagnosis, prognosis and patient management. We have validated and implemented a targeted, RNA-based NGS assay capable of detecting a variety of hybrid fusion transcripts; the platform also has the potential to detect aberrant transcript expression and selective SNVs/InDels. Previously, RNA sequencing to detect fusions in solid tumors has been successfully validated and implemented for clinical use [31‒33]. In comparison to prior studies of RNA sequencing to detect fusions in hematologic disorders [26, 27], we have demonstrated: (i) an increased number and type of hybrid fusion transcripts, (ii) the potential to identify aberrant transcript expression (EVI1 [MECOM], IKZF1), (iii) the potential to identify selective “hotspot” SNV/InDel variants (e.g., KIT, ABL1, NPM1), which are important in the context of gene fusion status, and (iv) rare/novel fusions of clinical significance in several illustrative samples.

In terms of gene fusions, the current NGS assay can detect various common hybrid fusion transcripts in a single reaction, as well as identify rare/novel hybrid fusion transcripts and clarify fusion partner genes, when cytogenetics/FISH may be limited/cryptic (e.g., NUP214::ABL1 [34‒42]). Also, by defining the exons involved in a fusion, this assay can inform appropriate exon-specific assay design for follow-up MRD testing by other methods (e.g., qRT-PCR), which are required since the analytical sensitivity (i.e., limit of detection) of this NGS assay for fusions is not adequate for MRD.

With respect to aberrant transcript expression, overexpression of EVI1 (MECOM) has been reported in inv(3) and t(3; 3) rearrangements [43] (which often lack a traditional hybrid fusion transcript) and is a prognostic marker [17, 44‒46]. With this NGS assay, we observed that normalized EVI1 (MECOM) expression in diagnostic patient samples with inv(3)/t(3; 3) MECOM rearrangements was increased. Thus, in diagnostic leukemia samples, this assay could prompt appropriate cytogenetic/FISH for EVI1 (MECOM) (and MLL rearrangement) if EVI1 (MECOM) expression is elevated. IKZF1 partial gene deletions and other IKZF1 abnormalities play a role in ALL [18, 47‒50]. Although some aberrant IKZF1 transcripts showing partial gene deletions may be seen with this NGS assay, multiplex ligation-dependent probe amplification and DNA-based sequencing are required to detect the full spectrum of IKZF1 alterations.

In terms of SNVs/InDels, this NGS assay could identify selective SNVs and InDels, such as KIT exon 8 and exon 17 mutations (p.Thr417_Asp419, p.Asp816, and p.Asn822), which are important in the context of CBF AML (e.g., RUNX1::RUNX1T1, CBFB::MYH11) [23, 51]; however, rare mutations in KIT exon 11 are not covered by this NGS panel. BCR::ABL1 p.Thr315Ile mutations, which are clinically significant and lead to resistance to some tyrosine kinase inhibitors [52, 53], were observed with this NGS assay; nevertheless, separate ABL1 tyrosine kinase domain sequencing is of greater sensitivity, due to the use of nested PCR, and is necessary to detect the full range of ABL1 tyrosine kinase domain mutations. Lastly, AML with mutated NPM1 is a distinct category in the latest ICC and WHO classifications [1, 3]. NPM1 mutations correspond to various types of frameshift insertion/deletion mutations in the c-terminus [54]. This NGS assay could detect NPM1 mutations (regardless of type) at the same time as assessing gene fusion status; indeed, since NPM1-mutated AML cases often have normal karyotype, the NPM1 variant data can inform AML classification and prompt the appropriate mutation-specific qRT-PCR required to monitor NPM1 MRD [55].

The limitations of this RNA NGS assay include the analytical sensitivity (i.e., limit of detection), which necessitates separate assays (e.g., BCR::ABL qRT-PCR, CBFB::MYH11 qRT-PCR, NPM1 qRT-PCR) for MRD testing; this is similar to the conclusion of another group using a similar platform [26]. Therefore, this NGS panel is ideally suited for diagnostic workup samples, not residual disease. Also, as a targeted NGS panel, at least one of the genes present in any fusion must be on this panel to be detected; thus, cytogenetics/FISH analysis is still necessary, especially to detect karyotypic abnormalities which do not generate hybrid fusion transcripts. Likewise, DNA-based NGS is still needed since this RNA-based NGS panel does not adequately cover all of the genes/exons needed for complete SNV/InDel evaluation of the broad variety of genes mutated in acute leukemia and myeloid disorders; in addition, given the analytical sensitivity (i.e., limit of detection) observed (approximately 10–15%) for the specific SNVs/InDels assessed, other more sensitive SNV/InDel methods (e.g., digital PCR, qPCR, nested PCR, etc.) are required for low-level mutation detection (e.g., KIT mutations in systemic mastocytosis, MRD monitoring, etc.).

In sum, our findings indicate that the current NGS assay is a complementary tool, along with other molecular and cytogenetic tests, for diagnostic samples in the clinical workup and research of acute leukemia (and other hematopoietic disorders with recurrent gene rearrangements), since it can detect rare/novel (as well as common) hybrid fusion transcripts and has the potential to identify aberrant transcript expression and selective SNVs/InDels of clinical relevance.

This work was performed as part of a clinical assay validation in the Molecular Hematopathology Laboratory, Department of Pathology, Weill Cornell Medicine. The study was approved by the Weill Cornell Institutional Review Board (Protocols #1007011151 and #1302013582), and written informed consent was obtained from the patient(s) for participation in the study.

The authors have no conflicts of interest to declare.

This work was supported by Weill Cornell Medicine/New York Presbyterian Clinical Genomics funds.

Y.L., K.D., and M.J.K. contributed to the conception of the work, acquisition, analysis, and interpretation of data. J.K., J.T.G., M.O., F.F., K.X., and G.R. contributed to the acquisition and/or interpretation. All authors had the opportunity to participate in drafting the manuscript, provided final approval of the version to be published, and agreed that questions related to the accuracy or integrity of the work will be appropriately investigated and resolved.

The data that support the findings of this study are included in this article and its supplementary material files. Further inquiries can be directed to the corresponding author.