The complexity of cancer can only be better understood and treated with a complex array of tools. In one case, cancer scientists use a single approach—circulating tumor DNA (ctDNA)—in many ways. Plus, even more uses of ctDNA lie ahead.
Even deep inside a person’s body, DNA can escape from a tumor and enter the bloodstream. If that ctDNA can be detected at a sensitive enough level, it can be used across much of oncology, from early detection and treatment selection to surgical follow-up and the likelihood of relapse. As Mark Sausen, PhD, vice president of technology innovation at Labcorp subsidiary Personal Genome Diagnostics (PGDx), says, “ctDNA can be used for a number of biomarker development and diagnostic applications in oncology.”
One key element in using ctDNA with cancer will come from collecting more data to confirm what insight such measurements might provide. “We are seeing a significant increase in compelling clinical data to support the use of ctDNA for the earlier detection of cancer compared to existing screening modalities, where curative intent intervention is possible with diagnosis occurring earlier in the cancer continuum,” Sausen says.
That data will be used to make ctDNA even more useful in the clinic. “Liquid biopsies that analyze ctDNA in blood are starting to become accepted into regular clinical oncology practice, and some would even consider these to be standard of care in certain circumstances,” says David Eberhard, MD, PhD, chief medical officer of Inivata. “A liquid biopsy as an alternative to tissue biopsy for identifying actionable mutations is the most well-established application.” Using ctDNA as a liquid biopsy, however, depends on very sensitive detection. For cancer screening and early detection, “these liquid biopsy tests are focused either on specific cancers or are designed to be pan-cancer,” explains Richard Chen, MD, chief medical officer and senior vice president of research & development at Personalis. “The key challenge here is that current solutions have modest sensitivity.”
Detecting DNA junctions
At the Mayo Clinic in Rochester, MN, George Vasmatzis, PhD, co-director of the biomarker discovery laboratory, explores the junctions in DNA created from genomic rearrangements in tumors. “Rearrangements happen in more than 90% of the tumors,” Vasmatzis says. “It is one way that the tumor evolves to alter genes that can be used to advance the tumor.” Most important from a clinical perspective, specific junctions can identify one patient’s cancer. “Every patient has their own unique set of junctions,” Vasmatzis emphasizes.
To make use of this information, Vasmatzis sequences a patient’s tumor to find the junctions. Next, it takes just a few days to make primers to those junctions, which can be identified with a PCR assay on a blood sample. This assay can be used for both detection and quantification of junctions.
Consequently, the approach developed by Vasmatzis can be used in many ways in oncology. For example, it could be used to monitor a patient for a relapse. It could also show if a treatment is working. If it’s not, the treatment could be changed.
Measuring residual disease
In searching for the recurrence of cancer, many companies look for tumor-specific mutations. This can be done with a standard panel of likely tumor-based mutations or developed to look for a patient’s own cancer mutations.
“The tumor-informed approaches can have a big advantage over tumor-naive approaches, because you are creating a personalized liquid biopsy panel that is specifically targeted to the unique mutations in each patient’s tumor, allowing for detection at high sensitivity,” says Chen. So, Personalis uses a tumor-informed approach with its NeXT Personal product.
With NeXT Personal, whole genome sequencing identifies up to 1,800 specially-selected somatic variants in a patient’s tumor to create a tumor-specific assay. “The targeted panel is then used to sequence millions of unique DNA molecules from a patient’s blood sample,” Chen says. “Aggregation of the signal across these sequences using proprietary algorithms enables us to achieve part-per-million sensitivity.”
Credit: Personalis
In the past, oncologists estimated the likelihood of recurrence after surgery with pathology and other methods, such as the TNM system, where the T represents the tumor, N is whether or not the cancer has spread to the lymph nodes, and M describes a cancer’s level of metastasis. This information can be combined to estimate a cancer’s stage, which correlates with the odds of recurrence. For any individual patient, though, this was not a very accurate indicator, but a ctDNA-based assay can be much more accurate. “For the first time, you can say that a patient is likely cured or that the risk of recurrence is 100% without additional treatment,” Alexey Aleshin, MD, head of oncology medical affairs at Natera, which makes the ctDNA-based Signatera test.
