Detect these cancers years earlier, at their earliest stages
For decades, HPV-associated head and neck cancers have grown in silence, announcing themselves only when treatment must be aggressive. Researchers at Mass General Brigham have now developed a blood test called HPV-DeepSeek that detects viral DNA shed by tumors years before symptoms emerge, achieving 99% accuracy in clinical study. The test represents a rare convergence of biological insight and technological precision — the possibility that a simple blood draw might one day intercept a cancer nearly a decade before it declares itself. If validated through ongoing trials, it could shift the entire logic of how these increasingly common cancers are found and treated.
- HPV-associated oropharyngeal cancers are rising faster than any other head and neck malignancy, yet unlike cervical cancer, they have no established early screening program — patients typically arrive at diagnosis already in need of aggressive treatment.
- HPV-DeepSeek sequences the entire viral genome rather than the one or two fragments targeted by existing liquid biopsies, dramatically widening the detection net and achieving sensitivity and specificity that outperform every current diagnostic method.
- In an unpublished study, the test detected cancer in 79% of people who would later be diagnosed — with the earliest positive signal appearing nearly eight years before clinical diagnosis — suggesting tumors shed traceable DNA long before they become symptomatic.
- Critical questions remain unresolved: how frequently to screen, what to do with a positive result in an asymptomatic person, and how to avoid the harms of overtreatment for cancers that might never progress.
- Parallel trials are now testing whether related assays can detect residual cancer after surgery within days, potentially guiding real-time decisions about whether patients need additional radiation therapy.
Researchers at Mass General Brigham have developed a blood test — HPV-DeepSeek — capable of identifying HPV-associated head and neck cancers with 99% accuracy, outperforming every existing diagnostic approach in a study of 152 cancer patients and 152 healthy controls. The test works by detecting fragments of the virus's genetic material circulating in the bloodstream, along with nine other biological markers. What distinguishes it from current liquid biopsies is its scope: rather than targeting one or two pieces of the viral genome, it sequences the entire HPV genome, casting a far wider and more reliable net.
The urgency behind this work is real. Human papillomavirus drives roughly 70% of oropharyngeal cancers in the United States, and these cancers are growing in frequency faster than any other head and neck malignancy. Unlike cervical cancer — also HPV-driven — there is no established screening program for oropharyngeal disease. Most patients don't seek care until symptoms appear, by which point treatment is typically intensive and carries significant side effects. Daniel Faden, the surgical oncologist leading the research at Mass Eye and Ear, set out to create a minimally invasive way to catch these cancers before they announce themselves.
The most striking evidence came from a separate, still-unpublished study involving 28 people who would later develop oropharyngeal cancer. HPV-DeepSeek detected cancer in 79% of them — while all healthy controls tested negative — and the earliest positive result appeared nearly eight years before diagnosis. These cancers develop over roughly 15 years, gradually shedding viral DNA as they grow, and this study suggests that window of detectable signal is far longer than previously understood.
The implications for treatment are significant. Catching cancer at its earliest stages could allow less intensive therapy, reducing side effects and improving quality of life. But important questions remain: how often should screening occur, what should be done when someone tests positive without symptoms, and how do clinicians weigh early detection against the risks of anxiety and overtreatment? The research team is now running clinical trials to address these questions, including work on detecting residual disease after surgery — a companion assay called MAESTRO has already shown it can identify microscopic remaining cancer within days of an operation, with measurable consequences for survival outcomes.
Liquid biopsy is moving steadily from experimental promise toward clinical practice, and whole-genome approaches like HPV-DeepSeek are a meaningful part of that shift. The path from research finding to standard care remains long, but for the first time there is a clear, early signal — one that could change everything about how these cancers are found and fought.
Researchers at Mass General Brigham have developed a blood test that can identify HPV-associated head and neck cancers with 99% accuracy—a breakthrough that could fundamentally change how these tumors are caught and treated. The test, called HPV-DeepSeek, detects fragments of the virus's genetic material that have escaped from tumors into the bloodstream, along with nine other biological markers. In a study of 152 cancer patients and 152 healthy controls, it outperformed every existing diagnostic method, including other commercial liquid biopsies currently in use.
The stakes of this work are substantial. Human papillomavirus causes roughly 70% of oropharyngeal cancers in the United States, and these cancers are increasing in frequency faster than any other head and neck malignancy. Yet unlike cervical cancer, which is also HPV-driven and has well-established screening programs, there is no early detection tool for oropharyngeal cancer. Most patients don't see a doctor until symptoms appear—a delay that often means more aggressive treatment, with all its attendant side effects. Daniel Faden, a surgical oncologist and principal investigator at Mass Eye and Ear, explained that the goal was to create a minimally invasive way to catch these cancers before they announce themselves.
