Liquid biopsy technologies have made substantial headway in recent years, sparking booming commercial interest in the development of potential clinical applications.
Liquid biopsy technologies have made substantial headway in recent years, sparking booming commercial interest in the development of potential clinical applications. Some industry analysts expect the global market for these assays to grow to $12 billion by 2025, a 38% increase over this year.1
FDA-approved liquid biopsies using circulating tumor DNA (ctDNA) and circulating tumor cells (CTCs) have been available for several years,2,3 but they have not been widely adopted into clinical practice, with the exception perhaps of lung cancer diagnostics.4,5 Advancements in the field are now allowing test developers to fine-tune the technology and overcome several of the initial hurdles. The field is exploding with competing technologies vying for the ultimate goal of establishing clinical utility.5
Among the potential data about tumors that liquid biopsies could offer, information on detecting cancer at an earlier stage is the most alluring, particularly in tumor types for which no effective screening tests are currently available. In 2016, the FDA approved the first liquid biopsy for cancer screening, Epigenomics’ Epi proColon.6,7 This year, several other liquid biopsy platforms received breakthrough device designations from the FDA,8,9 suggesting that others may soon join Epi proColon in this niche.
As liquid biopsy options grow, so, too, will the complexity of what these tests measure. The evolving field presents a diagnostic language that goes beyond the information yielded through tissue biopsy.
Cancer Clues in the Blood
Tissue biopsies are the gold standard for gathering information about a tumor to aid diagnostic and therapeutic decisions, but their invasive nature means they are not always feasible, depending on the location of the tumor and the patient’s performance status. Furthermore, they capture only a snapshot of a tumor at the time they are taken and can fail to accurately reflect the intratumoral heterogeneity and tumor evolution that can fuel treatment resistance.10
By contrast, liquid biopsies are simple, minimally invasive, and less expensive and can be readily repeated to provide a clearer picture of the tumor over time.11,12
The key to liquid biopsies is that within that vial of blood may be some material derived from the tumor, which can offer clues about the nature of the tumor itself. Several types of tumor-derived material can be present, including cells, cell fragments, and cell-free DNA and RNA, as well as exosomes, which are tiny, fluid-filled, membrane-bound sacks that pass out of the cell to expel waste or transport cargo to another cell13-15 (Figure15). Each provides different and often complementary information about a tumor. Although investigators are studying all for their potential in liquid biopsies, to date the f ield has largely centered on the capture and analysis of ctDNA and CTCs.
Naked DNA, free floating in the blood— so-called cell-free DNA (cfDNA)—was first identified in the 1940s15 and is derived from both diseased and normal cells. Investigators believe it results both from active secretion from live cells and as a by-product of necrotic and apoptotic cell death.16,17
Although ctDNA has been detected in most patients with advanced malignancies, the frequency varies widely according to cancer types, ranging from more than 75% of patients with advanced pancreatic, ovarian, colorectal, breast, and other cancers to less than half of patients with brain, renal, prostate, and thyroid tumors.18 When present, ctDNA makes up just a tiny percentage of the total cfDNA in a patient’s blood.5 It can be distinguished from that background noise by the presence of molecular alterations that are characteristic of the tumor genome. To that end, analyses of ctDNA in bloodbased liquid biopsies fall into 2 categories: polymerase chain reaction—based tests that identify known mutations and next-generation sequencing assays that detect a broader range of alterations.4
Once a blood sample is collected for liquid biopsy, it undergoes centrifugation to separate the serum from the plasma. Although ctDNA can be found in both, in general plasma is the preferable source; as such, ctDNA is often referred to as plasma ctDNA or cellfree plasma tumor DNA.19 The reason is that serum has been shown to contain higher levels of cfDNA than plasma (all cfDNA, not just that derived from the tumor). Investigators believe this is because of the lysis of white blood cells in the serum, but it creates greater background noise, making it more difficult to isolate the small amounts of ctDNA.20
The first description of CTCs, meanwhile, dates back to the 1800s, when cells that were morphologically identical to those within the tumor were identified in the blood of a patient with metastatic cancer.21 CTCs are thought to be released into the bloodstream to facilitate metastatic spread and can be present either as single cells or in small clusters.22,23 They have now been described across a wide range of metastatic cancers.12
