Cryptic KMT2A::AFDN Fusion Due to AFDN Insertion into KMT2A in a Patient with Acute Monoblastic Leukemia

Abstract

Background: KMT2A rearrangements occur in ~10% of acute myeloid leukemia (AML) cases and are critical for classification, risk stratification, and use of targeted therapy. However, insertions involving the KMT2A gene can evade detection using chromosomal analysis and/or fluorescence in situ hybridization (FISH). Methods: We present a case of a 22-year-old woman with acute monoblastic leukemia harboring a cryptic KMT2A::AFDN fusion identified by RNA sequencing. Initial FISH showed a 3′ KMT2A deletion, while conventional karyotyping and the automated bioinformatic pipeline for optical genome mapping (OGM) did not identify the canonical translocation. Results: To resolve these discrepancies, metaphase KMT2A FISH (break-apart fusion probe) was performed to assess whether KMT2A was translocated to another chromosome. However, the results did not support this possibility. As the fusion signal remained on the normal chromosome 11, with the 5′ KMT2A signal localized to the derivative chromosome 11. A subsequent manual review of the OGM data revealed a cryptic ~300 kb insertion of AFDN into the 3′ region of KMT2A, reconciling the discrepancies between chromosomal analysis, FISH, and RNA fusion results. Conclusions: This case highlights the importance of integrating multiple testing modalities with expert review when there is a discrepancy. Our findings emphasize the need for a comprehensive approach to genomic assessment to enhance diagnostic accuracy and guide therapeutic decision-making.

1. Introduction

KMT2A rearrangements are identified in approximately 10% of cases of acute myeloid leukemia (AML) that occur across all age groups and represent distinct genetically defined types of AML. KMT2A (lysine methyltransferase 2A), formerly known as MLL, has been shown to have over 100 identified fusion partners [1]. Among these, the seven most frequent fusion partners—MLLT3 (30%), MLLT10 (19%), ELL (10%), AFDN (8%), MLLT1 (4%), EPS15 (1%), AFF1 (1%), and partial tandem duplication (10%)—account for approximately 80% of KMT2A recombination events in AML [1]. KMT2A rearranged AML is associated with a poor prognosis, with a notable exception of t(9;11)(p21;q23)/KMT2A::MLLT3, which confers an intermediate prognosis [2]. KMT2A encodes a histone methyltransferase that regulates gene expression, and its rearrangements result in the upregulation of HOX family genes, driving leukemogenesis via arrested differentiation [3]. Targeting the KMT2A menin complex with menin inhibitors has been shown to be efficacious in reversing this arrest in differentiation. Fusion partners such as ABI1, AFDN, AFF1, MLLT1, and MLLT10 are associated with a higher risk of relapse, while other partners, including ELL, MLLT3, MLLT11, and SEPTIN6, are associated with standard risk [4]. Detecting KMT2A rearrangements is critical for disease classification, risk stratification, and therapeutic decision-making, as these rearrangements are targetable using menin inhibitors [5]. The menin inhibitor revumenib was recently approved for the treatment of relapsed/refractory acute leukemias with KMT2A rearrangements by the US Food and Drug Administration.

In many cases, detecting KMT2A rearrangements requires integrating multiple testing modalities. KMT2A rearrangement can escape detection when using conventional karyotyping techniques and the rearrangement may present as a simple deletion of chromosome 11/del(11)(q23), or as cryptic rearrangements with insertions or subtle breakpoints [1,6,7,8,9,10]. The KMT2A fluorescence in situ hybridization (FISH) break-apart probe is widely used in clinical laboratories to detect KMT2A rearrangements because this gene has numerous fusion partners, but this approach has limitations, including the inability to detect cryptic fusions, such as KMT2A::USP2 [7], or insertions [8,9], or partial tandem duplication [10], and FISH also does not provide information regarding specific fusion partners. Next-generation sequencing (NGS)-based RNA fusion panels provide higher resolution and can identify fusion partners and junctions with precision. However, these assays do not provide insights into underlying DNA alterations, especially complex rearrangements. Optical genome mapping (OGM) is a non-sequencing-based, whole-genome screening technology that detects complex structural variants with high resolution but cannot identify precise nucleotide-level junctions. Therefore, each testing modality has limitations, and discrepancies generated using these methods can present diagnostic challenges.

