Adding a Lower Cutoff Value for CA19-9 May Identify Additional High-risk Cases of Pancreatic Cancer

CA19-9 is a biomarker whose levels often correlate with pancreatic cancer stage and prognosis

PHILADELPHIA – A dual-threshold model for measuring the pancreatic tumor marker serum carbohydrate antigen 19-9 (CA19-9) identified patients with pancreatic cancer who had high-risk disease despite having low CA19-9 levels because of a genetic variation that impairs their ability to produce this biomarker, according to a study published in Clinical Cancer Research, a journal of the American Association for Cancer Research (AACR).

Pancreatic ductal adenocarcinoma is a highly lethal cancer in which approximately 80% of cases are identified at an advanced stage, contributing to a five-year survival rate of 13.7%. Physicians typically use CA19-9 levels to assess risk, help guide treatment decisions, and monitor treatment response in patients with pancreatic cancer.

Levels of CA19-9 (which is also known as sialyl Lewis antigen A) correlate with pancreatic cancer stage and prognosis in that higher levels indicate more advanced disease and a worse prognosis. Currently, a level of CA19-9 less than 37 units/mL is considered normal or, in the case of a known diagnosis of pancreatic cancer, prognostic for standard-risk disease.

However, approximately 10% of patients with pancreatic cancer do not have elevated CA19-9 levels, even in the presence of advanced disease. The Lewis antigen-negative status of these CA19-9 “nonproducer” patients comes from genetic polymorphisms in the FUT3 genes that impair the fucosyltransferase (FUT) activity required to produce CA19-9.

“When CA19-9 levels are in the normal range, clinicians without access to their patients’ genotyping cannot distinguish Lewis antigen-negative patients with advanced disease from those with genuinely low tumor burden,” said Yung-Yeh Su, MD, PhD, assistant investigator and attending physician at the National Health Research Institutes, National Cheng Kung University Hospital, and Kaohsiung Medical University Hospital in Taiwan, and senior author of the study.

To establish a CA19-9 cutoff that would capture Lewis antigen-negative patients and better predict their prognosis, Su and his coauthors examined the relationships among genotypically determined Lewis antigen status, CA19-9 levels, and prognosis. They used whole-exome sequencing to determine the FUT2/FUT3 genotypes of 615 patients with pancreatic cancer at National Cheng Kung University Hospital and Kaohsiung Medical University Hospital in Taiwan. Participants were stratified by FUT genotype into four groups ranging from FUT3-null (CA19-9 nonproducers/Lewis antigen-negative) to FUT-high.

The authors designated half the study population as a training cohort to determine whether combining FUT2/FUT3 genotypes with CA19-9 levels could help define new CA19-9 thresholds to identify nonproducer (Lewis antigen-negative) patients, thereby reducing misinterpretation and improving prognostic assessment.

This process identified a CA19-9 level of 7 units/mL or less as a new cutoff for identifying patients with pancreatic cancer who have the Lewis antigen-negative genotype. Validation of the cutoff in the rest of the study participants showed that this CA19-9 cutoff had 87.9% accuracy in identifying Lewis antigen-negative patients.

The study also found that Lewis antigen-negative patients with a CA19-9 level of 7 units/mL or less experienced outcomes comparable to patients with CA19-9 levels of greater than 200 units/mL. By examining overall survival (OS) based on CA19-9 levels, the authors determined that patients in the two intermediate strata had the longest OS: those with CA19-9 levels between 7 units/mL and 37 units/mL had a median OS of 23.2 months, while patients with CA19-9 levels greater than 37 units/mL to 200 units/mL had a median OS of 22 months.

In contrast, patients with CA19-9 levels of 7 units/mL or less had a median OS of 13.5 months, which was comparable to the median OS of 12.8 months seen in patients with CA19-9 levels of more than 200 units/mL, suggesting that CA19-9 levels of 7 units/mL or less may be prognostic of high-risk disease.

“These data tell us that the conventional normal CA19-9 range of less than 37 does not distinguish between true low tumor burden and Lewis-negative status,” Su said.

To compensate for the lack of routine genotyping in clinical practice, Su and his coauthors suggested that using a dual-threshold CA19-9 model in which a cutoff of 7 units/mL or less will identify high-risk Lewis antigen-negative patients, whose clinical outcomes will likely parallel those of patients with a CA19-9 level of more than 200 units/mL.

“Using a CA19-9 cutoff of 7, we achieved about 88% accuracy in identifying Lewis antigen-negative patients, whose prognosis is poor despite ‘normal’ or very low CA19-9 values,” said Su. “Using this cutoff was a practical surrogate for identifying these patients without the need for genotyping, thereby improving their risk stratification and preventing underestimation of their disease severity.”

The study’s limitations include the fact that the study was conducted entirely in Taiwan, meaning the findings’ implications for populations outside of East Asia remain unknown. An additional limitation is that CA19-9 assays can vary across laboratories. An international validation study is planned as the next step.

The study was funded in part by the Ministry of Science and Technology, Taiwan; National Health Research Institutes, Taiwan; National Cheng Kung University College of Medicine/National Health

Research Institutes Joint Research Program, Taiwan; Kaohsiung Medical University, Taiwan; and National Science and Technology Council/T-Star Center. Su declares no conflicts of interest.

Download a photo of Su Download a graphic related to the study