For the purpose of a liquid biopsy, CTCs are extracted from whole blood and can offer up a host of information on the tumor. One of the major advantages of CTCs over ctDNA is that all cell contents, beyond just the DNA, can be assayed to give a broader picture of tumor biology.11 The major challenge, however, is CTCs’ incredible rarity: On average, fewer than 100 CTCs are found in a 10-mL sample of blood, against a background of millions— even billions—of normal cells.24
To address the challenge, many platforms have been developed to isolate and enrich CTCs. To date, most are based on the idea of positive selection, which exploits the biological or physical properties that are specific to CTCs and absent on normal cells. For example, CTCs have a different size, density, and electric charge from those of normal cells, and they also express specific tumor-associated antigens on their surface. The most widely studied CTC is the epithelial cell adhesion molecule (EpCAM).25
Liquid Biopsies Now and in the Future
To make things even more challenging, the phenotype of CTCs is thought to be continually changing, and CTCs can lose expression of cell-surface antigens. EpCAM is downregulated when cells undergo epithelial to mesenchymal transition. An alternative strategy is to use negative depletion, removing all other cells from the sample until all that remains are the CTCs.12,26
A plethora of potential applications for liquid biopsies exist, although the clinically validated uses currently remain somewhat limited. Results of studies have consistently shown that both CTCs and ctDNA harbor the same patterns of mutations and other cancer-associated genetic and epigenetic abnormalities as the primary tumor.26,27 In the era of precision medicine, identifying patients with targetable alterations at initial diagnosis is increasingly important, and investigators are exploring liquid biopsies in this role.
In 2016, the Cobas EGFR Mutation Test v2 (Roche Diagnostics) became the first FDA-approved liquid biopsy. The agency approved it as a companion diagnostic to aid in identifying patients with advanced nonsmall cell lung cancer who have mutations in the EGFR gene that render them sensitive to treatment with the EGFR inhibitor erlotinib (Tarceva) as first-line treatment.3 Clinical trials found a high level of concordance between this test and tissue biopsy in identifying patients with EGFR positivity and demonstrated improved progression-free survival in these patients when treated with erlotinib.28,29 The test has subsequently been approved for use with a second EGFR inhibitor, gefitinib (Iressa), in the frontline setting and with the third-generation EGFR inhibitor osimertinib (Tagrisso) in first- and secondline settings.30,31
In addition to identifying individual mutations and alterations in panels of genes, liquid biopsy could also be used to evaluate tumor mutational burden (TMB). Results of studies have suggested the potential of TMB in identifying patients most likely to benefit from immune checkpoint inhibition across a wide range of tumor types.32
TMB has been evaluated typically using tissue biopsy; however, significant interest in developing a liquid biopsy for this purpose has emerged, and despite the complexities, it has shown promise.33 Several assays in development incorporate evaluation of TMB and other biomarkers, such as DNA damage repair deficiencies, into their readout.
Liquid biopsies, particularly CTC-based assays, at initial diagnosis also may be useful for prognostication. The number of CTCs present in the bloodstream has been shown to have prognostic significance in various tumor types. Three studies led to the first regulatory approval of a CTC detection platform, CellSearch (Menarini Silicon Biosystems).34-37
CellSearch captures CTCs on the basis of their expression of EpCAM, using magnetic particles coated with EpCAM antibodies, and then stains them with a variety of other f luorescent antibodies that help to further identify them as CTCs. It is FDA approved for the monitoring of patients with metastatic breast, colorectal, or prostate cancer; the presence of CTCs above a certain threshold predicts poorer survival.2,38
Changes in the levels of ctDNA and CTCs over time may also be indicative of prognosis as well as response to treatment. The level of ctDNA in the blood increases with growing tumor burden, and patients with distant metastases have higher levels of ctDNA compared with those with localized disease. Both CTCs and ctDNA decrease with treatment and rebound with recurrence. Liquid biopsies have been shown to detect recurrence prior to radiographic or clinical evidence of disease, which could give clinicians a bigger window of opportunity to adjust treatment regimens in real time.12,39
Because of the complexities posed in capturing CTCs, much of liquid biopsy development has focused on ctDNA, but various new CTC platforms harness microfluidic technology and do not rely on epithelial markers for isolation, features that could help to overcome the inherent heterogeneity that limits CellSearch.40
Investigators hope that in addition to more sensitive retrieval platforms, a different application of CTC-based liquid biopsies might lead to more acceptance in the clinic. They are evaluating the utility of characterizing CTCs to guide treatment decisions.