Here, we present a case of acute monoblastic leukemia in which different testing methods yielded discrepant KMT2A rearrangement results. Archer RNA fusion analysis identified KMT2A::AFDN fusion transcripts, and KMT2A FISH analysis showed a 3′ deletion suggestive of unbalanced KMT2A rearrangement. However, conventional chromosomal analysis and an initial analysis using our automated bioinformatic pipeline for OGM did not identify t(6;11)(q27;q23)/KMT2A::AFDN.

Due to our knowledge of the RNA sequencing result, we performed an additional manual inspection of OGM results at the KMT2A locus and identified a cryptic insertion of AFDN into the 3′ region of KMT2A, explaining the discrepant results. The RNA fusion was the result of an intragenic insertion event rather than an interchromosomal translocation.

2. Case Report

A 22-year-old woman presented with headache and a complete blood count (CBC) showed: WBC, 21.9 × 10⁹/L (reference range, 4–11 × 10⁹/L), hemoglobin, 8.3 g/dL (reference range, 14–18 g/dL), and platelets, 48 × 10⁹/L (reference range, 140–440 × 10⁹/L). A review of a peripheral blood smear showed 87% circulating blasts with monoblastic morphology characterized by open chromatin, variably conspicuous nucleoli, round to indented nuclear membranes, and moderate basophilic cytoplasm. No Auer rods were identified (Figure 1A). Computed tomography (CT) of the abdomen showed hepatomegaly (18.8 cm) and splenomegaly (16.4 cm) with no focal lesions. Bone marrow aspirate smears showed large blasts with morphologic features similar to the peripheral blood blasts (Figure 1B). The bone marrow biopsy specimen was hypercellular (100%) with sheets of immature cells (Figure 1C). Flow cytometry immunophenotyping performed on bone marrow aspirate detected blasts positive for CD34, CD117, CD33, CD38, TdT, HLA-DR, CD4, CD64, CD123 (increased), CD54 (partial), CD15, CD56, and MPO (dim, ~5% of cells) (Figure 1E–I) and were negative for CD13, CD19, surface and cytoplasmic CD3, and CD133. Immunohistochemistry performed on the bone marrow biopsy specimen showed that the blasts were positive for lysozyme (Figure 1D). These immunophenotypic findings support the diagnosis of acute monoblastic leukemia.

Conventional chromosomal analysis showed a complex karyotype: 47,XX,+8,del(9)(q21q31)[11]/47,idem,inv(11)(q14q23)[8] (Figure 2A). FISH for KMT2A showed deletion of 3′ KMT2A in 95% of interphase nuclei, suggesting an unbalanced rearrangement involving KMT2A (Figure 2B). The Archer RNA fusion panel identified an in-frame KMT2A::AFDN fusion transcript involving exon 8 of KMT2A (NM_005933.3) and exon 2 of AFDN (NM_001291964.1) (Figure 2C). The initial OGM analysis using our routine “rare variants” automated bioinformatics pipeline provided additional insights, detecting multiple numerical and structural abnormalities, including +8, del(9q21.31q31.2)(79,486,359_107,159,358) x1, interchromosomal fusion t(9;10)(q32;q25.1)(114,326,068;104,682,589), inv(11)(q23.3q23.3)(118,321,159_119,475,489) involving CBL, del(11q23.3)(118,485,078_118,873,559) x1 affecting the 3′ end of KMT2A, and two intrachromosomal fusions: fus(11;11)(q14.3;q23.3)(90,516,821;120,252,883) and fus(11;11)(q23.3;q24.3)(119,344,090;129,505,513). However, there was no evidence of the canonical t(6;11) translocation (Figure 3A,B). To address these discrepancies, metaphase KMT2A FISH (break-apart fusion probe) was performed to assess whether KMT2A could be translocated to another chromosome, but the results did not support this possibility. The fusion signal was observed on the normal chromosome 11, and the 5′ KMT2A signal was localized to the derivative chromosome 11 (Figure 2D).

These results prompted a manual review of the OGM data subsequently, which identified a cryptic ∼300 kb insertion of AFDN (exons 2–11) into the 3′ region of KMT2A (exon 8) (Figure 3C,D). This finding corroborated the KMT2A::AFDN fusion transcripts detected by the RNA fusion panel, reconciling the discrepancies between chromosomal analysis/FISH and RNA fusion results.

