Background: CA19-9 is a tumor marker for pancreatic cancer, cholangiocarcinoma, and other malignancies. However, its sensitivity and specificity is suboptimal in clinical practice, which we hypothesized limits its clinical utility.
Aims: To evaluate the clinical utility and limitations of CA19-9 as a tumor marker.
Methods: We performed a retrospective review of CA19-9 levels (U/ml) in 483 consecutive patients between 2006 and 2008 at two university hospitals. We abstracted clinical, radiographic, and pathological data and final diagnoses. Descriptive and non-parametric analyses were performed.
Results: Patients presenting with jaundice had the highest CA19-9 (420) compared to other complaints (<20) (p<0.01). The indications with the highest CA19-9 had evidence of biliary obstruction (71), liver mass (54), and pancreatic head mass (27) compared to other indications (<15) (p<0.01). The diagnoses with the highest CA19-9 (p<0.01) were cholangiocarcinoma (476), pancreatic cancer (161), and choledocholithiasis (138). Using a receiver operator curve to evaluate CA19-9, the area under the curve was 0.7 when evaluating all patients for pancreatic cancer or cholangiocarcinoma or patients with pancreatic head mass for pancreatic cancer.
Conclusions: This study found that for pancreatic cancer and cholangiocarcinoma, CA19-9 had poor clinical utility as a tumor marker and did not change patient management. Elevations in CA19-9 were associated with biliary obstruction based on clinical history, laboratory data, and diagnoses.