Targeted gene sequencing — coupled with complementary whole-genome sequencing, when needed — can unearth tumor-specific markers for monitoring metastatic breast cancer using the cell-free tumor DNA that circulates in the bloodstream, according to a proof-of-principle study published last week in the New England Journal of Medicine.
"This is going to be a very exciting field," the study's co-senior author Nitzan Rosenfeld, a researcher with the University of Cambridge and the Cancer Research UK Cambridge Institute, told Clinical Sequencing News.
"There's a lot to study," he explained, "and I think it's an area where the clinical questions and the maturations of technologies are coming together to create a very interesting field of activity and research."
In their current analysis of 52 women with metastatic breast cancer, Rosenfeld and colleagues from the UK and Australia described the targeted and whole-genome sequencing methods they used to find tumor-specific somatic mutations and/or structural changes in 30 of the cases.
These genetic glitches subsequently served as markers for detecting metastatic breast cancer sequences in individuals' circulating, cell-free DNA over the course of treatment, they explained. The presence of tumor sequences in the cell-free fraction of the patients' blood also made it possible to catch 89 percent of the progressive disease cases — in several instances providing the first indication of disease progression.
Indeed, when the researchers put the cell-free DNA method head-to-head against computed tomography scanning, it caught more than half of progressive disease cases earlier than the standard imaging-based method. It also compared favorably to other blood-based approaches, such as tests for a circulating cancer antigen called CA15-3 or for circulating tumor cells.
"We're obviously encouraged by this data and we're expanding this to other cancer, other genes, other treatments, earlier cancers," Rosenfeld said. "We think this might have potential in many different areas and fields in cancer and we're trying to accumulate the relevant data."
The recognition that tumor cells and cell-free tumor DNA can make their way into an individual's blood stream has spurred interest in finding ways to exploit these blood-based materials in the clinic, both for prenatal testing (see, for example, CSN 12/19/2012, CSN 11/14/2011, and IS 3/31/2009) and for finding and following cancers.
Almost a decade ago, for instance, researchers based at the University of Texas MD Anderson Cancer Center published a NEJM study looking at relationships between disease progression or survival and levels of circulating tumor cells in the blood in individuals with metastatic breast cancer.
And several groups have come up with ways of not just finding, but also characterizing cancer cells that circulate in the bloodstream. Earlier this month, for instance, German researchers reporting in Cancer Research outlined an array-comparative genomic hybridization and high-throughput sequencing scheme for assessing the genomes of individual tumor cells nabbed from circulation.
At Cold Spring Harbor Laboratory, meanwhile, a team is looking at the possibility of employing single-cell sequencing on circulating tumor cells to better diagnose and treat prostate cancer (CSN 9/7/2011). And others have shown that it's possible to get transcriptome information on one or a few circulating tumor cells using a messenger RNA sequencing method known as Smart-Seq (CSN 7/25/2012).
Researchers have been finding ways to discern tumor features from cell-free DNA in the blood, too (CSN 12/5/2012). But so far less work has been done to determine the effectiveness of following tumor dynamics with this circulating tumor DNA — something authors of the new study decided to explore in patients with metastatic breast cancer.
From their initial set of samples from all 52 women with metastatic breast cancer, the researchers began by doing targeted PIK3CA and TP53 gene sequencing, using the Fluidigm Access Array and Illumina's GAIIx or HiSeq 2000 instruments to do tagged-amplicon deep sequencing on matched tumor-normal samples from each individual.
That analysis was sufficient for finding trackable somatic mutations in 25 tumors — or nearly half of the cases.
But the team didn't stop there. They also did whole-genome paired-end sequencing on tumors and matched normal samples from nine of the women, including several individuals for whom targeted PIK3CA and TP53 sequencing did not uncover somatic tumor mutations.
For those experiments, the investigators did not have a firm depth of coverage in mind, Rosenfeld explained. "The study aim wasn't to comprehensively cover the genome, so we were flexible about that."
Instead, their objective was to pick up at least one distinguishing mutation per tumor so that they could test for the presence of DNA from the cancer in their subsequent analyses of cell-free DNA in the blood.
