NEW YORK (GenomeWeb) – Researchers from the University of California, San Diego have confirmed that testing circulating tumor DNA from cancer patients in a clinical setting can uncover actionable mutations and guide therapy, but also that tissue sequencing and liquid biopsy have unique and likely complementary value.
Though these types of studies have become ubiquitous in the realm of tissue-based next-generation sequencing, medical centers are still in the early days of tracking the results of clinically implemented liquid biopsy tests.
And while companies like Guardant Health, whose Guardant360 liquid biopsy test was used in the UCSD-led study, are reporting on results from clinical testing into the tens of thousands of samples, there have been few independent reports on the yield and clinical impact of these types of tests, at least compared with the breadth of evidence available for tissue-based NGS.
In their study, published in Clinical Cancer Research earlier this month, the UCSD team led by Maria Schwaederle took a fairly straightforward approach, collecting records on test results, clinical histories, and outcomes of 168 patients who had received Guardant's 54-gene panel ctDNA analysis as part of their clinical evaluation at UCSD's Moores Cancer Center between June 2014 and February 2015.
Schwaederle told GenomeWeb that liquid biopsy is being incorporated into clinical care at Moores at the discretion of each individual physician.
Her team's study focused only on patients tested by Guardant mainly for practical reasons — it would be hard to compare things like diagnostic yield if they were considering different tests with different panels of genomic targets — but also because use of other tests has been much more limited at the center so far.
In analyzing their 168-patient cohort, the researchers found that 58 percent (98 patients) had one or more ctDNA alterations, and among these 98, about 70 percent had one or more alteration that was potentially actionable by a US Food and Drug Administration-approved drug, mostly in an off-label context.
Of the 168 patients, 101 also had received NGS of their tumor tissue by Foundation Medicine, and so the researchers were able to compare the two to analyze concordance.
Since the tissue and plasma tests sequenced different gene panels, Schwaederle and her team only considered alterations in common for both tests, which encompassed all 54 genes in Guardant's panel (Foundation's test examines more than 300 genes). A total of 63 patients had a tissue alteration in one of these genes, but only 22 (35 percent) were in common between both tissue and ctDNA.
Interestingly, the time interval between tissue and ctDNA testing for these patients in whom the two samples concurred was significantly shorter — 2.7 months on average — compared to those for whom tissue yielded a result that ctDNA did not, or vice versa — 14.4 months.
In addition, when the group looked specifically at the most commonly mutated genes in the cohort, the concordance rates between the two methods was much higher — 70 percent for TP53 and EGFR, 88 percent for PIK3CA, and 93 percent for ERBB2 alterations.
According to Schwaederle, the main takeaway on ctDNA versus tissue concordance from her team's analysis is that it appears to be a two-way street. While the research field is accustomed to evaluating tests against a gold standard, in truth, neither tissue testing nor ctDNA is really the gold standard.
For instance, Schwaederle and colleagues wrote, "ctDNA tests can theoretically detect DNA shed from multiple metastatic sites, whereas tissue biopsy DNA tests would discern only alterations in the small piece of tissue evaluated. On the other hand, ctDNA tests may not be sensitive enough to detect alterations that are important and easily identified in a tissue test, and not all sites of disease may shed DNA into the circulation."
"Ideally they both should be used," Schwaederle said. "But if it's too risky to get a tissue biopsy, we think it's accurate to believe the ctDNA results alone."
Though the number of patients in the study was small, the group was also able to analyze data on the treatment that subjects ultimately received after getting their ctDNA results.
Within the overall cohort, 33 patients were ultimately treated with a molecularly matched therapy, while 39 were treated with another type of drug. The remaining patients were not evaluable for treatment after test results, mainly because they died before treatment, were only treated with surgery and adjuvant therapy, were still on a prior therapy, or were lost to follow up before treatment.
Because it was a retrospective study and because many subjects were tested using both Guardant360 and more traditional tissue sequencing or other molecular assays, Schwaederle said that it's impossible to know for sure that only ctDNA and not some other test result informed treatment in those patients for whom treatment was matched to some alteration in their ctDNA results.
Overall, 15 of the 33 patients who were treated with molecularly targeted drugs got a drug linked to an alteration specifically informed by the Guardant test. In the other 18, treatment must have been informed by tissue sequencing or some other assay, because the chosen drug targeted a gene or alteration not covered by the Guardant's panel.
Finally, the investigators were also able to evaluate outcomes for 12 of the 15 patients who got a potentially ctDNA-matched therapy. Three of these were too early in their treatment to assess, but five (about 42 percent) had achieved at least six months of stable disease by the study's analysis point.
Interestingly, in three of the five responders, the relevant actionable alteration was found first in ctDNA and later confirmed by tissue NGS.
In comparison, among patients treated with non-molecularly matched drugs whose outcomes could be evaluated, only about 7 percent achieved stable disease for more than six months.
Schwaederle and her colleagues admitted that the numbers of patients in this analysis was small, and prospective studies are needed to further define the value of ctDNA tests in guiding treatment.
Moving forward, she said that her team plans to continue following patients who receive ctDNA tests from Guardant and potentially other providers in a clinical setting.
Regarding clinical implementation at Moores, Schwaederle said that there is variability in adoption, with some physicians more keen to order this type of test than others, possibly because they are involved in other ctDNA research studies, or have experience with Guardant or another company.
Insurance coverage also likely plays a role, she said. "We don't have any specific guidelines at Moores, it's really up to the oncologist. But since I guess we began the experience with liquid biopsy here working with Guardant, I think most of them are using that. But now some clinicians are also involved in other studies with other technologies."
Another shift that will be interesting to track moving forward, Schwaederle said, is in the use of cancer sequencing tests — whether tissue-based or ctDNA based — earlier in the progression of a patient's disease.
"Oncologists before were waiting to the last minute to order any testing to match patients with therapies, after three or four or five lines of therapy and exhausting every other possibility," she said. "Now they are really ordering this testing earlier in the disease course, very close to diagnosis actually … and there is also more and more off-label use of drugs."
As these changes progress, so too may the statistics around the impact of testing on patient outcomes.