NEW YORK – The oncology community remains optimistic about the potential of genomic tests to improve early cancer detection and screening, and hopefully shrink the morbidity and mortality associated with a cancer diagnosis. But as evidenced at a session at the annual meeting of the American Association for Cancer Research this week, clinicians are increasingly struggling with patient interactions as commercial assays have become clinically available despite a lack of solid utility data.
Speaking as part of a panel on the subject, University of Pennsylvania physician and ethicist Angela Bradbury said that she has already had "a couple of patients" come into her office curious about Grail's Galleri assay.
"These tests are commercially available. There's at least one … and this is despite ongoing trials evaluating their effectiveness and clinical utility," she said. "We are bound to have patients if you have not already … come in and ask about these tests, is this something they should have."
Showing an image of Galleri informational material on the web, Bradbury describe a potential patient interaction. "They might pull out their phone and say, 'Look at what I found about these. Why would I not do this?'" she added. "Detect cancer early when it can be cured? If you're a patient reading this, it's very exciting."
She further noted that she works in a high-risk clinic, "so I have patients who are like, 'Oh my God, why have you not recommended this to me?' There's also this component of trust that might be lost when patients see something that looks so exciting and it has not been something recommended to them."
For Bradbury this isn't necessarily something new. Over her career, she's encountered the same struggle with other new technologies that lacked clinical utility data but were still commercially available, and she described the issue as a product of the US Food and Drug Administration's tradition of not enforcing its authority over laboratory-developed tests.
"I keep thinking why do we keep doing this? … and I think part of the challenge is that we have two different pathways for approval," Bradbury said. "If my patient comes to me and says, 'Oh, there's a new drug for ESR1 mutations. When can I get that?' It's very easy because I say you have to wait for FDA approval."
The same process is not in place for diagnostics, but a patient encountering a commercially marketed test would have no reason to know that.
Philip Castle, director of the Division of Cancer Prevention at the National Cancer Institute, argued that the LDT pathway was not intended for commercial tests, especially population screening tests, but rather for specialized tests in small labs that would "never have commercial value."
Physicians, with a mandate of both magnanimity toward their patient and the pledge to do no harm, find themselves between a rock and a hard place, Bradbury added, especially in regard to screening, where a large part of evaluating whether something is clinically useful is knowing the overall benefits and harms.
Jamie Brewer from the FDA's Office of Oncologic Diseases described the same tension from a regulatory standpoint.
"While there may be a desire to make sure that all individuals who could benefit from early cancer detection have access to this new technology, it's also important to ensure that these tests are used appropriately and that we are continuing to learn more about performance [and limitations]," she said.
Castle described the current moment as a mix of hype and hope, something he said stakeholders know that they must sort out.
"When we say clinical utility, it's really talking about the assessment of full benefits for the population. And that has to be done," he said. "We have to look at benefits and harms so that we can understand what these tests do and who they actually benefit, which may not be equal across the population. We have to do our due diligence."
According to Bradbury, these necessary answers have certainly not yet come in for MCEDs like Galleri. For example, she said, it is not yet definitively known whether these tests will shift what would be late-stage diagnoses to earlier stages, let alone early-enough stages to allow curative treatment.
"Cancers that are already shedding in the blood may not necessarily be early-stage … and we're already seeing that in some of the data, so I think there's still a question as to whether this is going to be as promising in reducing mortality as it has been claimed to be," she said.
Bradbury also cited a modeling study, published last year by Fred Hutchinson Cancer Center professor Ruth Etzioni and colleagues, which called into question the overall impact of shifting cancer detection to earlier stages.
Companies in the space have used modeling to propose a dramatic positive impact on cancer mortality in the context of widespread MCED implementation.
But according to Etzioni and her team, while stage shifting is linked to substantial improvement in mortality for some tumor types, other cancers are unlikely to see significant outcome improvement even in the context of substantial, or even near-total stage shifts, the group wrote. Moreover, expected mortality reduction based on cancer survival distributions has often failed to match the actual mortality outcomes of published trials.
"There's been a lot of hand-waving to justify these [tests] that is not scientifically valid. So just caveat emptor, buyer beware of what's being put out there," Castle said regarding modeling data.
