It’s an aspirational goal: develop a simple blood test that can detect cancer cells in an asymptomatic patient, locate the tissue of origin, and give that person an early diagnosis when treatment is more likely to be successful. The ideal test would be able to
do all that and do it with nearly 100 percent accuracy. The sequencing technology and advanced computing methods are already available. They have been used to create liquid biopsy tests that are being used in people with advanced cancer for therapy selection and to monitor disease progression. Next-generation sequencing allows for the analysis of minute quantities of cell-free tumor DNA (cfDNA) circulating in the blood. Machine learning algorithms are being used on large data sets to characterize the molecular signatures of various cancer types. These new tests represent minimally invasive and less costly alternatives to traditional tumor biopsy. Now, researchers are turning their attention to creating affordable, minimally invasive early detection tests (EDTs) based on liquid biopsy technology.
Guardant Health has been in the liquid biopsy business since 2014 when it launched Guardant360, a next-generation sequencing cfDNA assay for treatment selection in advanced solid tumor cancers. The test can be used in multiple tumor types, but is most frequently ordered in advanced lung cancer. From the beginning, developing early detection tests has always been part of the company’s game plan. “When we started Guardant, we imagined that some day patients would go through cancer screening at their annual check-ups,” said Amirili Talasaz, Ph.D., the company’s president and COO. “Our philosophy from the start was to begin with advanced cancer patients, understand the circulating tumor DNA, understand the biology better over time, and, in a systematic way, go to earlier stage testing.”
Of all the potential clinical applications of tumor-derived cfDNA, early detection is one of the most ambitious. The tests rely on the fact that mutations in tumor-derived cfDNA are highly specific markers for particular cancer types. This specificity is key in order to minimize false-positives. Large case-controlled studies are required to establish sensitivity and specificity to ensure accurate identification of patients with early-stage disease. The effort presents unique computational challenges, requiring multi-disciplinary teams working together. Researchers are rapidly developing new computational methods that increase the accuracy of NGS. Creating pan-cancer EDTs, ones that test for the most com- mon cancer types, also means of improving on currently liquid biopsy tests by being able to accurately determine the location of a tumor within the body.
Improving sensitivity
In December 2018, Guardant launched its LUNAR assay for use in biopharmaceutical and academic research for the detection of early-stage cancer and recurrence of disease in patients with lung, colorectal, ovarian, and breast cancer. (A clinical version of the test for investigational use only in prospective studies is expected to launch in the second half of 2019.) The technology is based on data from testing 80,000 advanced cancer patients using Guardant360. It includes all genes recommended for profiling by the National Comprehensive Cancer Network, including the 73 genes most relevant to clinical care. The test can detect tumor DNA in the blood with 90-95 percent accuracy.
In addition to Guardant360-generated data, the company also based LUNAR on whole-genome sequencing liquid biopsy data and information from publicly-avail- able databases, such as the TCGA and PubMed. The new test uses epigenetic data to increase sensitivity. “We realized that if we just looked at genomic data in blood, we were going to have limited sensitivity detecting tumor signatures at an early stage,” Talasaz explained. LUNAR includes assessment of methylation and DNA fragmentation patterns—ones that vary between normal cells and cancerous ones. Research findings using LUNAR, which provides results in seven days, are showing promise, Talasaz said. Findings from a colorectal cancer study that included 105 recently diagnosed cancer patients and 125 cancer-free, age-matched healthy patients were presented at the American Association for Cancer Research annual meeting in March 2019. “With just using a single blood test in patients that have been recently diagnosed, we can detect cancer with 90 percent sensitivity.”
GRAIL is also among those companies developing a pan-cancer EDT. Its prototype cfDNA sequencing assay detects a long list of cancer types: colorectal, lung, esopha- geal, head and neck, liver, ovarian, pancreatic, lymphoma, multiple myeloma, and certain aggressive breast cancers. “Many of these cancer types are not screened for today and often present at late stages,” said Kelsey Grossman, a spokesperson for GRAIL. Grossman also said that GRAIL has developed the capability to detect tissue of origin and will present that data at a future medical conference.
Optimizing EDT assays involves using bioinformatics and machine learning algorithms on large molecular data sets. Ultimately, GRAIL will collect data from 165,000 people. Its Circulating Cell-free Genome Atlas (CCGA) study is a prospective, observational, longitudinal study designed to characterize the landscape of genomic cancer signals in the blood of people with and without cancer. The study is fully enrolled with approximately 15,000 participants across 142 sites in the United States and Canada. The company’s STRIVE study is fully enrolled with approximately 100,000 women at the time of their screening mammogram and the SUMMIT study will enroll approximately 50,000 men and women age 50 to 77 years who do not have a cancer diagnosis at the time of enrollment.
These large numbers are needed because the fraction of tumor-derived cfDNA in the blood is very small compared to that of normal cells. “An early detection test needs to be backed by a body of evidence involving tens of thousands of participants to distinguish true cancer signals from biological noise and predict the cancer’s tissue of origin when cancer is present to enable doctors to appropriately direct next steps for diagnosis and care,” Grossman said. It is important that EDTs have a very low rate of false positive results. “While current screening options are very good at detecting certain cancers, false positive results are frequent, causing people anxiety and many unnecessary follow-up procedures. High false-positive rates often limit the use of screening tests.”
