Rapid identification of MSK-IMPACT cancer variants with real-time targeted nanopore sequencing

At the Nanopore Community Meeting 2024 in Boston, USA, Dr Stephanie Chrysanthou (Memorial Sloan Kettering Cancer Center, USA) showcased how nanopore sequencing using adaptive sampling of the targets in the MSK-IMPACT panel enables rapid calling of cancer-associated pathogenic variants.

MSK-IMPACT: the first test of its kind

Molecular profiling of tumours is critical for precision oncology. The identification of germline and somatic variants is important not only for cancer diagnosis, but also to inform prognosis, preventive surgery and therapies, and cancer surveillance.

MSK-IMPACT (Memorial Sloan Kettering integrated mutation profiling of actionable targets) is a targeted test that identifies both germline and somatic mutations across solid tumour cancers¹. In 2017 the lab-developed test, based on hybridisation capture and next-generation tumour-normal sequencing, became the first of its kind to be authorised by the US Food and Drug Administration (FDA).

Of the thousands of individuals participating in MSK-IMPACT, 17% have been found to harbour germline pathogenic mutations. The ability to detect these variants is not only critical to reducing risk to the individual but can also have ‘major health implications’ for their family.

In the current MSK-IMPACT workflow, DNA is first extracted from solid tumour and paired normal blood samples. The samples are then prepared for target enrichment via hybridisation capture of the 505 target genes in the MSK-IMPACT panel, representing actionable cancer predisposition genes, then sequenced using a short-read technology. A data analysis pipeline is used to detect germline and somatic mutations in the target regions, covering SNVs, structural variants (SVs), copy number variants (CNVs), and microsatellite instability information.

Where testing was previously limited to a smaller number of solid tumour types, MSK-IMPACT provides critical variant information across rare and common solid cancer tumours.

However, hybridisation capture is a labour-intensive and time-consuming enrichment assay that includes overnight probe hybridisation and, given the limitations of short-read sequencing, samples must then be batched, resulting in lengthy turnaround times. Furthermore, any updates to custom hybridisation probe panels require considerable labour, time, and costs to implement and revalidate.

Stephanie and her colleagues at the Integrated Genomics Operation (IGO) core laboratory, who use MSK-IMPACT for cancer research, investigated whether Oxford Nanopore adaptive sampling-based targeted sequencing could resolve these challenges. To test this, they benchmarked nanopore sequencing against prior MSK-IMPACT profiles generated using hybridisation capture and short-read sequencing.

Harnessing real-time nanopore sequencing

Adaptive sampling is a bioinformatics-based targeted sequencing method unique to Oxford Nanopore. Unlike hybridisation capture or amplicon-based techniques, Stephanie explained, the target enrichment step takes place ‘during sequencing, on the instrument, instead of on the bench’ — without lengthy library prep protocols. The PCR-free workflow is simple, rapid, and preserves long fragments, producing long nanopore reads that can resolve large, complex SVs, tandem duplications, and repetitive retrotransposon elements. With no need for bespoke probes, any changes — for instance, the addition of new targets — to a panel such as MSK-IMPACT using adaptive sampling would be as simple as editing the target BED file, saving considerably on time, cost, and complexity.

Stephanie and her team created a BED file targeting the complete MSK-IMPACT panel, plus three housekeeping genes for use as a control to call CNVs. In total, this comprised 2.4% of the human genome.

From extraction, then library preparation with the Ligation Sequencing Kit, to targeted sequencing on one PromethION Flow Cell, and finishing with data analysis, the ‘much shorter’ nanopore sequencing workflow takes just 3.5 days — a turnaround time Stephanie believes can be reduced to 2.5 days or less with additional optimisation. Crucially, there is no need to wait to batch samples for sequencing on Oxford Nanopore devices — they can be sequenced as soon as required, removing the waiting times associated with legacy short-read sequencing technology. For data analysis, the team found that the fastest method was to ‘make use of [Oxford] Nanopore’s own offerings’: on-instrument basecalling and alignment, followed by variant calling with the intuitive EPI2ME workflow, wf-human-variation.

Rapid identification of pathogenic germline variants

The streamlined MSK-IMPACT targeted nanopore sequencing workflow enabled Stephanie and her colleagues to rapidly identify germline variants from tumour-normal research samples. Stephanie shared examples of variants they successfully identified, including a 313-base deletion in BARD1 and an SNV in CHEK2.

Stephanie described one research sample that was known to contain three germline mutations. Using the EPI2ME workflow, they were able to call the missense variant in BRCA1 and pathogenic SNV in LZTR1. The third, an SV in RAD51C, was clearly visible in the aligned long nanopore reads, but was not initially detected. Noting that ‘one size doesn’t fit all’ for SV calling, the team tested several SV callers before deciding to incorporate the tool Severus² into their pipeline. This enabled calling of the third mutation: a deletion spanning over 36 kilobases. Utilising their optimised data analysis pipeline, they ‘easily’ identified a pathogenic tandem duplication in PALB2 in another sample.

Finally, Stephanie discussed a research sample with a pathogenic four-exon deletion in MSH2. Severus did not identify this SV because its breakpoints were in intronic regions — and their filter for MSK-IMPACT targets excluded nearly all introns. The team plans to revisit this filter as, unlike short-read sequencing of hybridisation capture panels, which focuses on exons, nanopore sequencing with adaptive sampling produces long reads that allows for spanning of both exons and introns — and thus the identification of ‘exactly where the breakpoints are, because it doesn’t just target exons’.

With their nanopore sequencing workflow and optimised data analysis pipeline, Stephanie and her colleagues achieved just under 97% concordance with the short-read MSK-IMPACT data across 25 paired tumour-normal samples. The team also plans to ‘take advantage of the methylation data that comes for free with nanopore sequencing’, providing deeper molecular characterisation of their samples.

Stephanie concluded that ‘you can indeed use long-read adaptive sampling to detect pathogenic germline variants’ with MSK-IMPACT. She emphasised the accessibility of the technology, noting its potential for ‘other smaller entities with less infrastructure than MSK’. Beyond MSK-IMPACT, Stephanie noted, the method could be adapted for use with ‘any gene panel’ for a simple, highly customisable targeted workflow.

Find out more about cancer research with nanopore sequencing