Skip to main content

NanoString-Based Pan-Sarcoma Fusion Gene Assay Shows Promise as First-Line Dx


SALT LAKE CITY (GenomeWeb) – A research team based in British Columbia has developed a sarcoma gene fusion test using NanoString's nCounter platform, which they believe can serve as a first-line clinical diagnostic, replacing multiple individual assays that make up the current standard of care for these cancers.

The group published a report on its validation of the test in the Journal of Molecular Diagnostics earlier this month, and lead author Tony Ng discussed the project further at a NanoString-sponsored workshop at the annual meeting of the Association for Molecular Pathology here this week.

Ng explained at the meeting that diagnosis of sarcomas is challenging because the larger designation includes multiple overlapping morphologies and phenotypes. And genetic diagnosis and subtyping has increasingly been shown to frequently amend or revise purely histological diagnoses, with important implications for treatment and other clinical decision-making.

Genetically, sarcomas are classified into two main groups: one containing cancers that are karyotypically complex and lack recurrent genetic alterations, and another characterized by tumors with specific lesions — chromosomal translocations and resulting fusion genes — that define specific subgroups.

But while tools have become available for genomic diagnosis of sarcomas based on the identification of these alterations, they leave significant room for improvement. One issue is that some sarcoma subtypes are characterized by multiple fusion combinations involving several partner genes. In addition, the specific breakpoint in each partner gene can vary, resulting in a variety of exon-exon fusion combinations at the transcript level.

According to Ng, because current clinically available fluorescence in situ hybridization or reverse transcriptase-PCR approaches can't be sufficiently multiplexed, costs and turnaround time can pile up with the use of multiple consecutive tests.

Recognizing a clinical need for a single pan-sarcoma gene fusion identification assay, he and his colleagues decided to investigate NanoString's nCounter platform, which uses nucleic acid hybridization probes with specific and distinguishable fluorescent bar code tags that allow detection of multiple specific short nucleic acid sequences within a given sample.

Ng said at the meeting that he and his colleagues aren't the only sarcoma specialists who have tapped NanoString. Another Canadian team from Hospital for Sick Children in Toronto also began working with nCounter in 2015. That group's assay is narrower in scope though — focused specifically on pediatric sarcomas.

In their JMD study, Ng and his BC colleagues analyzed samples from a cohort of 212 cases across many different sarcoma types, comparing the sensitivity of their nCounter-based approach with traditional FISH/PCR testing and assessing the capability of the new assay to improve diagnosis by detecting fusion genes in cases where conventional techniques fell short.

Additionally, the investigators used the study to examine per-sample cost and turnaround time for the NanoString test relative to the current standard.

According to Ng, development of the assay required somewhat of a "brute force" approach, in which the team conducted an exhaustive literature review to curate targets for which they could obtain a custom set of junction probes from NanoString.

Using the resulting assay, 96 cases out of the 212-patient cohort came out positive for fusion gene expression. This included all 20 Ewing sarcomas in the cohort, 11 synovial sarcomas, and five myxoid liposarcomas.

In 15 of the cases, the NanoString platform detected fusions that were missed by a comparator standard clinical assay. These included EWSR1-FLI1, EWSR1-ERG, BCOR-CCNB3, ZC3H7B-BCOR, HEY1-NCOA2, CIC-DUX4, COL1A1-PDGFB, MYH9-USP6, YAP1-TFE3, and IRF2BP2-CDX1 fusions.

There were no false positive results using the NanoString platform. However, the comprehensive test did miss four fusion-positive cases as detected by clinically-available FISH or PCR assays.

Ng said at the AMP meeting that these false-negatives were largely cases of unpublished fusions, or targets that the group had not incorporated into their nCounter codeset, suggesting that these discrepancies can be resolved by continuing to evolve the panel.

At the AMP meeting, Ng highlighted this ability to add iterative updates as a particular strength of the nCounter system. This is something the BC team is already working on, he added, both by adding newly described fusion variants, and by experimenting with a 5'-3' strategy for areas not well-covered by junction probes.

However, realistically, the panel can't constantly cover every new or emerging clinically relevant fusion target. For that reason, Ng said that the clinical workflow he and his colleagues have proposed recognizes this, by making sure that histology and IHC are still performed up front, helping to establish baseline histologic diagnoses that help funnel in cases that are most likely to be covered by the assay.

This should help keep the false-negative rate low.  In cases that don't show a result using nCounter, another approach like sequencing could then be used to try to identify a missed or novel alteration.

Importantly, Ng highlighted, the study showed that the approach is highly sensitive for the sarcoma subtypes for which molecular diagnosis is the most clinically critical, namely Ewing's sarcoma, which requires particularly aggressive treatment.

In regard to cost effectiveness, Ng and his colleagues reported that when batched as six cases, the per-sample reagent cost using nCounter was completely comparable to conventional techniques, with a technologist hands-on time of 1.2 hours per case and assay time of 36 hours.

This evidence of cost effectiveness and timeliness suggests that an nCounter approach could "replace the majority of current clinically-available FISH and reverse-transcriptase PCR assays as the initial molecular diagnostic test for sarcomas," the group wrote in their JMD paper.

"We feel that this is something that we can certainly use in the clinical setting," Ng added at the meeting. "Data analysis is a piece of cake … and the cost effectiveness is [clear]."