NEW YORK – Investigators at Toronto's Princess Margaret Cancer Center have developed a new technique for combined microsatellite instability and mismatch repair deficiency testing, which they believe could replace the complex multi-step workflows that remain common in clinical practice.
In a study published last month in The Journal of Molecular Diagnostics, investigators also reported that the method, dubbed MultiMMR, was able to determine causes of MMR deficiency that may be missed by standard testing.
Testing for MMR deficiency and the microsatellite instability it causes is used to identify hereditary cancer syndromes, most notably Lynch syndrome, that arise when genes involved in this process are altered or inactivated. More recently, MSI has also been shown to associate with better responses to immunotherapy treatment, culminating in a companion diagnostic indication for pembrolizumab (Merck's Keytruda) in 2017.
Trevor Pugh, a Princess Margaret senior scientist and lead investigator developing the MultiMMR method, said that as the field has evolved testing practices for MSI/MMRd have become complex and variable from lab to lab. As a result, not all patients receive equivalent care.
The identification of Lynch syndrome and the discovery of its precise cause can necessitate multiple assays that use different technologies and expertise to perform, including immunohistochemistry, methylation analyses, gene-specific assays, and multi-gene sequencing panels.
"It can be very fragmented … [having] up-front IHC then a totally separate assay for MMR, then another assay for one gene … and a methylation test, and then a large battery of other genes. The motivation was really to have a one-stop shop to bring all the different aspects around MMR into one spot," Pugh said.
Making things more pressing, the study authors wrote, is a shift in professional recommendations endorsing things like universal testing for Lynch syndrome in certain cancer types or reflex germline testing for cancer patients of any type with somatic MSI findings.
As this evolution of clinical practice continues, current cascade testing protocols may not be able to keep up, the group wrote. "It's become a major movement in the field, and another reason to centralize a lot of this testing into a single platform," Pugh said.
Meanwhile, the testing landscape has become further complicated by the emergence of immuno-oncology drugs for which MSI/MMRd is a predictive efficacy biomarker. "Eligibility for immunotherapy is often contingent on MMR status, so timely and robust testing is important," Leslie Oldfield, the study's lead author said in a statement.
"MultiMMR streamlines the process, … can scale well with increasing demand, and can provide clinicians with important information to inform patient management and treatment decisions," she added.
Although analysis of MSI is increasingly incorporated into comprehensive tumor genomic profiling tests, individual panels differ from one another and largely lack the extensive validation that MMR IHC and PCR methods have undergone. As a result, NGS is still treated cautiously as a source for the diagnosis of mismatch repair defects by clinicians and even professional guidelines.
For example, recommendations drafted last year by The College of American Pathologists (CAP), the American Society of Clinical Oncology (ASCO), the Association for Molecular Pathology (AMP), and Fight Colorectal Cancer (Fight CRC) favor MSI by PCR over NGS when testing patients who are being considered for immunotherapy.
Pugh said that while oncology may be headed for a future of comprehensive genomics, even whole-genome sequencing, as a standard of care, it's not there yet. In the meantime, comprehensive but still niche-specific targeted assays can provide a bridge.
The team's MultiMMR assay simultaneously tests for promoter methylation, mutations, copy number status, copy neutral loss of heterozygosity, and microsatellite instability using a single aliquot of DNA. This is enabled by a strategy the team developed to generate a single library that is then split for methylation sequencing on one side and gene mutation and MSI analysis on the other.
In the study, Pugh and his colleagues used the method on DNA samples from 142 specimens (82 normal and 60 tumor samples) from 82 patients with MMR-associated colorectal, endometrial, and brain cancers. For a subset of 45 patients, the investigators then compared the MultiMMR results with previous clinical testing using conventional assays.
They also used MultiMMR on a commercially available DNA reference sample that includes 11 variants known to be challenging to detect with next-generation sequencing.
According to the authors, MultiMMR showed about 95 percent concordance with patients prior to clinical testing. In detecting variants responsible for the MMR deficiency, MultiMMR matched the previous clinical testing results in 23 out of 24 cases and identified a mismatch repair deficiency in 29 patients with incomplete or inconclusive testing.
The test also identified all 11 target mutations in the synthetic reference sample mix, even over multiple sequencing runs. In a separate study, laboratories had tested 10 workflows against this same synthetic control and only six found all 11 MMR variants, the authors wrote.
Importantly, Pugh said, MultiMMR was able to identify causes of MMR that can be missed by current stepwise practices. In one case, pathology review had identified MSH2 protein expression loss, but this was inconsistent with the somatic MLH1 promoter methylation detected by MultiMMR. Reexamination of the IHC results showed focal MLH1 protein loss in addition to MSH2.
"We spent a lot of time trying to figure out some of these diagnostic dilemmas where IHC is not consistent with molecular testing, and often it's actually explained by methylation of promoters that aren't typically looked at. When we do the methylation testing [with MultiMMR], we get information on all of the promoters … so we have knowledge on all the genes and proteins of interest. That helps tease apart some of these unexpected pairs of IHC chemistry findings," Pugh said.
Apart from logistical and comprehensiveness benefits, the group believes that MultiMMR offers cost advantages as well.
"Price is a moving target," Pugh said "So, even in the time of journal review for our paper, the price of Illumina sequencing dropped something like 20 percent. … It makes it very hard to do apples to apples price comparison, because this is now dramatically cheaper than single-arm gene panels were even year and a half ago."
A better focus, he argued is on workflow costs. "There is a lot of sort of systemic friction in getting a sample, getting a test, getting another test, getting a third test. … Sometimes those are in different labs."
With MultiMMR, "I think the real savings are perhaps in reagent costs, but even greater in logistical costs and the ability to issue a comprehensive report from a single molecular test," Pugh said.
According to the study authors, the team's latest version costs about $557 in library preparation and sequencing reagents per tumor/normal pair. Operationally, technicians spend about 63 hours of labor to process 18 cases, something that could be improved further with laboratory automation.
The Princess Margaret team is not yet using MultiMMR clinically, but Pugh said that the physicians the group has been working with are very interested in seeing the technology through to a clinical launch. "I don't think it's very, very far off, but we do need the final polish-up round of technology improvement," he said.
Currently, the team is working on the latest version of the panel, focusing on improving throughput and making it easier to run the methylation and the conventional genetic arms in parallel. "Right now, I think it's about a four-day protocol, but we think we can do it in a day and a half or two days," Pugh said.
Additional clinical validation will also be key, he said. And if that goes well, the team may potentially seek to make the protocol available as a kit for other labs to implement.
Other things the researchers are exploring include applying MultiMMR to liquid biopsy samples, something many of the team's clinical collaborators are especially interested in. This will require pushing the DNA input requirement down, but Pugh said that the team already has early data showing that its library construction protocols are compatible down to 10 nanograms of DNA input.
Finally, because RAS mutation testing for colorectal cancer patients is becoming recommended by more professional societies, Pugh and his coauthors also wrote that they are working on adding KRAS and NRAS to their capture panel, which would further eliminate the need for patients to receive multiple tests.