Nanoparticle-based blood test outperforms invasive biopsy for pancreatic cancer detection

A dielectrophoresis (DEP)-driven liquid biopsy distinguished pancreatic cancer from benign and precancerous pancreatic disease in a blinded cohort, exceeding the performance of the current standard diagnostic test, which is the invasive endoscopic ultrasound-guided fine needle aspiration (EUS/FNA) tissue biopsy, according to research published in the journal Small last month.

Stuart Ibsen, Ph.D., from the Cancer Early Detection Advanced Research Center, Knight Cancer Institute, Oregon Health and Science University, Portland, Oregon, working with collaborators at the University of California San Diego’s Moores Cancer Center and the biotechnology firm RyboDyn, applied DEP, a label-free electrokinetic technique, to simultaneously recover cancer-derived extracellular vesicles and cell-free DNA (cf-DNA) nanoparticles from a single plasma sample enabling quantification of associated biomarkers.

In a blinded cohort of 36 patients, the combined two-biomarker panel produced a sensitivity of 0.92, a specificity of 0.83, and an area under the curve (AUC) of 0.93. A test with an AUC of 0.93 would point to the right person about 93 times out of 100. Among patients older than 50, the age group accounting for more than 90% of pancreatic ductal adenocarcinoma (PDAC) diagnoses, the AUC rose to 0.97. By comparison, the gold-standard endoscopic ultrasound-guided fine needle aspiration (EUS/FNA) tissue biopsy carries a reported AUC of 0.79.

Pancreatic ductal adenocarcinoma, the most common and aggressive form of pancreatic cancer, remains one of the deadliest malignancies in the United States, in part because the disease is rarely caught early. The only widely used blood marker, CA 19-9, is recommended for monitoring known disease rather than for screening. EUS/FNA, the current diagnostic procedure, is costly, requires a trained interventional endoscopist, and carries a roughly 1 in 100 risk of post-procedural acute pancreatitis, an event with a reported 10% mortality rate. Between 60% and 76% of patients referred for EUS/FNA ultimately do not have pancreatic cancer.

Those statistics establish the managed care relevance of this research. The DEP-based test relies on a routine blood draw, requires no plasma dilution and completes nanoparticle capture in 15 minutes, the authors reported. Both biomarker collection and quantification occur on the same microfluidic chip, eliminating fluid transfers that can degrade signal in low-volume samples, a feature the authors said is particularly useful for early-stage cancers, where biomarker abundance is limited.

The panel also separated pancreatic ductal adenocarcinoma from low-grade intraductal papillary mucinous neoplasm (IPMN), a precancerous lesion that often shares molecular features with invasive disease. The authors said the work is among the first to demonstrate that a combined DNA-and-protein threshold can distinguish pancreatic ductal adenocarcinoma from these precursor lesions. The lone IPMN sample classified as cancer was a high-grade lesion that would typically be surgically removed.

Glypican-1 is a protein that sits on the outside of cells. It's anchored to the cell's outer membrane and helps the cell communicate with its surroundings, things like grabbing onto growth signals and helping the cell move. Healthy adult cells have very low amounts of it. But certain cancer cells, especially pancreatic cancer cells, make a lot of it. That difference is what makes it interesting as a cancer marker.

Glypican-1 alone produced an AUC of 0.77, and cf-DNA alone produced an AUC of 0.67. Both markers were needed to reach the panel’s 0.93 AUC, the authors said, in part because the two were uncorrelated, capturing independent biological information. The test also flagged a patient initially classified as non-cancer who was diagnosed with stage 2 hepatocellular carcinoma six months later, suggesting a potential broader signal for occult malignancy.

The authors positioned the platform as a candidate triage tool for individuals at elevated pancreatic ductal adenocarcinoma risk, including those with new-onset diabetes, familial history or germline mutations such as BRCA2. Larger cohort validation and integration of additional biomarkers are planned.