A few years in, a new picture for liquid biopsy​

Charna Albert

July 2021—Liquid biopsy has entered a more confident era, with two FDA-approved next-generation sequencing assays for comprehensive tumor mutation profiling, evidence of its clinical utility, and broadened patient access.

“Looking back, it’s remarkable how we’ve evolved,” says Geoffrey R. Oxnard, MD, vice president and global medical lead of Foundation Medicine’s liquid biopsy portfolio. Three years ago, he coauthored a joint ASCO-CAP review on circulating tumor DNA analysis in patients with cancer (Merker JD, et al. J Clin Oncol. 2018;36[16]:1631–1641). The tone of the review, he says, was enthusiastic but cautious. There was “clear excitement about the potential of liquid biopsy but uncertainty if the data would come together.” Today, “we have a lot more confidence,” says Dr. Oxnard, a thoracic oncologist who is also in the section of hematology and medical oncology, Boston University School of Medicine.

This confidence stems, in part, from the FDA’s approval in 2020 of two liquid biopsy tests—Foundation­One Liquid CDx and Guardant360. The approvals speak to the rigor with which the assays were developed and their strong analytic validity, Dr. Oxnard says. “I recognize that not every liquid biopsy is FDA approved and many labs offer lab-developed tests. But the ability of some liquid biopsies to make it through FDA approval highlights how the technology has matured.”

Also making a difference are the data from the BFAST, TRITON2, and SOLAR-1 trials. In the 2018 ASCO-CAP review, Dr. Oxnard and his coauthors cautioned that data on clinical utility were lacking. Liquid biopsy was being used organically and it was meeting a need for clinicians, he says. “But was that need leading to good patient outcomes? That is the second piece we’re now seeing come to fruition.”

Greater patient access is piece No. 3. “Liquid biopsy is now a reality. With FDA approval, these tests are covered by Medicare and increasingly by other insurers. They’re covered across oncology. It’s moving past lung cancer, and it’s creating access to next-generation sequencing wherever the patient may be.” And the NCCN recently updated its prostate cancer guidelines, he notes, to say that ctDNA testing should be considered when tissue is not available. “What I’m hearing from community doctors is that a bone biopsy at recurrence just isn’t a scalable solution for prostate cancer patients and liquid is actually a compelling approach to try first.”

Access to genomic analysis is at the heart of a study Dr. Oxnard and others coauthored, which found that in patients with metastatic castration-resistant prostate cancer (mCRPC), comprehensive genomic profiling of ctDNA is a “compelling clinical complement to tissue CGP” (Tukachinsky H, et al. Clin Cancer Res. 2021;27[11]:​3094–3105). Using plasma from 3,334 patients with mCRPC, they evaluated the landscape of genomic alterations detected in ctDNA and assessed concordance with tissue-based comprehensive genomic profiling. In the concordance analysis, 72 of 837 had BRCA1/2 mutations detected in tissue, 67 (93 percent) of which were also identified in ctDNA, including 100 percent of predicted germline variants. ctDNA identified more acquired resistance alterations than tissue.

“Let’s be forthright about liquid biopsy,” Dr. Oxnard says. “It can be a great specimen and it can be a bad specimen. Fundamentally, we don’t know how much tumor is in one. But it can be a great specimen.”

“It’s not always going to work out,” he says. “But in that patient with bone mets, you could have high shed, high signal. You could find the actionable variants you’re looking for. So let’s see this as an alternative option when tissue is not meeting our need.” Detection of BRCA mutations wasn’t perfect, he acknowledges. “But sensitivity was high, and we know that when we don’t find it in the blood, we do need to be ready to use tissue as a backup.” There are two ways to obtain answers for two kinds of patients, he adds: “When you’ve got a great [tissue] specimen, test that specimen, no question. And when you don’t have a good specimen, here’s the alternative to create access.”

Less clear, he says, is what’s indicated for a tissue specimen of uncertain quality and tumor content. “Should you try tumor first? Should you try liquid first? There’s still a lack of clarity there, and that’s where we need to collaborate with our pathology colleagues.”

Other advances of the past few years include the clinical trials like BFAST, the Blood-First Assay Screening Trial, which is the first prospective study to treat patients with advanced non-small cell lung cancer based on blood testing alone. ctDNA testing with FoundationOne Liquid identified ALK fusions at similar frequencies historically detected in tissue. Of 2,219 patients, blood-based NGS yielded results in 2,188 patients, and 119 had ALK-positive disease and 87 of those patients were enrolled and treated with alectinib, with an objective response rate of 87.4 percent (Gadgeel SM, et al. Ann Oncol. 2019;30[suppl 5]:v918). “When you treat patients based on liquid biopsy, you can get great results,” Dr. Oxnard says.

In the phase three SOLAR-1 trial, which evaluated the alpha-specific P13K inhibitor alpelisib in combination with fulvestrant in men and postmenopausal women who have PIK3CA-mutated HR-positive/HER2-negative advanced breast cancer, a significant progression-free survival benefit was seen regardless of whether the PIK3CA mutation was identified by a tumor tissue test or ctDNA test (Andre F, et al. N Engl J Med. 2019;380[20]:1929–1940).

“We very much see use in the community,” particularly in advanced lung cancer, Dr. Oxnard says of ctDNA testing. “What we are getting from academia is creativity in how to use it next.” He cites as an example liquid biopsy-based detection of BRCA reversion mutations in BRCA2-associated pancreatic cancer (Kondo T, et al. Pancreas. 2020;49[10]​:e101–e103). “That’s a space that’s scientifically interesting. The clinical story is still playing out.”

Despite greater access to and use of liquid biopsy, he says, current testing rates for actionable biomarkers are too low. The MYLUNG (Molecularly Informed Lung Cancer Treatment in a Community Cancer Network) Consortium chart review study, reported in June at the ASCO annual meeting, found that most metastatic NSCLC patients received at least one biomarker test before first-line systemic therapy, but less than 50 percent were tested for the recommended ALK, BRAF, EGFR, ROS1, and PD-L1. Testing with NGS was done in less than 50 percent of patients though it increased over the periods examined. “We have such a cumulation of evidence for multigene testing. It’s on us to better educate and communicate the reality of how to help patients,” Dr. Oxnard says. “It’s in our grasp but we need to deliver.”

A good liquid biopsy is sometimes richer in DNA than a tumor biopsy, Dr. Oxnard says. “Certainly it can be richer in tumor DNA than a poor-quality tumor biopsy.” FoundationOne Liquid CDx measures specimen quality using tumor fraction. When tumor fraction is elevated, he says, clinicians can trust and act not only on a positive result but also on a negative result. “If you have a great, rich liquid specimen, you don’t need to worry that the negative is a false-negative.” And “when it’s a high-shed specimen, it’s a more aggressive cancer—act on that negative result. When it’s a low-shed specimen that can be a less aggressive cancer, that’s when you want to follow through and get the tissue.”

His worry is that clinicians are using liquid on its own more and more because it’s convenient, and he wants them to understand, he says, that for a low-tumor-content specimen, reflexing to tissue may be especially valuable, whereas for a specimen of high tumor content, “the relative value of follow-through might not be as high.”

“It’s also appropriate to point out that we need to do better” at improving test sensitivity, he says. “What’s great about the constraints of an FDA approval is there is an expectation of continuous improvement and an expectation that as you evolve the chemistry and algorithms you can improve sensitivity.”