3. Discussion

In this case, the cytogenetic and molecular findings were initially intriguing due to the complexity of the genomic event at 11q23. Conventional chromosome analysis did not detect the canonical balanced translocation t(6;11)(q27;q23), which results in KMT2A::AFDN fusion. Instead, interphase FISH identified a 3′-KMT2A loss, suggesting the presence of an unbalanced translocation. The RNA fusion panel detected the KMT2A::AFDN fusion transcript, providing transcriptional evidence of the rearrangement. However, metaphase FISH did not resolve the discrepancy between the cytogenetic findings and the RNA fusion results. This discrepancy prompted a manual review of the OGM data that initially identified an unknown insertion at the 3′-KMT2A region. Upon detailed re-analysis, incorporating RNA fusion panel findings, we identified a cryptic insertion of AFDN into the 3′ region of KMT2A as a part of a complex genomic event on chromosome 11. This finding underscores the critical role of professional review in the OGM sign-out process and highlights the pitfalls of over-reliance on automated bioinformatics pipelines in analyzing complex genomic events.

To our knowledge, this is the second reported case of a cryptic insertion involving AFDN into the 3′-KMT2A, alongside one previously described case [12]. In the prior report, conventional karyotyping appeared normal, but further analysis using a whole chromosome 11 paint detected one KMT2A::AFDN fusion. An AFDN cosmid probe indicated that the other fusion resulted from the insertion of a submicroscopic portion of chromosome 6, including part of AFDN, into an apparently normal chromosome 11, a finding confirmed by RT-PCR [12]. Furthermore, three reports have described the insertion of 5′ KMT2A into AFDN, with two of them being cryptic, leading to KMT2A::AFDN fusions [6,9,11]. In a study of 30 cases with KMT2A::AFDN fusion, there was only one case with insertion, highlighting its rarity [6].

Moreover, in a study of 183 hematologic malignancies with t(4;11)/KMT2A::AF4, presented at the European 11q23 Workshop, only five cases involved variant or complex translocations, none of which were insertions [13]. This suggests that while insertions represent an uncommon mechanism of KMT2A rearrangement, they are a recognized phenomenon.

Among insertion-induced KMT2A fusion events, insertions of the 5′ KMT2A segment into partner genes are more frequently observed than insertions of partner genes into the 3′ KMT2A segment (80% vs. 20%) [14]. Our case, involving the insertion of the partner gene AFDN into 3′ KMT2A, represents a rare event. Cryptic insertions like the one in our case evade detection by traditional cytogenetic assays, including chromosomal analysis and FISH, and may be underestimated in routine diagnostics [8,15]. RNA-based assays—such as RNA fusion panels or transcriptome sequencing—are often preferred for uncovering cryptic rearrangements resulting in fusion transcripts [15].

In AML with KMT2A::AFDN, breakpoints typically occur in intron 8 or intron10 of KMT2A and intron 1 of the AFDN, findings that align with the case we report [7]. The AML in the patient we report exhibited monoblastic differentiation, consistent with prior reports that AML with KMT2A::AFDN typically exhibits myelomonocytic or monocytic morphologic and immunophenotypic features. KMT2A-rearranged acute monoblastic leukemia is usually negative for CD34; however, this neoplasm case falls within the uncommon 5% of cases that demonstrate CD34 positivity [16].

Although t(6;11)/KMT2A::AFDN was the sole abnormality in two-thirds of cases reported, trisomy 8 is one of the most common additional abnormalities, with a frequency of 10% [6], as observed in this case. AML with KMT2A::AFDN is associated with a poor prognosis and a high risk of relapse, making the identification of this cryptic fusion critical for classification, risk stratification, and patient management.

This patient underwent induction chemotherapy with fludarabine, cytarabine, granulocyte colony-stimulating factor with idarubicin, cytarabine, and venetoclax (FLAG -IDA+VEN) and achieved complete remission with minimal residual disease negativity by flow cytometry, followed by consolidation therapy with FLAG-IDA+VEN. She is currently ~2 weeks status post allogeneic stem cell transplantation in first complete remission.

In summary, this case highlights both the promise and challenges of integrating novel molecular and genomic tools—such as RNA fusion panels and OGM—alongside traditional techniques such as chromosome analysis and FISH. While these new technologies have the potential to provide detailed characterization of genetic alterations at both the RNA and DNA level, enabling detection of cryptic abnormalities, their routine clinical application is not without limitations. Each assay has its own strengths and constraints, and assay discordance remains a key challenge in multi-modal testing. Discrepancies between methodologies can arise due to differences in sensitivity, coverage, or bioinformatic interpretation, necessitating careful evaluation by trained professionals to resolve inconsistencies. Additionally, the increased complexity of multi-modal testing requires expertise in integrating diverse genomic data, ensuring accurate classification, risk stratification, and prognostication of AML, ultimately leading to improved patient outcomes.