A Prevalence Study of Gene Fusions in 2958 Acute Myeloid Leukemia Patients from the Community Using FISH and NGS Testing

Background: Gene fusions have profound implications and significantly influence clinical decisions in cancer diagnosis, prognosis, therapeutic selection, and patient quality of care in Acute Myeloid Leukemia patients among other hematological malignancies. Gene rearrangements have been classically tested in the clinical setting by FISH, RT-PCR, and, more recently, also by NGS. This study interrogates the prevalence of real-world clinical fusion alteration detection from three perspectives: tumor-agnostic, Acute Myeloid leukemia-specific, and lastly by comparing between NGS and FISH laboratory techniques.

Methods: Clinical samples (n=29,467) from hematologic cancer patients were tested for RNA fusions using an enriched, amplicon-based, dual DNA/RNA sequencing platform with a collection of cDNA-specific primers that target RNA breakpoints at 229 genes and can detect known and novel fusions. The comprehensive NGS panel-tested samples were analyzed to understand the distribution of fusion detection agnostic of cancer-type associations. Samples were then filtered based on indicated Acute Myelogenous characteristics to determine the fusion prevalence for this malignancy lineage. This subset of AML samples was further analyzed against FISH methodologies to compare performance and identify any detection gaps.

Results: The total fusion positivity prevalence in all cases tested from all heme malignancies was 19.3% (n=5697), with the four most frequent fusions including BCR::ABL1-3.7%, TFG::GPR128- 2.4%, RUNX1::RUNX1T1- 1.0%, PML::RARA- 0.9%. We then focus on the prevalence of fusions among AML patients (n=2958) where we found 507 positive cases- 17.1% (507/2958). The 10 most prevalent fusions in AML patients were: RUNX1::RUNX1T1- 5.1% (n=152), CBFB::MYH11 (n=80), TFG::GPR128 (n=75), NUP98::NSD1 (n=69), BCR::ABL1 (n=46), ZEB2::PTMA (n=46), PML::RARA (n=44), ELL::KMT2A (n=23), ZEB2::BCL2L11 (n=23), KMT2A::MLLT3 (n=22), all others- 16.8% (n=497). When we analyzed unique fusion genes, we found there are some recurrent genes fused with multiple other partners, such as RUNX1, which was present in 10% (n=296) patients, while its best known and most frequent partner, RUNX1T1 (n=152), was only fused to RUNX1. Additionally, KMT2A (n=125) is most frequently paired with ELL (n=42) but also with other genes, such as MLLT3 (n=23). When we explored 97 NGS fusion-positive cases with concurrent FISH analysis, we found 48 cases had a positive FISH rearrangement result, where 44 of these cases had both a concordant NGS fusion detection with a FISH detection. For two additional patients, the NGS test provided an alternative alteration detection to the PML-RARA positive FISH results.

In one of these patients, there were ploidy alterations detected on chromosomes 11, 17, and 21 from probes intended for TP53, KMT2A, and RUNX1, and the NGS test only detected a BCL6::SLC9A9 gene fusion. In the other patient, FISH testing detected an abnormal yet unspecified RARA probe signal in chromosome 17, but NGS analysis revealed that the translocation was a KMT2A::MLLT6 fusion. Interestingly, the location of MLLT6 is mapped to chromosome 17 (q12), roughly 2Mb apart from the RARA gene, and no RARA fusion was detected. Moreover, there were 49 patients harboring RUNX1 (38), BCR::ABL1 (6), KMT2A (5), and PML (1) fusions detected by NGS that were undiscernible by FISH. Thus, the real-world data show that both tests offered equivalent results on most (44/48) patients between detected FISH rearrangements and NGS fusion detection.

Conclusion: This study showed that gene fusions are very common in AML patients, and NGS testing unveils a high gene fusion prevalence (17.1%) in these patients. A prominent finding was that RUNX1 and KMT2A fused genes occur most frequently in AML patients and that the overall fusion prevalence is higher than what can be inferred from FISH testing. It is important to note that these results cannot be confused with an accuracy study but rather taken as proof of the value of NGS testing concurrently with FISH methodologies to gain a comprehensive rearrangement profile on AML patients to help with their diagnosis, prognosis, and therapeutic intervention.

Duong:Neogenomics: Current Employment. Krawczyk:Neogenomics: Current Employment. Montgomery:Neogenomics: Current Employment. Lyle:Neogenomics: Current Employment, Current equity holder in publicly-traded company. Lopez-Diaz:Neogenomics: Current Employment.