Karen Titus
August 2025—There’s the cart before the horse. There are certain equines that should not be flogged; others should not be backed. There’s closing the stable door at an unfortunate time. And in the case of multicancer early detection testing, there’s enough caution to temper excitement that it’s reasonable to suggest not putting the starting gate in front of the newborn foals.
When it comes to multicancer early detection, or MCED, testing, feel free to pick your own metaphor. The opportunity to detect any number of cancers before a patient is symptomatic, perhaps even years before another type of diagnostic test can do so, might be a paradigm shift. A sea change. Or a Pandora’s box.
Most immediately, however, call this the testing that’s launched a hundred questions.
“It’s an interesting area, and it’s a very exciting area,” says Sounak Gupta, MBBS, PhD, vice chair of oncology practice, Division of Laboratory Genetics and Genomics, and vice chair of research, Division of Anatomic Pathology, Mayo Clinic. “But I also find it very difficult to navigate because it’s so multidisciplinary.”
“Patients are certainly getting these tests,” says Jennifer Keating Litton, MD, MHCM, vice president of clinical research and professor and chair ad interim of breast medical oncology, University of Texas MD Anderson Cancer Center. “But what to do with them, and what they mean, is really unclear.”
The tests, broadly speaking, could enable patients to be screened for multiple cancers using liquid biopsies to detect biomarkers, such as circulating tumor DNA, and machine learning algorithms trained to identify the likely tumor or cancer signal of origin.
To date there is one commercially available (though not FDA approved) test, the Grail Galleri. It’s unlikely to be lonely for long. “Everyone is in a race to development,” Dr. Litton says.
There are any number of studies exploring the testing characteristics of MCEDs, she says. But these are early days. “Obviously the gold standard is overall survival,” she says, but producing such data takes years.
By comparison, mammograms have been shown to decrease breast-specific mortality; likewise, low-dose CT scans have lowered lung cancer mortality. “But so far, to my knowledge, we have a lot of tests that are promising, and showing sensitivity and specificity, but none of them have yet matured enough to show an endpoint other than testing characteristics,” she says.
MCED testing is not standard of care at MD Anderson, Dr. Litton emphasizes, for a very clear reason: Data do not support it as a standard of care. But given the high level of interest, the institution has “had to organize around the reality that patients are requesting it,” she says. This has fallen under the purview of MD Anderson’s longstanding Suspicion of Cancer Clinic, which was established to address the needs of patients who don’t have a specific cancer diagnosis but do have symptoms or concerns on, say, an imaging test that require a more complete workup. If a specific cancer is diagnosed, the patient is then referred to the appropriate clinic.
For now, says Dr. Litton, this is the most reasonable habitat for MCED testing. The goal was to standardize the approach as much as possible, given that there is nothing standard about MCED testing currently. “There’s no validated workup,” she says. “What we wanted to avoid was getting different answers from different providers” when test results were positive. “So we can standardize our approach—we’re seeing patients in the same place, so that we can follow them and do kind of standardized workups and standardized education. Then we can collect that data and learn from it.”
The organizational challenges alone are huge, she says. If a cancer is found, “I suspect that the [ensuing] workup might end up being tailored to the specific test, depending on how well it detects [each type of cancer] and at what stage.”
This is not being done around any one specific test, she adds, and she and her colleagues are keeping an eye on next-generation tests. “We’re following, very carefully, the data as it becomes available on multiple tests currently under development, to follow for those things that are most meaningful—changing outcome for a patient.”
Dr. Litton is particularly interested in MCED testing that could be used on patients whose cancers currently don’t have screening modalities. “That would obviously be a huge benefit,” she says.
But for now, the unknowns are huge. Among her questions:
- What are the endpoints that would justify putting this on a recommendation?
- What might be the unintended consequences of MCED testing?
- What do you do with a positive test result?
When she talks with colleagues, she says, “I think the biggest concern is that people might take this test, that’s yet unvalidated, in lieu of well-studied screening tests.” What happens if patients forego colonoscopies, for example, or mammograms? MCED testing is no substitute for such testing.
Eventually it may be. It’s possible that MCED testing could change outcomes, particularly for cancers that don’t now have good screening methods. If testing could detect cancers in a way in which cures were more possible, in a way that were accessible to all patients as part of routine health care, Dr. Litton says, “That would be wonderful.”
But the current reality, in her view, contains more worry than wonderful—that patients will avoid testing that shows actual improvement in outcomes. She also expresses concern about what she calls the “financial toxicity” of this testing—that patients could find themselves overwhelmed by the expense of both the test itself and subsequent workups.
“And then obviously the emotional drain of having a positive test,” she adds, “that then turns out to be a false-positive.”
CED testing has the potential to take a seat at the table and to upend the table completely.
