A genetic twist hidden in the DNA of roughly one in ten pancreatic cancer patients could be masking their disease as less dangerous than it actually is — and a new two-threshold approach to measuring a standard biomarker might finally catch what traditional tests miss.
Pancreatic ductal adenocarcinoma remains one of the world's most lethal cancers, with approximately 80% of cases diagnosed only after they've advanced significantly. The five-year survival rate stands at just 13.7%, making early and accurate risk assessment crucial. For decades, physicians have relied on a single measurement — serum carbohydrate antigen 19-9, or CA19-9 — to gauge how aggressive a patient's disease is and guide treatment decisions. Higher CA19-9 levels typically signal more advanced cancer and a grimmer outlook. But this approach has a critical blind spot.
About 10% of pancreatic cancer patients produce little to no CA19-9 despite having advanced disease. These "nonproducer" patients carry a genetic variation in the FUT3 gene that prevents them from producing the Lewis antigen needed to generate this biomarker. When their CA19-9 levels appear normal, clinicians face an impossible choice: they cannot distinguish between someone with genuinely early-stage cancer and someone with dangerous advanced disease hidden behind normal lab results.
Now, researchers from Taiwan's National Cheng Kung University Hospital and Kaohsiung Medical University Hospital have published a solution in Clinical Cancer Research. Using whole-exome sequencing to determine the FUT2 and FUT3 genotypes of 615 pancreatic cancer patients, they developed a dual-threshold model that could change how doctors interpret CA19-9 levels.
The team identified a new cutoff of 7 units/mL or less as a marker for Lewis antigen-negative patients — those genetic nonproducers hiding behind falsely reassuring test results. When they validated this threshold in half their study population, it identified Lewis antigen-negative patients with 87.9% accuracy. More tellingly, the survival data proved the clinical significance: patients with CA19-9 levels at or below 7 units/mL experienced outcomes comparable to those with levels greater than 200 units/mL, suggesting that very low CA19-9 may actually signal high-risk disease in certain patients.
The findings reveal a stark paradox in the conventional approach. Patients with intermediate CA19-9 levels — between 7 and 37 units/mL, or between 37 and 200 units/mL — had longer median overall survival times of 23.2 and 22 months respectively. But those at the extremes, the very low and very high producers, had median survival times of 13.5 and 12.8 months. As lead investigator Yung-Yeh Su, MD, Ph.D., noted: "The conventional normal CA19-9 range of less than 37 does not distinguish between true low tumor burden and Lewis-negative status."
This dual-threshold model offers a tangible path forward. By combining genetic information with biomarker measurement, physicians can now avoid the dangerous misinterpretation that has left many nonproducer patients without appropriate intensity of treatment. For patients whose genes prevent them from producing CA19-9, this research offers something more than a new number — it offers recognition, clarity, and a chance at treatment matched to their actual risk.