Researchers from Peking University have developed a new liquid biopsy technology that can detect and trace disease signals using only a very small blood sample. The platform, called cf-EpiTracing, can analyse epigenetic signatures in as little as 50 microlitres of plasma, approximately a single drop of blood.
The research was led by Professor He Aibin of Peking University’s College of Future Technology and Professor Jing Hongmei from the Department of Hematology at Peking University Third Hospital.
Identifying the origin of disease signals
Liquid biopsies are increasingly used to detect disease markers in blood samples. However, existing methods often struggle to determine which tissue or organ the detected signals originate from, limiting their clinical usefulness.
The cf-EpiTracing platform addresses this challenge by analysing detailed epigenetic “fingerprints” present in fragments of cell-free chromatin circulating in the blood. By combining these epigenomic features with machine-learning algorithms, the system can identify the tissues involved in disease processes and distinguish between different disease subtypes. According to the researchers, the technology can also help predict patient outcomes more accurately than many existing clinical tests. The study was recently published in Nature.
Promising results in cancer detection
In tests focusing on colorectal cancer screening, cf-EpiTracing demonstrated high diagnostic performance. In training datasets the model achieved an accuracy of up to 97.6 percent, while validation using an independent cohort produced an accuracy of 92.2 percent.
The technology also revealed new insights into lymphoma biology. Researchers observed that patients with diffuse large B-cell lymphoma showed elevated signals from CD34-positive cells in their plasma, which may indicate bone marrow involvement and more aggressive disease.
The research team believes that integrating cf-EpiTracing with other cell-free diagnostic approaches, such as DNA methylation analysis, mutation detection and chromatin topology, could further improve diagnostic precision. Such a multi-omic approach could enable earlier disease detection and more accurate monitoring of disease progression and treatment response. If validated in larger patient populations, the platform may contribute to broader use of non-invasive diagnostics in oncology and other disease areas.
Blood test innovation
In recent years, we have seen further developments in the field of blood tests. For example, last year a new type of blood test was investigated for the early diagnosis of Alzheimer’s disease. Research by the Mayo Clinic shows that the test can detect Alzheimer’s in people with cognitive symptoms with approximately 95% accuracy. T
he test measures two biomarkers in the blood, Aβ42/40 and p-tau217, which are associated with amyloid plaques in the brain. p-tau217, proved to be highly reliable. The blood test offers a less invasive and cheaper alternative to existing diagnostic methods, such as PET scans and lumbar punctures, and, according to researchers, could contribute to earlier and more accessible detection of Alzheimer’s.
In 2024, researchers at the Peter MacCallum Cancer Centre developed an AI-based method that can detect malignant tumours using only a blood sample. The technology, called MisMatchFinder, analyses circulating tumour DNA (ctDNA) to identify mutation signatures specific DNA patterns caused by cancer.
Traditionally, this information requires genome sequencing of tissue obtained through biopsy, a process that is costly, time-consuming and sometimes difficult due to tumour location. By extracting the necessary genomic information directly from blood, MisMatchFinder could make advanced cancer diagnostics more accessible and enable faster, more personalised treatment decisions. The approach also allows physicians to monitor disease progression without repeated invasive biopsies.