Doctors have also questioned the emotional or psychological impact of MCED testing on patients, considering the uncertainty that they could face with a positive result. If follow-up scans don't identify a tumor, are patients rescanned? If so, how often?
"What is the impact on patients of that uncertainty and sitting and waiting to know whether this signal meant anything or maybe it didn't mean anything?" Bradbury asked.
MCED test makers are working hard to answer these questions in large studies like Grail's PATHFINDER. But Bradbury spoke to the frustration among physicians being put in a position where they must face patient interest in these tests before all the data is in.
"I think there are those that would argue [that] patients understand if you tell them the risks and benefits … you could argue out of respect for a person's autonomy [and] help our early adopter patient get access to something that they want," she said. "But we also need to remember … that as healthcare providers, and really as members of the public, that we also have to think about justice and equity and being stewards of the healthcare system and being cautious about costs and making sure that we do that equitably."
In her own practice, Bradbury said she advises patients to have patience. "I treat patients who are at high genetic risk or at risk based on family history … and [we] have been waiting for something like this," she said. "But what I would say right now is this is not ready right now, but hopefully soon. Hopefully we can have a clinical trial or perhaps a registry that you could enroll on so that we're collecting generalizable knowledge for the next generation."
Hoping to resolve this clinical uncertainty, the NCI recently developed a plan for a study it calls Vanguard, which NCI's Castle described as an initial feasibility study that would then transition into a large randomized controlled trial.
"The basic idea is can we operationalize this? Can we get people randomized? There are some real logistical issues that have to be worked out before we scale up," he said.
The first goal will be to enroll about 24,000 people, about one-tenth of what NCI hopes it can then build up to in a second stage. The plan is for the study to have at least two MCED arms and standard-of-care control arm, and to analyze the death rates for the targeted cancers of each MCED compared to the control arm.
"It's not a MCED-versus-MCED trial because that's almost impossible to do because they target different cancers and we don't have enough sample size, except for maybe lung cancer, to do individual cancer or even groups of cancer. So those are the realities of this," Castle said.
If there is enough money and a third promising MCED test, it would be great to add another arm, he added. "Our really ideal situation is that we [evaluate enough tests] that we feel like we've validated at least one of them into the marketplace," he said. "How many of those will it take? We don't know. I mean, if we knew how well these worked, we wouldn't be doing this, so it's a little bit of a black box for us as well as for you."
Investigators are also going to study patients' willingness for randomization and adherence to testing and diagnostic follow-up, which Castle called a key aspect, noting that "we shouldn't ever talk about screening without the completion of care, and I shouldn't even have to qualify that."
Castle said that he and his team are also thinking about adding some high-risk cohort trials in parallel to Vanguard, as well as collecting real-world evidence, not just to evaluate tests in terms of clinical outcomes but also to better understand pathways to care and barriers to access.
A first step for NCI in getting the ball rolling has been to build a population screening network that can effectively get representative populations onto screening trials. This has progressed rapidly, Castle said, with applications in for an initial 10 to 15 accrual enrollment sites, as well as a coordinating center and statistical data management center.
According to Castle, this probably won't be enough to support a trial the size of what NCI hopes to do in the expanded Vanguard follow-up, but there are other ways to support that planned trial. He further noted that NCI is in discussions with the US Department of Veterans Affairs, US Department of Defense, and even colleagues in the UK about an expansion cohort "to really accommodate a much larger study when the time comes."
University of Colorado professor Marie Wood discussed progress her team has made in putting together a blinded reference set that the NCI can then use to verify performance characteristics of the MCED tests that go forward in the Vanguard study or other efforts.
Eligibility for both cases and controls in this trial is between ages 40 and 75 to match current screening parameters for the control group of individuals with no prior cancer diagnoses and no signs of disease.
"We've had pretty good success getting [cancer case subjects] enrolled prior to any definitive treatment," Wood said, noting that they began the study in August 2022, and as of April 10 have enrolled almost 1,500 patients comprising 1,000 controls and almost 500 cases.