In January, at the American Society of Clinical Oncology’s 2022 Gastrointestinal Cancers Symposium, Natera presented interim results from more than 1,000 patients in the GALAXY cohort of the CIRCULATE-Japan trial. “Patients with stage-two or -three colon cancer who have a positive Signatera test have a hazard ratio around 13 times more likely to recur than those with a negative test,” says Aleshin. “Also, patients with a positive test who did not receive adjuvant therapy were nine times more likely to recur, which is pretty profound.” Plus, patients with a negative Signatera test do not seem to benefit from adjuvant therapy, regardless of the stage of their cancer.
Other companies are also developing very useful ctDNA-based liquid biopsies. As an example, Inivata’s RaDaR is intended to detect low levels of ctDNA in a blood sample after a treatment. “In the LIONESS study of patients with head and neck squamous cell cancer, RaDaR could detect ctDNA at levels as low as 6 parts per million,” says Eberhard. “At the same time, RaDaR is highly specific.” He adds, “In the published LIONESS cohort of stage III-IVB patients, there were no false positive results for predicting later tumor recurrence.”
could be used across cancer care. Credit: Inivata
Some scientists even use ctDNA liquid biopsies to predict a patient’s survival. One team of scientists from Genentech, Roche Sequencing Solutions, and the University of California, Davis used ctDNA liquid biopsies from patients who were being treated for non-small-cell lung cancer.
In 94 patients, the scientists measured allele frequency and mutant molecules per milliliter of blood at the start of the study and then every three weeks. The results showed that the level of ctDNA mutant molecules per milliliter of blood six weeks after the treatment correlated with a patient’s likely overall survival. That information could be used by clinicians, in consultation with patients, to select the best course of action.
Personalizing monitoring and care
For years, even a couple of decades, many advances in treating cancer revolved around the phrase personalized medicine. In many cases, though, personalized treatment meant a group of patients, sometimes many of them. Scientists at Gritstone bio hope to change that—zeroing in on treating patients as individuals.
Using its proprietary EDGE platform, scientists at Gritstone bio use a tumor biopsy, sequencing, and artificial intelligence to find cancer-related antigens that could be targeted with a personalized vaccine. The company’s first product, GRANITE, is being tested in early-stage clinical trials for patients with microsatellite stable colorectal cancer or gastro-esophageal cancer. “In GRANITE, we primarily use ctDNA for monitoring during treatment,” says Matt Davis, PhD, director of molecular biology and sequencing at Gritstone bio.
This monitoring of the ctDNA proves especially important with immunotherapies. As Karin Jooss, PhD, executive vice president and head of R&D at Gritstone bio, explains, “If the T cells recognize the tumor, they infiltrate, and sometimes your next scan might show an increase in the tumor’s size—termed pseudoprogression—because of the T cells.” Without ctDNA monitoring, such an increase in tumor size could be interpreted as treatment failure, but just the opposite could be true. An increase in the tumor’s size soon after treatment could mean that the T cells are targeting and attacking the cancer, but that can only be confirmed with a decrease in ctDNA from a liquid biopsy. Even Gritstone bio’s ctDNA monitoring is personalized based on a patient’s tumor. “We are able to track the evolution of a patient’s unique targeted neoantigens, passenger and various cancer drivers—both clonal and subclonal—over time,” says Davis. “This is very important for understanding the genome of late-stage cancers and how they evolve with a very targeted and personalized therapy.” This information could be used to change a treatment or to adjust its dose. “You can actually personalize your treatment regimen,” Jooss says.
All of the information that Gritstone bio collects can reveal details of the molecular interactions between a treatment and a tumor. For example, a scientist can explore how a tumor escapes the impact of treatment if the ctDNA starts to increase. “The team is able to look at the mechanism of escape,” Jooss says. “What did the tumor down regulate or up regulate?” Then, Gritstone scientists look for ways to take away the tumor’s mechanism of evading the treatment. “We can explore combination treatments that don’t allow the tumor to use this mechanism,” Jooss says.
Other companies also make use of advanced analysis in assessing ctDNA-based data. When asked about PGDx’s latest advances in using ctDNA as a liquid biopsy, Sausen says that it’s the “development and application of machine learning-based algorithms for detection of ctDNA, including clinically actionable biomarkers associated with sequence and structural alterations, along with genomic signatures.” He adds that this “has enabled significant improvements in analytical performance as well as fully-automated the downstream analysis of raw sequencing data, allowing decentralized access to comprehensive genomic profiling with the PGDx elio platform.”
