Clinical usefulness of carbohydrate antigen 19-9 as a screening test for pancreatic cancer in an asymptomatic population

Abstract

Background and aim: Although the prognosis for pancreatic cancer is generally poor, it is well known that the survival rate for resected pancreatic cancer is much higher than that for more conservative treatment. The importance of early detection is emphasized for resection of pancreatic cancer. Measurement of serum carbohydrate antigen (CA) 19-9 has shown satisfactory sensitivity and predictive value in symptomatic patients, but no available data has been found on healthy asymptomatic subjects. Thus, the authors aimed to determine the clinical usefulness of CA 19-9 as a screening tool for pancreatic cancer in asymptomatic subjects.

Methods: From December 1994 to November 2000, 70 940 asymptomatic persons visiting the Health Promotion Center at the Samsung Medical Center, Seoul, Korea, participated. All subjects underwent abdominal ultrasonography and serum CA 19-9 measurement. The authors analyzed the sensitivity, specificity, and predictive values of CA 19-9 for detecting pancreatic cancer. Also, those subjects who had a serum CA 19-9 level above the cut-off value were followed up using a serial check of CA 19-9, computed tomography, or endoscopic retrograde cholangiopancreatography.

Results: The number of subjects with a level of CA 19-9 above the cutoff of 37 U/mL was 1063 (1.5%), including four cases diagnosed with pancreatic cancer. The prevalence of pancreatic cancer over the age of 30 years is 13.66 per 100 000 population in Korea. Therefore, the sensitivity is 100% and the specificity 98.5%. However, the positive predictive value of CA 19-9 for detecting pancreatic cancer is only 0.9% in the asymptomatic population.

Conclusion: Mass screening for pancreatic cancer using CA 19-9 levels in asymptomatic subjects is ineffective because of a very low positive predictive value, despite its high sensitivity and specificity.