RESUMO
BACKGROUND AND AIM: Colorectal cancer (CRC) remains the second most frequent cause of cancer deaths in the United States. Blood tests using tumor-related antigens aid in diagnosing CRC. However, higher sensitivity and specificity are needed before an acceptable tumor antigen blood test for CRC is clinically useful. This study describes the diagnostic accuracy of an enzyme-linked immunosorbent assay for the CA11-19 serologic tumor antigen for the detection of CRC. METHODS: Serum specimens were obtained from 522 colonoscopy-confirmed subjects in institutional review board-approved studies. Specimens were blind coded. CA11-19 levels were determined by using enzyme-linked immunosorbent assay analysis. The results were tabulated for categories of normal, hyperplastic polyps, benign GI, adenomatous polyps, and CRC based on their final diagnosis. RESULTS: When a cutoff of 6.4 units/mL for normal is used, the CA11-19 level was elevated in 128 of 131 of CRC subjects, for an observed sensitivity of 98% (95% confidence interval, 93.1%-99.5%). Normal levels were found in 87% of normal subjects (90/103) and 83% of those with benign GI diseases (185/223). When combined, this yields an observed specificity of 84% (95% confidence interval, 80.0%-87.9%). CONCLUSION: CA11-19 is a serologic tumor marker for the diagnosis of CRC with a sensitivity of 98% and specificity of 84%. This high sensitivity means that this test will detect 43 of 44 cases of CRC presented. For those older than 50 years of age, it has a positive predictive value of 3.6% and a negative predictive value of 99.98%. Additional prospective studies are needed to further clarify the use of CA11-19 as an aid in the diagnosis of CRC.