With the genome sequence data, the group found distinctive structural variants in eight cases. These included five cases that did not yield mutation markers during the targeted sequencing step, bringing the tally of mutation-trackable tumors up to 30.
Using the digital PCR assay and/or the tagged amplicon-sequencing approach, the researchers went on to test for these genetic glitches in cell-free DNA from 141 blood samples that had been collected from the 30 women over the course of their treatment.
Overall, the method picked up the presence of metastatic breast cancer in all but one of the 30 cases and in 82 percent of the blood samples tested for circulating tumor DNA.
When researchers tested blood samples for circulating tumor cells using Veridex's CellSearch System, meanwhile, they detected disease in 87 percent of the cases, though tumor cells did not turn up in 40 percent of the 126 blood samples taken over time in the same 30 women whose blood was tested for tumor DNA.
An immunoassay test for the tumor antigen CA 15-3 in the blood found 78 percent of cases.
Levels of both circulating tumors cells and CA 15-3 antigen in the blood seemed to coincide with increasing or decreasing tumor prevalence by imaging, as did the circulating tumor DNA, study authors noted. But the cell-free DNA approach "showed a greater dynamic range, and greater correlation with changes in tumor burden, than did CA 15-3 or circulating tumor cells."
The circulating tumor DNA-based approach also revealed progressive disease between two and nine months earlier than the standard imaging approach in 10 of the 19 progressive cases assessed using both methods.
The cell-free DNA method was not applied to the 22 cases in which tumor-specific mutations were not identified initially, though Rosenfeld noted that more extensive analyses of these tumors — be it targeted testing of a wider gene set or whole-genome sequencing — would be expected to yield suitable markers for those tumors, too.
"I'm definitely expecting that if we'd done whole-genome sequencing on these we would find something to track," he said. "I'd be very surprised if there would be a case where we would sequence the tumor's whole genome and not find any mutations to track."
So far it appears that either structural alterations or somatic sequence changes in the tumor can be tracked in circulating DNA as a means of establishing the presence of tumors within the body.
But Rosenfeld noted that it may be worth looking more closely at whether there's a benefit to focusing on apparent driver mutations, such as those detected in candidate cancer genes, or to unearthing tumor markers that are easy to spot, such as certain amplifications.
"As we show in the paper, both can provide information," he said. "It's a bit to early in the field to set more explicit guidelines, but I think both are useful and worth attention."
There's also the question of price, since whole-genome sequencing remains roughly an order of magnitude more expensive than the targeted PIK3CA and TP53 gene testing, despite drops in sequencing costs.
So while whole-genome sequencing "is probably going to work in just about every case you're going to run into," Rosenfeld explained, "it's obviously more expensive in terms of finding the mutations."
It also remains to be seen whether circulating tumor DNA monitoring would be effective in patients with primary rather than metastatic cancer, since those individuals are expected to have less tumor DNA in the cell-free fraction of their blood.
"The more advanced cancer patients, unfortunately, as part of the [disease] definition, have a higher tumor burden and it's easier to find their tumor DNA in the plasma," Rosenfeld said. "There is, at the moment, very little data about earlier [stage] cancers, but the expectation based on some data is that this is going to be much lower."
Similarly, he explained that there are still questions about how levels of circulating, cell-free tumor DNA relate to factors such as tumor biomass, tumor growth, or tumor cell death during treatment, if at all.
"There are many open questions in this field and we're just starting," Rosenfeld said.
In an effort to continue generating the type of data needed to determine the clinical utility of cell-free tumor DNA, he and his colleagues plan to use similar strategies to test other types of cancer and stages of disease. They are also considering the option of using expanded sets of genes for targeted sequencing during the initial search for somatic mutations in the tumor.
For example, while the team saw somatic mutations in about half of the cases by looking at just PIK3CA and TP53, Rosenfeld noted that newer studies have uncovered other frequently mutated genes in breast cancer, suggesting there should be other candidate genes to target during the search for marker mutations in future circulating tumor DNA studies.
"Having some intermediate panel is probably going to be very useful," Rosenfeld said. "One doesn't have to jump immediately to whole-genome sequencing. One can do panels of genes."