J Gastroenterol Hepatol. 2004;19(2) 

In This Article


The prognosis of pancreatic cancer is very poor and >90% of cases cannot be resected curatively at the time of the diagnosis. Clinical diagnosis of pancreatic cancer is often difficult because symptoms are vague and frequently non-specific, and thus it is generally recognized at a very advanced stage. Surgery at the stage of asymptomatic and small-sized pancreatic cancer remains the only conceivable method for curing this disease. To improve the prognosis and therapeutic results for pancreatic cancer, it is essential to diagnose pancreatic cancer in the early stages. Thus, various kinds of radiological studies and serum tumor markers are being used in the clinical situation. Abdominal USG and measurement of the CA 19-9 level are popular screening methods for detecting pancreatic diseases. The sensitivity, specificity, and positive predictive value of CA 19-9 are known to be very high in diagnosing pancreatic cancer in patients suspected of pancreatic diseases.[5,6,7,8,9,14,15,16]

The CA 19-9 level is elevated in 87% of pancreatic cancer patients, and serum CA 19-9 concentration is highly correlated to the tumor size in most, if not in all, patients with pancreatic cancer.[17,18]

However, the location of the tumor in the pancreas and the degree of differentiation of the tumor are not related to the level of CA 19-9.[13]

Approximately 13% of the patients with pancreatic cancer have normal levels of CA 19-9, which was shown in cases with large-size tumors as well as small-size tumors. The cause of this is uncertain, but it has been explained that about 5% of Caucasians do have not Lewis A antigen, which synthesizes CA 19-9.[19,20] CA 19-9 is also a useful tumor marker for follow up of pancreatic cancer after operation. The patients whose levels of CA 19-9 had increased before the operation and then decreased to within the normal level 2 weeks after the removal of the tumor showed a higher survival rate than patients whose CA 19-9 levels remained increased after the operation. During the follow up after the operation, an increase in CA 19-9 without abnormal findings on USG or CT scan can be used as a marker for early detection of recurrence.[5,21] The increased level of CA 19-9 can be detected in other malignancies, such as liver, biliary, gastric, or colonic cancer.[1,2] Also, the present study showed that an increased CA 19-9 level (>100 IU/mL) could be diagnosed as lung cancer, colon cancer, gastric cancer, hepatocellular carcinoma, or ovarian carcinoma. It may be useful to evaluate other possible malignant neoplasm besides pancreatic cancer in cases with a highly elevated serum CA 19-9 level.

Similar to the present study, a study was carried out in Japan to examine the usefulness of CA 19-9 as a screening test.[22] Among 12 840 asymptomatic individuals over the age of 40 years, pancreatic cancer was detected in only four cases and curative resection was possible in only one case. In contrast, among 8706 symptomatic outpatients, 104 patients were diagnosed with pancreatic cancer and 47 patients underwent curative resection. Therefore, these authors reported that the screening test of CA 19-9 was not useful in the asymptomatic population and has low cost-effectiveness, but is useful for symptomatic patients.

In the present study, only four out of 70 940 asymptomatic patients were detected with pancreatic cancer and only two underwent curative resection.

In Korea, the official prevalence rate of pancreatic cancer provided by the government is not yet available. Therefore, the prevalence rate of pancreatic cancer in persons over the age of 30 years was obtained based on the National Health Insurance statistical data.[13] The CA 19-9 level was increased in all four patients diagnosed with pancreatic cancer and the sensitivity of CA 19-9 was 100% eventually. Even if the size of the population is 70 940 persons, only one pancreatic cancer patient diagnosed during a routine check can influence the sensitivity greatly, because pancreatic cancer is a very rare disease.

However, the sensitivity and specificity should have regularity regardless of the subjects. In order to supplement the defect, the authors investigated the CA 19-9 levels of 83 pancreatic cancer patients who were admitted during the past 2 years. Seventy out of 83 patients showed increased CA 19-9 levels, which revealed the sensitivity of CA 19-9 to be 84.3%. The positive predictive value is 0.9% when the sensitivity is 100%, which shows little difference from 0.87% when the sensitivity is 84.3%. These data show that when the level of CA 19-9 is increased in the asymptomatic population, fewer than one out of 100 persons is diagnosed with pancreatic cancer.

The present study was based on abdominal USG, which has a sensitivity of about 80% in diagnosing pancreatic cancer. Therefore, it is possible that the person exhibiting a normal CA 19-9 level and a normal USG finding may later be revealed to have pancreatic cancer.[23,24] Moreover, because very small cancers in the early stages have normal CA 19-9 levels, it is difficult to use CA 19-9 as a marker for early diagnosis of pancreatic cancer.[25,26]

Therefore, serum CA 19-9 measurement is not useful as a screening test for pancreatic cancer in the asymptomatic population.

CLICK HERE for subscription information about this journal.