MCED is just one of a handful of promising applications of cell-free DNA in oncology testing. Looking at the broader context, says Mayo’s Dr. Gupta, “Liquid biopsies are 10 different things, but most people have the impression that it’s one single thing. And even within cfDNA testing, there’s a lot of nuance between second-, third-, and fourth-generation tests.”
Multiply those factors across screening for multiple cancers, and the possibilities, and concerns, expand like a hot air balloon.
“This is complex,” says Dr. Gupta.
Traditional screening has been focused on five tumor types: breast, lung, colorectal, prostate, and cervical. All have clear-cut treatments once a mass has been identified.
The tables would start to shift if MCED testing—as has been proposed—were used for patients who are presymptomatic (a term that appears to be gaining ground). The closest comparison Dr. Gupta can draw is hereditary oncology, where, say, germline testing in a patient with a family history of breast cancer reveals a BRCA1 or BRCA2 mutation. In this case, physicians have had decades of experience to guide how they counsel and follow up with the patient.
Then there’s the cost component—and not only for patients. “Screening needs to be cost-effective,” Dr. Gupta says, given that it’s meant to be scaled up across entire populations. For now, however, although costs may come down, MCED testing can cost as much as $1,000.
Moreover, he says, there’s the cost benefit of screening in terms of reducing cancer-specific deaths. “If it’s marginal, is that investment worth it? That’s the mindset I would approach this with.”
Previous success stories have emerged from researchers synthesizing data to put together evidence-based screening programs. “The most difficult part of that,” Dr. Gupta says, “has been to generate data on cancer-specific mortality,” which can take a decade or two.
The pressure of these larger forces is substantial. The historic metrics of cancer screening can’t be tossed aside. Neither can disruptive technologies. “And many people are trying to compress timelines—understandably so—to bring these to clinic, to the general population,” Dr. Gupta says.
But compressed timelines and outcomes data can act like oil and water. In his specialty of prostate cancer, outcomes data can take five to 10 years to accumulate. “Can we wait that long for some of this MCED data to mature?” he asks. What is the best way to balance speed with scientific rigor?
Data are emerging, albeit slowly.
A Research Brief in the American Association for Cancer Research journal Cancer Discovery, for example, recently assessed use of an MCED test (Wang Y, et al. Cancer Discov. Published online May 22, 2025. doi:10.1158/2159-8290.CD-25-0375). The proof-of-principle study, the authors note, shows the possibility of detecting ctDNA more than three years prior to clinical diagnosis.
But early studies have focused primarily on testing sensitivity and specificity, as Dr. Litton notes. “While important, none of them have yet shown an endpoint that changes a survival or intervention outcome for patients.” Sensitivity and specificity are merely the starting blocks of a much longer race.
Dr. Litton and others voiced their concerns in an article published this spring (Guerra CE, et al. Am Soc Clin Oncol Educ Book. 2025;45[3]:e473834). And a New England Journal of Medicine Perspective column teased out the tensions of MCED testing, noting it is both alluring and bewildering to cancer screening experts (Robbins HA. N Engl J Med. 2024;391[4]:292–294).
Those tensions are familiar to Dr. Gupta. “Early cancer detection is just one piece of the puzzle.” In rereading a recent study of MCED platforms, he says, he noted that success was shown for patients who had a cancer recurrence. “I would say that should not have even been part of the trial,” he says. “That’s a completely different thing.” In short, when it comes to the data, a “Handle with care” sticker would not be out of place.
As the field considers how MCED testing might enter clinical practice, Dr. Gupta offers another obvious but critical reminder: Testing applications and approaches are specific to clinical scenarios. “I’m not sure we can cross those lines easily.” He gives the example of a lung cancer biopsy that doesn’t contain enough material for molecular testing, “in which case, a liquid biopsy becomes very attractive.” But such cases don’t belong in studies for early cancer detection, he warns. “It’s completely different.”
Similarly, Dr. Gupta says, minimal residual disease testing is a close cousin, but not an identical twin, to MCED testing. “The way those assays are designed are very different from early cancer,” he says. Likewise, polygenic risk scores are valuable, but the majority of them are focused on a specific cancer type and would likely not be applicable to multicancer detection approaches. And while cfDNA is used for early cancer detection, “Generally, again, it’s usually very tumor specific.” None of which is to say these approaches won’t evolve into MCED applications at some point. “But we’re not there now,” he says.
Dr. Gupta offers some ideas for thinking about MCED testing as more options become available.
“Let’s say we have a test that’s being proposed by a company or an academic institution. The first question is: How do we judge if this test is working or not?” Dr. Gupta asks.
The current paradigm has been to look at cancer-specific mortality. In the case of MCED testing, he says, the more appropriate endpoint might be all-cause mortality at the population level.
