Detecting liver cancer early can save lives, but finding the disease before symptoms appear has always been difficult.
Researchers at the Johns Hopkins Kimmel Cancer Center have now developed stronger evidence that an artificial intelligence-powered blood test could help doctors detect liver cancer earlier and more accurately.
The study was published in Cell Press Blue.
The new work builds on an earlier technology called DELFI, which studies tiny pieces of DNA that naturally circulate in the bloodstream.
These DNA fragments are released when cells die, and the pattern of the fragments can reveal whether cancer or other diseases may be present. Unlike many traditional blood tests, DELFI looks across the whole genome rather than searching for a single mutation.
Liver cancer is one of the deadliest cancers worldwide because it is often diagnosed too late for the best treatments.
Many cases develop after years of liver damage caused by hepatitis viruses, heavy alcohol use, obesity, diabetes, or fatty liver disease. Current screening usually relies on ultrasound scans and the blood protein alpha-fetoprotein, but these methods can miss early cancers.
In this study, scientists tested the DELFI approach in 377 people from Guatemala and Romania. These countries were chosen because liver cancer develops for different reasons in each population.
Romania had more cases linked to hepatitis and alcohol, while Guatemala had many people with metabolic liver disease and exposure to aflatoxin, a natural toxin that can contaminate crops.
Even though the causes were different, the AI-powered blood test performed well in both groups. When researchers combined the DELFI results with AFP, age and sex, the test detected both early-stage and advanced liver cancer more accurately than current blood testing alone.
The team also wanted to understand why the test worked so well. They developed another method called MethID to identify where the DNA fragments came from. They found that the blood test collected information not only from tumor cells but also from healthy liver tissue, blood vessels and immune cells responding to the cancer.
This broader biological picture gives doctors more information than a simple yes-or-no cancer result. It may also help scientists better understand how liver cancer develops and how the body’s normal tissues react as the disease grows.
The researchers also found molecular patterns that reflected local environmental risks. For example, DNA changes linked with aflatoxin exposure were seen mainly in participants from Guatemala. Despite these regional differences, the overall AI system still identified liver cancer accurately, suggesting it can work in many different populations.
The scientists believe this technology could eventually become part of a larger platform capable of detecting several diseases from a single blood sample. Related research has already shown that similar fragment analysis can identify liver fibrosis, cirrhosis and even lung cancer through other DELFI-based tests.
Review and analysis: This study is important because it validated the technology in two very different populations rather than only one group, making the findings more convincing.
However, larger prospective clinical trials are still needed before the test becomes routine medical care. If future studies confirm these results, AI-powered fragment DNA testing could become a valuable tool for earlier liver cancer detection and better survival.
If you care about liver health, please read studies that refined fiber is link to liver cancer, and the best and worst foods for liver health.
For more health information, please see recent studies about how to boost your liver naturally, and simple ways to detox your liver.