The Oncotype DX AR-V7 Nucleus Detect (Epic Sciences) test identifies the AR-V7 protein, a constitutively active androgen receptor (AR) splice variant that can drive resistance to AR-based therapies, in the nucleus of CTCs isolated from whole blood samples. Study findings published in 2018 demonstrated that high-risk patients who had this variant present in the nucleus lived longer when treated with taxanebased chemotherapy versus AR inhibitors. Patients without the variant, however, had improved outcomes when treated with AR blockade compared with those treated with taxanes.41
Early Detection of Cancer
Perhaps the most tantalizing potential application for liquid biopsies is in the screening and detection of cancer. However, investigators need to overcome many complexities, not the least of which is that the levels of ctDNA and CTCs are at their lowest in the earliest stages of disease.
In the past several years, studies have reported promising results on liquid biopsy platforms designed to achieve a place in the screening and detection of cancer. The FDA seems to share the enthusiasm, awarding multiple breakthrough device designations designed to facilitate more rapid development.8,9
The answer may lie in tests that look at parametrics beyond genomic alterations, as well as in the implementation of artificial intelligence (AI)—based analytical approaches. GRAIL, a spin-off from Illumina, is one of the leading companies in this space.5 Its multicancer early detection test is focusing on epigenetic modifications, specifically methylation signatures in ctDNA (ie, methyl tags added to the DNA molecule to alter gene transcription).
Results of studies have shown that the DNA methylation signature identified in a tumor tissue biopsy is very similar to that found via ctDNA analysis. More important, the advantage of looking at methylation signatures is that they tend to be less heterogeneous than mutations.42 Several studies of GRAIL’s platform are ongoing, including the Circulating Cell-free Genome Atlas study, which was designed to enroll 15,000 patients with and without cancer from 142 sites in the United States and Canada (CCGA; NCT02889978).
Results from a preplanned substudy were presented at the American Society for Clinical Oncology’s Breakthrough Global Summit for Oncology Innovators in Bangkok, Thailand, in October. The analysis covered findings from 2301 patients, including 1422 with cancer of more than 20 types. The test was able to detect a strong signal across cancer types, including notoriously tough-to-treat ovarian and pancreatic cancers, at early stages, with detection rates ranging from 59% to 86% and a specificity of 99% or more. The detection rates were 34%, 77%, and 84% for stage I, II, and III cancers, respectively. In addition, the test correctly identified the tissue of origin of the cancer in more than 90% of cases.43
A precedent exists for this kind of assay in cancer screening. Epigenomics’ Epi proColon, a blood-based screening test for colorectal cancer, was approved by the FDA in 2016 for average-risk patients over 50 years of age who are unable or unwilling to be screened with other recommended tests. Epi proColon detects methylation of a particular gene, Septin 9, rather than whole-genome methylation signatures.6,7
Meanwhile, Freenome is casting its net wider by using a multiomic assay that examines multiple parametrics simultaneously: mutations, methylation signatures and protein biomarkers from both tumors, and the immune system. Both Freenome and Grail funnel their results into AI-driven analytics.9,44,45 Recently published f indings demonstrated that among patients with colorectal cancer, this approach had a mean sensitivity of 85% at 85% specificity, with sensitivity generally increasing with tumor stage and larger tumor fraction.45 The prospective AI-EMERGE study in colorectal cancer, which aims to recruit 3000 participants, is ongoing (NCT03688906).