Additionally, most cancer screening approaches to date follow characteristics of each particular tumor type. Within cfDNA, the most common approach would be to look at individual, cancer-specific mutations. But with recent expansion into third- or fourth-generation liquid biopsy, he says, the type of mutations physicians look for is expanding.
One MCED test looks at methylation patterns, for example. “But you can also look at how the histone proteins sit atop the DNA fragments,” he says. “A gene that is being actively transcribed to mRNA may have less DNA that is protected by histone proteins, versus another gene that is not being transcribed that has a lot more histone proteins. These proteins influence how circulating tumor-derived DNA gets fragmented and has led to the field of fragmentomics—that looks at fragment sizes.”
It’s also important to consider that cfDNA is only one component of liquid biopsy, he says. Other elements to consider include circulating tumor cells, including exosomes. “So the complexity can go up on a log scale.” Tumor-agnostic tests leverage not only analytes that differ from cfDNA, including CTCs and exosomes, but also other, “for lack of a better word, -omic elements,” he says, including the aforementioned methylation and fragmentomics, in addition to genomics.
“That’s probably what makes a tumor-agnostic algorithm viable,” Dr. Gupta says. But the design is different from the genomics testing that has become relatively common in most clinical laboratories.
Another sticking point is the always-present divide between research-level data and clinical data. While some studies have shown promise in using methylation signatures for diagnostic purposes in neuro-oncology, the assumption that such signatures would translate easily across tumor types is just that—an assumption. “Again, we need the data,” Dr. Gupta says.
Which is why, he reiterates, it’s important for the scientific community as a whole—think of it as a panscientific community for pancancer testing—to become engaged. “Even for cell-free DNA-type testing, just for cancer profiling,” he says, “there are tumor types where studies have shown the efficacy of cell-free DNA testing may not be very high because the amount of circulating tumor DNA may be lower for a given tumor type.” It’s possible a similar scenario will play out in MCED. “Maybe this type of testing,” whatever the approach, “does not work as well for certain tumor types. And that has to be taken into account before we start marketing population-level multicancer detection testing.”
At the back end, Dr. Gupta continues, the metrics used to judge a new test’s success will need to differ as well. “That can get extremely nuanced,” he says. Like biblical lineage, testing begets testing.
He notes that one study looked at, among other things, how patients were followed up when the MCED test was positive. As Dr. Gupta recalls, the researchers estimated that, on average, it took some 160 days to resolve whether a patient did indeed have a cancer.
Says Dr. Gupta: “You have to put all this into context for what kind of tolerance we’re going to have for scaling up false-positive results, in terms of lab testing, imaging, biopsies, and other procedures, as well as patient anxieties. What’s the net benefit” as false-positives play out in real life?
False-negatives bring their own burdens. Echoing Dr. Litton, he asks, “If I get a negative result on one of these tests, and I have some symptoms, what would my perception be?” he asks. “Would I still seek care? Or would I assume that I had a cutting-edge test that told me I’m cancer-free, and just trust those results?”
Others in the field have raised those similar questions, he says. “People want to know what you do with an equivocal result. Does the patient come back? Do they get that testing again? Who pays for it? The patient? Medicare?” The economics of scaling up an expensive molecular screening test, and additional downstream testing, are not trivial. And without National Comprehensive Cancer Network and other guidelines, he says, there is no clear path for who should be screened. Should it be everybody? Should it be a specific population?
Though Mayo develops and performs liquid biopsy testing, it does not offer MCED testing currently. “In the absence of strong data, scaling up complex testing to the population level has many barriers,” Dr. Gupta says.
The larger challenge, he continues, would be clinical validation. While developing a test for minimal residual disease or molecular profiling is relatively easy to validate analytically, he says, “To show that a screening test works clinically [entails] large-scale, clinical-trial-type questions, which would be very challenging for an individual lab to pull off. So I think it will be very difficult to do this outside of a clinical trial setting”—ideally one that is multi-institutional.
“So at present we’re not actively developing multicancer early detection tests,” Dr. Gupta says. “But we are keeping a close eye on it, trying to follow the literature as it develops.”
They’re not the only ones. Patient interest is high. As Dr. Gupta notes, the American Cancer Society offers a patient-friendly Q&A on MCEDs on its website. His appraisal: “I think they’re being extremely cautious.” That’s fine with him. “Because at the end of the day, the scientific community—pathologists, oncologists, everybody—has to be collectively engaged. This is a big investment, from the entire medical community, and these tests should receive a high level of scrutiny.”
He continues: “I think as long as we have these conversations, as long as everything is in the peer-reviewed domain, that is the best way to move forward. It may not be the fastest way forward, but in my opinion, that’s the best way forward.”
Karen Titus is CAP TODAY contributing editor and co-managing editor.