UCLA scientists just published a method for detecting more than 50 cancer types from a single inexpensive blood draw. Here is how it actually works.
Somewhere in your bloodstream right now, cells from your organs are dying -- liver cells, lung cells, cells from your gut lining -- and as they die, they release tiny fragments of their DNA. This cellular shedding happens constantly, in everyone, for an entire lifetime. For most of medical history, that DNA was invisible noise. In April 2026, a team at UCLA published a paper in the Proceedings of the National Academy of Sciences showing how to read it.
Every cell that dies in your body leaves behind a chemical signature. When liver cells break down, when lung tissue sheds its lining, when the intestinal epithelium renews itself -- all of these processes release fragments of DNA into the bloodstream. Scientists call this cell-free DNA, or cfDNA. It is everywhere in the circulatory system, a constant stream of molecular debris that carries information about what is happening in every organ.
The problem has always been signal-to-noise. When you take a blood sample and search for cancer DNA, roughly 90% of what you find comes from white blood cells -- healthy cells that dominate the background and obscure the actual disease signal. If you could somehow remove that noise, you would have a clear view of what is happening in the tissues where cancer actually develops: the liver, lungs, pancreas, ovaries, stomach, and dozens of other organs.
This is where methylation changes everything. Methylation is a process as old as life itself. Cells attach small chemical tags called methyl groups to specific sites along the DNA sequence. These tags do not alter the genetic code itself -- you still have the same genes -- but they regulate which genes are active and which are silent. Think of it as a dimmer switch for genes rather than an on-off switch. A liver cell and a lung cell carry identical DNA sequences, but their methylation patterns are completely different, because each cell type needs different genes turned on and off.
Cancer cells have their own characteristic methylation patterns. When normal cells transform into cancer cells, their methylation profiles change in recognizable ways. The UCLA team, led by researchers including those at the university's David Geffen School of Medicine, engineered a method called MethylScan that exploits this difference. The method uses specialized enzymes to cut away the unmethylated DNA -- the white blood cell noise -- and preserves the methylated signal from solid organs. The remaining fragments are sequenced across the entire genome and compared against a reference library of tissue-specific methylation patterns. The result is a molecular portrait of organ health, constructed from a single inexpensive blood draw.
What makes MethylScan distinctive is that it was explicitly engineered to be affordable. Rather than sequencing every DNA fragment in a blood sample -- which is expensive and generates enormous amounts of noise -- the enrichment step concentrates the signal before sequencing happens. This allows the test to be performed for a fraction of the cost of competing approaches. The team estimates that as manufacturing scales, the test could be performed for under $100 per patient. For comparison, the most common cancer screening tools today -- colonoscopy, mammography, CT scanning -- cost hundreds or thousands of dollars per test, are invasive or involve radiation, and screen for only one cancer type at a time.
The UCLA team tested MethylScan on blood samples from 1,061 people. The group included patients diagnosed with liver cancer, lung cancer, ovarian cancer, and stomach cancer. It included people with liver disease -- hepatitis B, hepatitis C, alcohol-related liver disease, metabolic-associated fatty liver disease. It included people with benign lung nodules (abnormalities that turned out to be non-cancerous) and healthy controls with no disease. This is the gold standard for validation: testing on people who represent the full spectrum of disease states you actually want to detect.
The results are the most important numbers in early cancer detection this year. At a specificity of 98% -- meaning the test correctly identified 98% of healthy people as healthy -- MethylScan detected approximately 63% of cancers across all stages combined. For early-stage cancers, the ones where treatment is most effective and survival rates are highest, the detection rate was approximately 55%. That is not perfect, but it represents a massive advance over the current standard of care, which is often no screening at all.
Liver cancer tells a more striking story. In patients at high risk -- those with cirrhosis or hepatitis B infection -- the test detected approximately 80% of cases. This matters profoundly. Liver cancer in these populations is lethal, often because it is discovered only after it has become advanced and spread. A reliable way to catch it early could save tens of thousands of lives annually.
But the paper revealed something else that no existing blood test can do. MethylScan did not just detect cancer. It distinguished between different types of liver disease, correctly classifying about 85% of patients. This sounds technical until you understand what it means clinically. A patient with chronic hepatitis B needs antiviral therapy. A patient with metabolic fatty liver disease needs dietary intervention and weight management. A patient with cirrhosis needs to be monitored closely for liver failure and liver cancer. A patient with early-stage hepatocellular carcinoma needs surgical resection or transplantation. These are completely different treatment pathways. Today, distinguishing between them requires a liver biopsy -- an invasive procedure where a needle is inserted into the liver and tissue is extracted. With MethylScan, the same blood draw that screens for cancer also tells you what disease you actually have.
In cancer screening, two numbers matter: sensitivity (the percentage of actual cancers you catch) and specificity (the percentage of healthy people you correctly identify as healthy). At first glance, it seems like you want both to be as high as possible. The tradeoff is far more subtle.
Imagine a screening test with 99% sensitivity but only 50% specificity. That means it catches almost all the cancers, but it also generates an enormous number of false positives -- telling healthy people they have cancer. These people undergo biopsies, imaging, surgery, and suffer anxiety, all for a disease they do not have. Cascade effects ripple through the medical system. The test becomes counterproductive.
