A blood test built with AI may be getting better at spotting liver cancer before it is too late.
The test, called DELFI, short for DNA Evaluation of Fragments for Early Interception, was developed at Johns Hopkins’ Kimmel Cancer Center and showed promise in finding liver cancer earlier in people already considered high risk, according to a Johns Hopkins University School of Medicine article shared by Medical Xpress.
Johns Hopkins said the results were published July 31 and showed the assay accurately identified liver cancer in two separate high-risk groups.
Researchers evaluated blood from 377 participants in Romania and Guatemala, including people with and without hepatocellular carcinoma, which Johns Hopkins described as the most common form of liver cancer.
The two groups had very different risk patterns. Romanian participants often had liver disease linked to viral hepatitis or alcohol use, while many Guatemalan participants had metabolic liver disease, obesity, diabetes and exposure to aflatoxin, a toxin linked to liver cancer.
Researchers said DELFI performed well in both groups.
They also reported that sensitivity for detecting early-stage and late-stage disease improved when the test was combined with alpha-fetoprotein, or AFP, along with basic information including age and sex, compared with blood testing alone.
Victor Velculescu, M.D., Ph.D., the Cancer Genetics and Epigenetics Professor and co-director of the cancer genetics and epigenetics program at the Johns Hopkins Kimmel Cancer Center, said, “This study demonstrates that the approach works with high performance across different patient populations while revealing the biological signals in the bloodstream that make this type of detection possible.”
Researchers said liver cancer is already among the major causes of cancer death globally, and the burden is rising as metabolic liver disease becomes more common.
They said earlier detection can expand treatment options, but routine screening, mainly ultrasound with AFP, still misses some cancers at an early stage.
The study also pointed to a possible reason the test performed well. Using a technique called MethID, researchers found that fragment patterns in blood reflected not only tumor DNA but also changes involving liver cells, blood vessels and immune cells responding to the disease.
Researchers said signals from multiple parts of that process could help the test perform better in practice, because liver cancers can vary across patients and regions.
Beyond this study, the Johns Hopkins team said it is working to move fragmentome screening toward broader use in disease detection, not only liver cancer.
Johns Hopkins said earlier research using the same general technology also showed it could detect liver fibrosis and cirrhosis, conditions that commonly come before liver cancer.
Researchers also said a related blood test for lung cancer screening, FirstLook Lung, is already being used in some health systems around the country through DELFI Diagnostics.
The team said the next steps are prospective clinical validation and more work on tests that combine fragment analysis with protein biomarkers and standard risk factors.
Existing screening recommendations remain in place, especially for people with known liver disease or other risk factors.
“As a result, these DNA fragments contain much more information than whether cancer is present,” said Zachariah Foda, M.D., Ph.D., a medicine faculty member at the Johns Hopkins University School of Medicine and a co-senior author of the study. “It tells us where these fragments originate and how they change during cancer development, allowing us to better understand the biology of the disease and improve our ability to detect it.”
Read more from The Cool Down.