What makes HPV-DeepSeek different from existing blood tests is its scope. Current liquid biopsy approaches hunt for one or two pieces of the viral genome. HPV-DeepSeek sequences the entire HPV genome, casting a much wider net. This whole-genome approach is what allows it to achieve such high sensitivity and specificity—the test correctly identifies cancer when it's present and correctly rules it out when it's absent.
But the most striking finding came from a separate, still-unpublished study. Researchers tested HPV-DeepSeek in 28 people who would later develop oropharyngeal cancer and 28 healthy controls. The test detected cancer in 79% of those who eventually got sick, while all the controls tested negative. The earliest positive result came nearly eight years before the person was diagnosed with cancer. This suggests that HPV-associated cancers release detectable DNA into the blood years before tumors become clinically apparent—a window of opportunity that has never been clearly visible before.
The biological timeline matters here. These cancers typically develop over about 15 years, gradually shedding viral DNA as they grow. If that DNA can be caught in the blood years before a tumor becomes symptomatic, treatment could be radically different. Faden noted that catching cancer at its earliest stages could allow doctors to use less intensive therapy, reducing side effects and improving quality of life. But significant questions remain unanswered: How often should people be screened? What should happen to someone who tests positive but has no symptoms? How do you balance the benefit of early detection against the anxiety and potential overtreatment of finding cancers that might have remained dormant?
The research team is now running multiple clinical trials to address these questions. One line of work focuses on detecting microscopic disease remaining after surgery—the kind of residual cancer that might indicate whether additional radiation therapy is needed. In a related study published in the same journal, Faden's group tested a different assay called MAESTRO on non-HPV head and neck cancers and found it could accurately identify residual disease within days of surgery. Patients with detectable residual disease had significantly worse survival and recurrence outcomes, suggesting that these sensitive blood tests could help guide treatment decisions in real time.
The broader implication is that liquid biopsy—reading cancer's genetic signature from a simple blood draw—is moving from experimental curiosity toward clinical reality. Whole-genome sequencing approaches like HPV-DeepSeek and MAESTRO allow physicians to search for hundreds or thousands of genetic signals rather than just a handful, drastically improving the odds of catching what's actually there. For patients, this means more accurate diagnoses and the possibility of truly personalized treatment. The work is still in early stages, and the path from research finding to standard clinical practice is long. But for the first time, there is a clear signal that HPV cancers can be detected in blood years before they cause symptoms—and that detection might change everything about how they're treated.
Notable Quotes
If we can detect these cancers years earlier, at their earliest stages, it could drastically change how we treat patients; for example, being able to use less treatment leading to fewer side effects.— Dr. Daniel Faden, Mass Eye and Ear
Whole-genome sequencing liquid biopsy approaches are enormously powerful, allowing physicians to look for many hundreds or thousands of needles in haystacks as opposed to just a few, drastically increasing sensitivity.— Dr. Daniel Faden
The Hearth Conversation Another angle on the story
Why does HPV-associated head and neck cancer need a new detection method? Don't doctors already know how to find these tumors?
They do, but only after the patient shows up with symptoms—a sore throat that won't heal, difficulty swallowing, a lump in the neck. By then, the cancer has already been growing for years. There's no screening program like there is for cervical cancer, even though the same virus causes both.
And that delay matters because?
Because it means more aggressive treatment. Surgery, radiation, chemotherapy—all of it carries real side effects. If you catch the cancer when it's tiny, still just releasing DNA into the blood, you might be able to treat it with much less intensity.
The test detects viral DNA in the blood. How does that work?
As the tumor grows, it sheds fragments of the HPV genome into the bloodstream. HPV-DeepSeek sequences the entire viral genome, not just one or two pieces like other tests do. It's like the difference between looking for a specific person in a crowd versus scanning everyone's face.
The study found it could detect cancer eight years before diagnosis. That seems almost too good to be true.
It does. That's why they're running more trials now. The question isn't whether the test works—it clearly does. The question is what you do with that information. Do you treat someone who has no symptoms? How do you avoid unnecessary anxiety or overtreatment?
So this isn't the end of the story.
It's the beginning. The test works. Now comes the harder part: figuring out how to use it wisely.