Improving the characterisation of acute paediatric leukaemia worldwide

There are many different types of blood cancers that impact various stages of the blood production cycle, with the most common type in children being leukaemia^1,2. B cell acute lymphoblastic leukaemia (B-ALL) is one of the most common types of leukaemia, which causes uncontrolled growth of non-functional B cell lymphocytes^2,3. Acute myeloid leukaemia (AML) is another type of blood cancer, leading to the fast growth of monocytes and granulocytes⁴. Both lead to the build-up of non-functional cells in the blood and bone marrow, preventing normal blood cells from developing^1,4.

Not only is B-ALL the most common type of leukaemia, but it is also the most common type of paediatric cancer, with AML making up 15–20% of paediatric cancers³. Due to both being acute leukaemias, disease progression is fast⁴ and requires timely classification into genomic subtypes to determine prognosis and treatment plans³.

The arsenal of genetic tests used in the standard-of-care pipeline

Currently, B-ALL and AML are classified into genomic subtypes through a range of complicated and expensive genetic tests known as the ‘standard-of-care’ diagnostic pipeline. Tests include karyotyping, fluorescence in situ hybridisation (FISH), PCR, targeted panels, and whole-genome sequencing³. However, the specific genetic tests used in the diagnostic odyssey vary between different clinical facilities, especially in low- and middle-income countries (LMICs) where resources are limited^3,5. Despite these limitations, most cancer treatment centres across the globe are in LMICs³ with significantly lower incidence rates⁵, highlighting the need for accessible sequencing.

To improve access to testing, Geyer et al. investigated how to create a single, sequencing-based classification workflow to simplify the standard-of-care pipeline, with the aim of reducing testing costs and time. Typically, short-read-only sequencing methods are used to perform genomic subtyping of AML and B-ALL to resolve structural variants (SVs). However, Geyer et al. investigated Oxford Nanopore sequencing — a platform capable of generating reads of any length, from short to ultra long — because it has the ‘potential to generate faster and more cost-effective results compared with short-read sequencing methods’³ on a single consolidated platform.

Figure 1. The complete workflow from DNA extraction to sample classification. Figure from Geyer, J. et al.³ and available under Creative Commons license (http://creativecommons.org/licenses/by/4.0/).

Moving the standard-of-care pipeline to a single platform

Oxford Nanopore sequencing can generate reads of unrestricted length, providing more comprehensive data than short reads, because the long nanopore reads can span complex and repetitive genomic regions. Geyer et al. noted that nanopore data is streamed in real time, enabling rapid access to results, and that adaptive sampling — a unique targeted sequencing method without additional sample preparation — provides ‘comprehensive genomic sequencing, offering a more efficient use of resources and time’.

Geyer and colleagues performed Oxford Nanopore whole-genome sequencing on 57 acute leukaemia research samples. The samples were sequenced in singleplex and multiplex for up to 72 hours on MinION and PromethION 2 Solo devices; however, they found results could be inferred from as little as 15 minutes and up to six hours of sequencing (Figure 1). During sequencing, adaptive sampling was employed to enrich 59 genes commonly involved in translocations or fusions in B-ALL and AML. For six samples, 152 genes were targeted with adaptive sampling to enrich for additional genes and regions associated with AML, B-ALL, and T cell ALL, another type of acute lymphoblastic leukaemia.

Generating more comprehensive data than traditional genetic tests

From sample receipt to data analysis, the Oxford Nanopore sequencing-based pipeline took between six to nine hours and provided data that was ‘100% consistent with clinically derived genomic subtype classification’. Using a novel bioinformatics pipeline, Geyer et al. analysed nanopore data to identify genetic variants from chromosome-level abnormalities and large-scale SVs down to copy number and single nucleotide variants in a single rapid assay.

Geyer et al. achieved good depth of sequencing coverage across the genome (average of 12x), and an average of 86x coverage over targeted genes. This provided enough comprehensive data to identify genomic variants that typically require multiple traditional molecular tests to characterise. Using both whole-genome and targeted sequencing data, the team accurately identified clinically relevant karyotype profiles and gene fusions in acute leukaemias. Additionally, targeted data led to the identification of small-scale structural variants, copy number variants, and single nucleotide variants. Taken together, these results demonstrate that the Oxford Nanopore assay reliably detects clinically relevant genomic variation and tumorigenic drivers, which have the potential to improve future diagnostic outcomes.

Furthermore, long Oxford Nanopore reads provided additional information inaccessible to legacy sequencing methods, such as data to identify complex genomic variants. For example, DUX4 rearrangements are the cause of 14% of B-ALL cases and are not well-characterised using traditional methods⁶. Geyer et al. detected an SV and corresponding DUX4:IGH gene fusion using nanopore sequencing from a single run, demonstrating the potential of Oxford Nanopore sequencing for advanced diagnostic benefits in the future.

Democratising access to sequencing

Geyer et al. highlighted that the nanopore sequencing-based workflow is more cost-effective than the traditional standard-of-care pipeline comprising multiple molecular tests, due to simple nanopore workflows that require few reagents. The authors concluded that further ‘optimization and automation of the sequencing pipeline are essential to democratizing this approach’. However, this study demonstrates that Oxford Nanopore Technologies has a role in potential improvements in the current variable complex standard-of-care assays for classifying paediatric acute leukaemias by providing a cost-effective and rapid single-platform assay.