Driving better drug testing
As already shown, companies use ctDNA-based liquid biopsies to assess the impact of cancer treatments, but the breadth of that application in clinical trials is only getting started. Two of the key improvements will come from faster and more economical drug testing.
Tools made for the clinic can often be used in clinical trials, as well. “In addition to the clinical advantages of NeXT Personal, there are significant advantages for biopharma monitoring patients in clinical trials,” Chen says. “For biopharma companies, the greater sensitivity can lower the time and costs of clinical trials.” As an example, a clinical-trial sponsor can use a liquid biopsy to enroll patients who have ctDNA in their bloodstreams. “Having 10 to 100 times higher ctDNA sensitivity means that one does not need to screen as many patients to find positives, therefore lowering time and costs of initiating the clinical trial,” Chen says. During the clinical trial, such a ctDNA-related tool can also be used to monitor the response of patients to the treatment being tested.
In running trials for potential cancer vaccines, Jooss points out: “If you don’t have any way to stratify the high-risk patients, you’re running really large trials, and that has been the problem.” The more patients that a clinical trial requires, the longer that it will take to run it and the more it will cost. “Biotech just never was able to run these studies, because it was just a huge undertaking in the past, but now we can select actually high-risk patients, and that will be great for them,” Jooss says.
The Phase III ZEST study, for example, enrolled 800 patients with specific forms of breast cancer. These patients had already had a surgery or adjuvant therapy. In this trial, the Signatera test identifies the patients with ongoing levels of ctDNA, and they will receive a follow-up treatment with niraparib or a placebo.
By using a ctDNA-based test, a clinical-trial sponsor can “really selectively pick a group of patients with exceedingly high risk of recurrence, because they still have active disease in their body,” Aleshin explains. “If you can see that disease, you can improve those patient outcomes.” He adds: “This is a win-win for everyone that allows for much quicker approval of these drugs in the adjuvant setting, which traditionally required studies that take forever because you’re treating a lot of patients who are unlikely to benefit from the treatment—ones who are already cured.” With a ctDNA-based test, patients who need further treatment can be identified, and so can the ones who would not benefit from additional cancer therapy.
Developing more data
The future of ctDNA in clinical trials and the clinic depends on developing an extensive database of information. When asked what improvements in using ctDNA as a liquid biopsy would be the most beneficial and why, Eberhard says, “At this point, we need solid clinical data demonstrating how to best use liquid biopsies in ways that can guide the most therapeutically-effective and cost-effective cancer management decisions for patients and oncologists, by guiding the choice of optimal therapies, by detecting cancers early so that they are more effectively treated, and by being accessible to everyone who needs them.”
In addition, ongoing improvements in the capabilities of ctDNA-based liquid biopsies will expand the uses of this approach. “From a technical perspective, continued improvements in analytical performance will enable a broader range of diagnostic opportunities across early- and late-stage disease settings,” Sausen says. “From a clinical perspective, the increasing utilization and publication of ctDNA-based analyses from investigator- and industry-sponsored clinical trials will help to further inform the clinical utility and actionability of ctDNA-based liquid biopsy analyses, driving adoption and utilization to further improve patient outcomes.”
Although ctDNA can be measured in many ways, this general approach provides a tool that can be used across oncology, from the early detection of cancer to monitoring the disease after treatment. Equally important, ctDNA-based methods can be optimized for a specific patient. At some point in the future, ctDNA-based methods could benefit everyday patients around the world. Such a test might become easy enough and affordable enough to be included in an annual physical. In that way, people with no symptoms of any sort, might be diagnosed and treated far earlier than with traditional methods. As a result, oncologists might finally achieve the longstanding goal of detecting cancers at the very earliest stages, all with the hopes of treating the disease as effectively as possible, and saving more lives.
Mike May, is a freelance writer and editor with more than 30 years of experience. He earned an M.S. in biological engineering from the University of Connecticut and a PhD in neurobiology and behavior from Cornell University. He worked as an associate editor at American Scientist, and he is the author of more than 1,000 articles for clients that include GEN, Nature, Science, Scientific American and many others. In addition, he served as the editorial director of many publications, including several Nature Outlooks and Scientific American Worldview.
