Researchers at the Johns Hopkins Kimmel Cancer Center have validated an AI-powered blood test capable of detecting liver cancer across diverse international populations. The study confirms that the DNA Evaluation of Fragments for Early Interception (DELFI) platform successfully identifies cancer in patients regardless of the disease's underlying origin.
Investigators tested the platform on 377 individuals from Guatemala and Romania. These populations presented different risk factors, ranging from viral hepatitis and alcohol consumption to metabolic liver disease and aflatoxin exposure. Despite these biological variations, the blood test remained consistent in its ability to pinpoint liver cancer. When combined with traditional protein markers and patient clinical data, the tool demonstrated higher sensitivity than existing standard screening methods.
Beyond simple detection, the team used a method called MethID to trace the source of circulating DNA fragments. The analysis revealed that the test detects signals not just from tumor cells, but also from liver tissue, blood vessels, and the immune system. This allows for a deeper view of how the disease develops within the body. By capturing these molecular signatures, the technology can identify patterns specific to certain environmental exposures, such as aflatoxin, while maintaining a baseline accuracy for general cancer identification.
This work expands the potential for noninvasive liquid biopsies. The researchers noted that their fragmentome technology is already being applied to lung cancer screening in some health systems. Future plans involve prospective clinical validation to further refine how these tests combine with other biomarkers to catch liver cancer earlier when treatment options are most effective. This development represents a shift toward more accurate, adaptable screening tools that could be deployed across different healthcare environments globally.