Skip to main content
Premium Trial:

Request an Annual Quote

At AACR, Experts Discuss Benefits and Challenges of Incorporating NGS in Clinical Trials


SAN DIEGO — Incorporating next-generation sequencing in the context of oncology clinical trials has many potential benefits as well as hurdles, according to a number of researchers that presented this week at the American Association for Cancer Research meeting in San Diego.

Researchers are starting to use sequencing both retrospectively to analyze exceptional responders to drugs — a strategy that can be employed to rescue failed trials — and also prospectively to match patients to a specific arm of a trial based on their mutational profile.

A trial by the National Cancer Institute, called MATCH, will use a next-gen sequencing panel developed on Life Technologies' Ion Torrent PGM system to randomize patients into one of many different phase II trials. The trial is expected to begin in November and will be based on another NCI trial called M-PACT, which was recently launched and will study the impact of next-gen sequencing on guiding treatment decisions for cancer patients.

Separately, a group from Memorial Sloan Kettering Cancer Center has developed a next-generation sequencing panel called Impact, and it plans to genotype all incoming patients with the aim of studying whether the patient's treatment was impacted by the NGS assay.

Researchers at MSKCC have also used NGS to analyze so-called "exceptional responders" to drugs that either failed or showed limited therapeutic response in clinical trials, an effort which has since prompted the NCI to begin its own Exceptional Responders Initiative.


The NCI's MATCH trial is slated to begin in November and will use a targeted sequencing assay developed on Life Technologies' Ion Torrent PGM to direct patients into different phase II clinical trials for targeted therapies.

According to Mickey Williams, director of the molecular characterization laboratory at the Frederick National Laboratory of Cancer Research and who is contracted by the NCI, the MATCH study is based on the M-PACT study. It will have multi-arm phase II trials. Patients will be placed into a specific trial arm based on the results of the NGS assay.

The trial arms will include both single agents and combination therapies, FDA-approved and investigational drugs, and will have two primary endpoints — response rate and progression-free survival.

The PGM assay will include both targeted DNA sequencing and targeted RNA sequencing to identify gene fusions. Four CLIA-certified laboratories will run the NGS assay — the Frederick National Laboratory of Cancer Research, MD Anderson, Massachusetts General Hospital, and Yale.

Patient consent will also include the use of genomic information for research, so if enough sample is available, more comprehensive sequencing can be done, such as exome, whole-genome, or whole-transcriptome sequencing.

Williams said that the team at NCI has been working closely with the US Food and Drug Administration on the design of the trial as well as the validation of the NGS assay to ensure that it meets their requirements. In order to use next-gen sequencing in such a manner, sensitivity, specificity, and accuracy, including for both variants of high and low frequency, will have to be demonstrated. The labs running the assay will have to demonstrate that it is reproducible both within the lab and between labs.

He said the NCI chose the PGM over the Illumina MiSeq system for the study because the researchers will use material extracted via fine needle biopsies, and based on the commercial protocols of both Illumina and Life Technologies, the "PGM requires 25-fold less input DNA," he said.

Identifying 'exceptional responders'

According to David Solit, a director of the center for molecular oncology at MSKCC, there is a huge need to accelerate drug development and to figure out which drugs will benefit which patients.

Often, in developing targeted therapies, targets are initially identified by genomically characterizing tumors and cell lines, but perhaps another way is to identify the genetic basis for rare extraordinary clinical response to guide trials in select sub-populations of individuals, Solit suggested.

He cited several examples of patients that did exceptionally well in a drug trial that otherwise showed no or minimal clinical response. In one case, a 73-year-old patient with bladder cancer achieved a complete response to the mTOR inhibitor everolimus. The patient was one of only two of the 45 patients to show any response and the drug did not meet its statistical endpoint.

Retrospective whole-genome sequencing of the patient identified a frameshift mutation in TSC1 that led to inactivation of TORC1, which was the target of the drug, Solit said.

In a separate example, Solit described a patient with ovarian cancer that had been enrolled into a trial for a MEK inhibitor, which also ended up being a "disappointing study" with only a 12 percent success rate. The patient was the only one to achieve complete response. Five years later, she is still on the drug and cancer free, he said.

Researchers at MSKCC retrospectively performed an in-house-developed targeted sequencing panel called Impact that assesses 341 cancer-related genes and found a 15-bp in-frame deletion on a gene within the MEK signalling pathway. She was "the perfect patient to give a MEK inhibitor to," Solit said.

The success of Memorial Sloan Kettering's initial sequencing of the metastatic bladder cancer patient has since spurred the NCI to begin an Exceptional Responders Initiative. According to Barbara Conley, associate director of NCI’s Cancer Diagnosis Program, between 1 percent and 10 percent of patients respond well to drugs that do not end up receiving FDA approval and "mutations or changes in gene expression may explain" that response. As such, the NCI is planning a retrospective study to perform next-generation sequencing on such patients, including exome, whole-genome, targeted, and transcriptome sequencing.

While the "use of whole-genome sequencing of outliers to salvage drugs with a low response rate" is one strategy, Solit said, the field also needs an "assay to prospectively look for patients."

The primary hurdle to developing targeted therapies is "identifying the patients," he said, noting that the center plans to begin sequencing all its patients who have recurrent or metastatic disease.

Solit said MSKCC has an IRB-approved protocol in place by which the patients can be genotyped via the center's targeted NGS Impact test, and consented to have whole-genome or exome sequencing performed for research purposes. While the primary goal of the study will be to see whether the NGS panel impacts treatment, a second goal will be to identify and return pathogenic germline variants.

Lack of standards

One major challenge with using next-gen sequencing in the clinic is the current lack of standards around the technology. According to Williams, one major question that has not been addressed is whether different platforms will give the same result.

For instance, he said, "if someone is screened using Foundation Medicine versus the PGM panel that we're using for MATCH, will we get the same result? Will the same clinical decision be made?" he asked. "It's [a question] we need to address."

This is one reason why the NCI has been working so closely with the FDA in its design of the MATCH trial and why the group plans to file a pre-submission to the FDA for the assay. "We want to engage the FDA early and discuss risks," he said.

Aside from the sequencing platform itself, every step within the process has the potential for introducing variability, including sample prep, the bioinformatics, and the clinical decisions that are made based on the results, Williams said. As such, while four labs will run the NGS assay, only one will do the nucleic acid purification, and there will be a central data analysis pipeline that will call actionable variants. Those variants will have to be signed off on by the lab director, but the idea is to automate the process as much as possible. Finally, rather than a tumor board that makes clinical decisions, the group is designing a rules-based engine to automatically place the patient in a specific arm of the trial, Williams said.