By contrast, a test with 63% sensitivity and 98% specificity catches a majority of cancers while almost never alarming people who are healthy. The false positive rate is low enough that when the test says you have cancer risk, it is probably accurate. This creates confidence. Patients follow up with confirmatory testing (which remains the gold standard). Clinicians have actionable information. For a population-level screening tool, 98% specificity is the critical threshold that makes the test usable in the real world.
MethylScan is not the first multi-cancer early detection test to be developed, nor is it the only one now in clinical use. Grail, a subsidiary of Illumina, developed the Galleri test, which detects signals from more than 50 cancer types and is currently available for self-pay purchase in the United States. Grail completed enrollment in the PATHFINDER 2 trial, which enrolled nearly 36,000 participants, and results are being analyzed. Exact Sciences, the company behind the Cologuard colorectal cancer test, has developed Cancerguard, which also analyzes DNA and protein markers for 50+ cancer types and is moving toward clinical launch.
What distinguishes MethylScan is not that it detects more cancer types -- it does not -- but that it was designed from the ground up to be low-cost and to detect non-cancer disease simultaneously. The team did not chase the highest sensitivity. Instead, they optimized for clinical utility and affordability. The enrichment approach that makes MethylScan inexpensive also makes it simpler to scale manufacturing. This matters enormously in global health. If a blood test costs $1,000, it will never be deployed in low-income settings. If it costs under $100, the calculus changes. You can imagine a future in which every person, regardless of income, gets annual screening for dozens of conditions from a single inexpensive blood draw.
The broader technology trajectory is accelerating. A decade ago, DNA sequencing a human genome cost over $1,000 and took weeks. Today, sequencing costs less than a hundred dollars and takes days. Machine learning algorithms have become far more sophisticated at finding patterns in methylation data. Our library of tissue-specific and cancer-specific methylation patterns grows larger and more precise each year. These three currents -- cheap sequencing, better algorithms, and deeper knowledge of the methylation landscape -- are converging into a wave of innovation in liquid biopsy. The question is no longer whether such tests can work. The question is how quickly they can be validated, deployed, and made accessible.
The most consequential trial now running is the Grail-NHS partnership in the United Kingdom. This study is enrolling over 140,000 participants and will report results in 2026. This is not a small academic study. This is a population-level health system trial involving one of the world's largest medical databases and hundreds of thousands of people over multiple years. The results will answer the question that matters most: in a real-world setting, does screening for cancer with a blood test actually reduce cancer mortality? That answer will determine whether these tests become standard medical practice globally.
The far horizon of this research is not a better cancer test. It is a routine annual blood panel that monitors the health status of every major organ simultaneously. One draw. One test. Results that tell you the state of your liver, lungs, pancreas, ovaries, stomach, and beyond. Results that detect not only cancer, but also early signs of liver disease, kidney failure, and potentially dozens of other conditions, all from the same sample.
This is theoretically possible because the methylation signature of illness is written into cfDNA. Cancer changes methylation patterns. So does liver inflammation. So does kidney disease. So does the early stages of heart failure, where stressed cardiac cells release cfDNA with altered methylation profiles into the bloodstream. The same technology that reads the cancer signal can, in principle, read the signals of any condition that changes a cell's gene expression program. Cardiac stress, kidney disease, neurodegeneration, autoimmune activation -- all leave methylation signatures in circulating DNA that are, at least theoretically, readable.
But theory and clinical reality are different things. MethylScan's results are promising, but they come from a single study of 1,061 patients. This is not a small study, but it is not a population study either. Confirming performance at true scale -- and understanding how the test performs across different ethnicities, age groups, socioeconomic backgrounds, and health profiles -- requires trials of extraordinary magnitude. The Grail-NHS study is one such trial. The UCLA team is already planning larger validation studies of their own. These will take years. This is not impatience with the pace of science; it is appropriate caution. A test that performs well in a research study might behave differently in clinical practice, in different populations, with different storage conditions, with different versions of the assay.
What is clear is that the fundamental technology works. We can read the health signal in blood. We can distinguish cancer from non-cancer disease. We can do it affordably. The question now is whether the clinical and regulatory infrastructure can keep pace with the technology. History suggests it will not move as quickly as the underlying science might suggest, but it will move.
The UCLA paper concludes with a statement that deserves careful reading. A blood test that simultaneously screens for dozens of conditions -- cancers, liver diseases, kidney dysfunction, and conditions yet to be discovered -- could reduce the need for invasive diagnostic procedures across an enormous range of diseases. No more colonoscopies for everyone. No more liver biopsies for diagnosis. No more expensive CT scans looking for nodules that might be cancer. These procedures could be reserved for patients whose blood test indicates genuine risk. The cascade of downstream benefits -- reduced anxiety, reduced complications from unnecessary procedures, reduced cost, increased equity in access to screening -- extends far beyond the test itself.
We are in the early phase of a transformation in how we think about health surveillance. For most of human history, our bodies were largely opaque to us. Disease was discovered after symptoms appeared. Now we have the ability to read the continuous health broadcast from our own cells and act before disease becomes symptomatic. This is not science fiction. It is published in the Proceedings of the National Academy of Sciences. It is being tested in clinical trials involving hundreds of thousands of people. The technology is real. The question now is how quickly, and at what cost, we integrate it into the practice of medicine.
© 2026 Lisa Pedrosa · lisapedrosa.com
All articles cited to primary institutional or peer-reviewed sources
Buy me a coffee