Seeking cancer
While GRAIL and Guardant are working on pan-cancer assays, PapGene is taking a stepwise approach. In 2018, the company was granted Breakthrough Device Designation from the FDA for an EDT that screens for ovarian and pancreatic cancer. Like most EDTs, the assay uses a combination of tumor-derived cfDNA and protein biomarkers to detect cancer in individuals over the age of 65. “This is the first step toward our vision of making pan-cancer detection a reality for patients and dramatically improving cancer survival rates,” said PapGene CEO How- ard Kaufman at the time in a statement. The com- pany’s test is based on the technology behind CancerSEEK, a pan-cancer EDT developed by researchers at the Johns Hopkins Kimmel Cancer Center. In 2018, the cancer center’s researchers published a paper in Science describing an EDT that tests for eight common cancer types: ovarian, liver, stomach, pancreas, esophagus, colorectal, lung, and breast cancers. Screening tests are currently available for only the three of these cancers: breast, colorectal, and lung. Researchers explored several hundred genes and 40 protein markers, eventually settling on 16 genes and eight proteins. They also reported that the test, which was evaluated on 1005 patients with non-metastatic cancers in stages I to II, had a sensitivity that ranged from 98 percent for ovarian cancer to 33 percent for breast cancer. Unlike available liquid biopsy tests, CancerSEEK accurately determines the location of the tumor. Researchers estimated that the future commercially available test would cost about $500. “This test represents the next step in changing the focus of cancer research from late-stage disease to early disease,” said Bert Vogelstein, a professor of oncology and one of the authors of the Science paper, in a press release.
Fast-moving field
Not all liquid biopsy companies are trying to move into the EDT space, but are instead developing new applications for liquid biopsy in patients with cancer at various stages. “We still see that there is a lot of work for us to do in the advanced cancer space,” said John Simmons, Ph.D., vice president of translational medicine at Personal Genome Diagnostics (PGDx). For example, the company has early development programs focused on detecting minimal residual disease and monitoring. “Physicians want to know information about what kinds of resistant mechanisms have emerged and whether those are targetable by approved drugs or eligible for clinical trials,” Simmons said.
In March, the company announced that it received European safety certification, the CE mark, for its PGDx elio plasma resolve. The test is the first kitted plasma-based NGS oncology test to have that certification. The qualitative test uses NGS to detect single nucleotide variants, small insertion/deletions, amplifications, rearrangements, and microsatellite instability in a broad multi-gene panel in circulating cfDNA isolated from blood samples. It includes several clinically actionable variants across tumor types, enabling more informed treatment decisions. “The majority of patients with advanced stage cancer are not receiving any NGS testing,” Simmons said. In fact, he added, many cancer patients are not having the complete compendium of molecular tests performed that could direct their therapy.
Foundation Medicine, like PGDx, is still focusing their efforts on using liquid biopsy for treatment selection and has not made public any plans to move into EDT development. “The immediate impact for liquid biopsy today is one of looking for therapy selection and, in some regards, that also refers to looking at the development of resistance,” said
Prasanth Reddy, M.D., Foundation Medicine’s VP of medical affairs. Reddy also said that an important application for liquid biopsy is monitoring tumor burden to determine therapeutic efficacy.
In 2018, the company announced the launch of FoundationOne Liquid. The test, which requires two 8.5 ml samples of blood, provides genetic information and microsatellite instability (MSI) status for use in treatment selection. It analyzes 70 genes known to drive cancer growth, including homologous recombination deficiency genes, and reports the genomic biomarker for microsatellite instability to help inform the use of checkpoint inhibitor immunotherapies and multiple targeted therapies, including poly (ADP-ribose) polymerase (PARP) inhibitors. The test allows physicians to gain insights into the molecular drivers of cancer in an individual patient. “We can use that information to determine whether a potential treatment may be suitable for a specific patient at a specific time in their cancer journey,” Reddy said. As for EDTs, Reddy said: “There are some things that need to be worked out on the clinical end, (including) how to integrate very sensitive ways of detecting potential tumors in a patient.” For example, physicians will need to know what to do if a patient tests positive using an EDT, he said.
In general, keeping oncologists informed about the lat- est technology is a common challenge in the field of liquid biopsy, according to Guardant’s Talasaz. “One of the bottlenecks of precision oncology is physician education about new technology in the field, how much we know about the genomics and the molecular properties of tumors, and the novel therapeutics that are commercially available. There are some gaps there. It’s a field that is moving very fast.” Guardant’s strategy has been to build collaboration with key opinion leaders and generate lots of clinical publication on liquid biopsy. More than 110 studies have been published that used the Guardant360 test and collaborators have used it in 40 clinical utility studies to-date. “The that thing really helped us is that the unmet need is so huge. Physicians were hoping for this to work.”