This week, cancer detection took a step forward when, for the first time, an FDA advisory panel reviewed a multi-cancer blood test for approval and voted to support GRAIL’s Galleri test for screening adults 50 and older 6-4. A single blood draw, read for signals from many cancers at once, moved closer to routine medicine. We should celebrate this as a beginning, with important opportunities ahead that extend well beyond screening.
The researchers and clinicians who brought multi-cancer screening to this point took on an extraordinary challenge: finding cancer in people who have no reason to suspect they have it. Their work has broadened what medicine can aspire to achieve, and it should raise our ambitions for patients who already know something is wrong.
Much of the panel’s discussion centered on the word “early,” which deserves attention beyond a test’s label.
For someone who feels well, early means finding cancer before symptoms appear. For someone who already has symptoms, it means reaching an answer sooner. Both matter, and the second deserves a larger place in the conversation about what blood-based detection could accomplish.
I think back to my years as a surgeon, reading patients’ charts backward through the biopsy, the referral, the scans, and the visits that began with a first symptom, a sign or finding. Those records often contained months of investigations that told a story of what wasn’t there, before the test that finally explained what was.
The next question for the field is how to help the people who are already worried, sitting across from a physician who must decide which of several possible explanations to pursue.
Diagnostic delay by design
Every year, roughly 40 million Americans visit a doctor with a symptom or finding that could be cancer, often beginning with fatigue, persistent pain, or unexplained weight loss. Most will have another explanation, but about 80% of cancers in the United States are diagnosed after symptoms appear, making these visits a central opportunity for earlier detection. On average, a patient can wait up to 5 months before a diagnosis, while mortality can increase by each month that passes.
We usually describe this as diagnostic delay, but some of that delay is built into the workup. Clinicians rightly investigate the more common explanations first, limiting unnecessary imaging, invasive procedures, anxiety, and expense. For a patient who does have cancer, that sensible approach can mean months of delay before malignancy becomes the leading concern.
Looking backward through a chart makes the pattern easier to recognize than it was at the first appointment. The physician making those initial decisions has to judge when a familiar symptom warrants urgent investigation, often without an objective measure of cancer risk to help distinguish that patient from the many others with similar complaints.
Molecular testing in those initial evaluations could provide useful information before that sequence has run its course, helping clinicians recognize who needs a more urgent cancer evaluation. These patients are already seeking care, and we should bring the same scientific ambition to shorten their search for an answer that we bring to finding cancer before symptoms begin.
It would be tempting to reach for a test similar to the one the FDA panel discussed, but evaluating a patient with symptoms presents a different clinical problem from screening someone who feels well. Among asymptomatic individuals, the low prevalence of cancer makes false positives a particular concern, requiring screening tests to operate at very high specificity. The thresholds needed to achieve that can also make cancers harder to detect when the tumor signal in the blood is faint.
In patients whose symptoms, signs or findings (where bedside clinical, radiological, biochemical or pathological) already raise suspicion of cancer, the physician is deciding how urgently to investigate and where to look. The higher likelihood of cancer changes how a molecular result should be interpreted and how the consequences of missed cancers and false positives should be weighed. A test developed for this population could help guide those decisions, with its performance established in the patients it is meant to serve.
Considering cancer first
At Harbinger, we are developing RESOLVE around this second meaning of early, bringing molecular information into the evaluation when a patient first presents with a suspicion of cancer. The aim is to support the decision and direction of the workup for a physician who suspects cancer and still deciding what should happen next.
A high-risk result is intended to support rapid imaging or specialist evaluation, with an indication of where the signal is most likely coming from, while a low-risk result is intended to support a proportionate workup with continued clinical vigilance.
We include an intermediate category, recognizing that cancer biology may not provide a decisive answer on the day a patient happens to have blood drawn or may not yet have revealed a molecular signal if very early on. In appropriate patients, a repeat draw a few weeks later is intended to help assess whether the signal is stable or rising, adding information as the clinical picture develops.
Following this signal over time could give physicians another way to assess an unresolved concern, with the patient’s symptoms, examination, and other findings continuing to determine the urgency of investigation.
Changing the months that follow
Through this approach, information available at the first appointment could change the months that follow. A patient who reports unexplained weight loss and persistent fatigue might have a blood test that helps their physician recognize the need for urgent investigation, while another patient might continue through a more measured evaluation with a clear plan for follow-up.
Over the next decade, this could change what patients expect when they seek help for symptoms they cannot explain. Physicians could have a clearer basis for arranging investigations and referrals from the outset, while patients would know what is being pursued, why it matters, and when they should expect the next decision.
That future would give both meanings of early a place in everyday care. Screening would continue to look for cancer before symptoms appear, while tests developed for symptomatic patients, or those with clinical findings and signs would aim to shorten the path from concern to diagnosis.
It could also expand what physicians learn from a blood test, with repeat measurements in appropriate patients adding information about how a signal changes over time, establishing when those changes are clinically useful is an important part of the work ahead.
The larger opportunity is to organize care around reaching an answer. A result that identifies increased risk should lead to timely imaging or specialist review, with responsibility for follow-up clearly assigned. Better information at the first appointment will have limited value if the patient then spends months waiting for the investigation it indicates.
I expect Cancer Suspicion Clinics to become a larger part of healthcare in the U.S. and worldwide, building on services already established at a handful of centers that have evolved from rapid cancer diagnosis and early cancer detection programs. Their wider adoption could give clinicians with any ‘suspicion of cancer’, whether in primary, secondary or tertiary care, a coordinated route for investigating symptoms, signs or unexplained clinical or radiological findings that raise concern without pointing clearly to a particular cancer. Within these services, multi-cancer blood tests such as RESOLVE, specifically developed for this population, could help clinicians decide the pathway and the urgency with which to investigate. The starting point would remain the clinician’s suspicion, even when symptoms, signs or findings are vague, with liquid biopsy information supporting clinical judgement. Ultimately, this approach aims to detect cancer more efficiently through established clinical pathways, while also reducing unnecessary tests and referrals and shortening the wait for diagnosis.
What progress will require
New approaches should earn their place through evidence gathered where they are intended to be used. Studies need to establish whether testing helps physicians choose appropriate investigations, avoids unnecessary procedures, and shortens the time to diagnosis, then determine where those changes improve patient outcomes.
Test of this type will need to directly support clinicians’ decisions and judgments in line with their suspicion and understanding and coordinate with their well-established pathways of care. The economic case must account for that full course of care, including repeated appointments, additional procedures, delayed diagnoses, and treatment for advanced disease alongside the price of testing. Earlier information should be judged by what it enables physicians to do and what it changes for the person waiting for an answer.
The scrutiny accompanying this week’s advisory review matters to everyone developing blood-based cancer tests. It pushes the field to be precise about the patients a test is intended to help, the evidence supporting its use, and the care that follows its result. Those expectations should strengthen both screening and evaluation of patients with symptoms.
Back to the chart
Reading patients’ charts backward meant tracing a diagnosis through the investigations and visits that preceded it. The opportunity now is to bring better information to the beginning of that history, when physicians are still deciding what to investigate and how urgently to act. The decisions made at that first visit can shape the months that follow, giving us a compelling reason to put more useful information in the physician’s hands earlier.
Screening gives us a chance to find cancer before symptoms appear, while RESOLVE is being developed to help physicians act sooner once clinical symptoms, signs or findings bring a patient into care.
Together, these approaches could make both meanings of early more achievable, helping more patients reach an earlier answer while they still have time and choices ahead of them.