Although the goal is to achieve a 30 percent accrual of underrepresented minorities, the population has been largely white so far, but Wood said the team has been "slowly moving the needle" and will enroll another 1,000 patients solely from underrepresented groups.
FDA's Brewer said that equity is crucial in the progression of MCED strategies. "An area that's really important to me is making sure that with this new technology there's adequate representation of all patients who could benefit … and that we are developing our understanding of MCEDs in a patient population that looks like the US population."
"In six months, they've really put together pretty close to the full reference," said Castle. "It's really been quite amazing, and we're very appreciative of it."
Looking to the future
As Vanguard and the individual efforts of companies like Grail and Exact Sciences move forward, the hope is that the crucial question highlighted by Bradbury will eventually be answered, but with testing already commercially available, the field is left scratching its head about what best practices may eventually emerge — or evolve in the interim.
Panelist Luis Diaz, a physician and researcher at Memorial Sloan Kettering, grappled with the question of MCEDs in the context of existing standard-of-care screening.
"It's a complicated question because there's some real enthusiasm for the eventual positive state of this. But this is a high-risk test. And the way I view these MCEDs is more like a drug where we have to understand the potential harm," he said.
"Just like a drug, we don't replace the standard of care on day one," he added. "We treat patients who are the sickest, the patients who are treatment refractory. And then systematically, once we understand the parameters, the positive or negatives of the drug, we move it to early guidelines."
Some standard-of-care tests couldn't or at least shouldn't be replaced even if MCEDs are proven out, Diaz added, citing tests that actually prevent cancer. "In colorectal cancer, two-thirds of the benefit of colonoscopy is finding adenomas and treating them, so we don't want to give that up. Believe me," he said.
Similarly, for cervical cancer the vast majority of benefit is in detecting and treating HPV. "These are important benefits that we cannot afford to lose, or we may undo some of the benefits or maybe all of the benefits of adding MCEDs into that," Diaz said.
Diaz called the implementation of MCED as competing with or even serving as an adjunct to standard-of-care screening premature and "unbelievably ambitious."
Early implementation in high-risk populations would be "easier to take at this point … and there are plenty of high-risk patients right now that could benefit," he said.
Bradbury agreed but cautioned that these groups "already live with a lot of fear."
"I think high-risk populations make sense, but I think they also may be potentially unique in how they experience a diagnostic follow up — the stress and anxiety that they may experience. So I just think it will be really important that we're watching all of those outcomes," she said.
Castle returned to the issue of hype, saying that getting studies rolling for high-risk patients might be a way to address these populations' eagerness and temper the risks posed by the commercial availability of MCEDs.
"The interesting thing about this is that every big health center has a cohort of these folks who are desperate for something, so let's do it in a controlled way where we can really understand the impact on them," Castle said.
Bradbury added that even with as many people put on trials as possible there are still going to be others who can't get access. "One way to fill the gap would be to make sure that everyone else is at least on a registry of outcomes so that, again, we're collecting data," she said.
Addressing the issue of the follow-up of positive MCED results, Diaz shared that MSK started a clinic about a year ago called the Clinic for Cancer Signals, which deals with not just MCEDs but also other procedures like full-body scans that patients can now access.
"There's a lot of crazy stuff out there, and patients will come to their primary care doctor and say, 'What do I do with this result?' The primary care doctor has no idea, so we opened up this clinic, and we opened the doors to see these patients in order to begin to just collect data and figure out how to start that diagnostic odyssey," he said.
As necessary as large efforts like Vanguard are, Diaz said that smaller programs and studies are going to be important as well. "I think we might be biting off a lot more than we can chew. I don't think we can afford to do all these studies. We have to start small because we're going to bankrupt many systems if we do it this way."
Castle added that company-sponsored studies will also begin to read out before Vanguard can be completed, hopefully beginning to answer crucial questions.
But he said the field should remain cautious. "We've been burned by things like PSA and other interventions. So maybe we're more conservative, the pendulum has swung a little in the other direction, for good and for worse, so I think we're going to have to wait now. But the truth is, we're not going to be able to stop this, which is why having a registry to collect that real-world data is going to be crucial."