Study objective: To determine the utility of the serologic marker CA19-9 in the diagnosis of pancreatic cancer in patients suspected of having a pancreatic disorder.
Design: Blinded study of frozen pedigreed serum samples collected at time of diagnostic evaluation with follow-up review at a mean of 8 years.
Setting: A general university teaching hospital serving both primary and referral patient populations.
Measurements and main results: Sera collected prospectively from 1978 to 1980 from 261 patients undergoing imaging studies of the pancreas (ultrasound, computed tomography, endoscopic retrograde cholangiopancreatography) for a variety of symptoms were assayed for CA19-9 levels, and the results were compared with earlier determinations of other potential markers for pancreatic cancer. In 54 patients ultimately shown to have pancreatic cancer, the CA19-9 assay showed a sensitivity of 70% with a median value of 349 u/mL (normal less than 70 u/mL) and range, 7.3 to 2,859,964 u/mL, whereas specificity of the marker in this population was 87%. The positive predictive value was 59%, and the negative predictive value was 92%. Results of CA19-9 testing in the small group of patients with definitive staging information showed no difference in sensitivity between patients with local/regional disease (n = 6) and those with distant metastases (n = 14), 50% compared with 71% (P = 0.613).
Conclusion: CA19-9 was found to be a more sensitive and specific marker of pancreatic cancer than other serologic markers and should be a useful test in the patient with suspected pancreatic disease.