Abstract

Background: Gastrointestinal cancer antigen CA19-9 is known as a valuable marker for the management of patients with pancreatic cancer.

Methods: The analytical and clinical performance of the Access^® GI Monitor assay (Beckman Coulter) was evaluated on the UniCel^® DxI 800 Immunoassay System at five different European sites and compared with a reference method, defined as CA19-9 on the Elecsys System (Roche Diagnostics).

Results: Total imprecision (%CV) of the GI Monitor ranged between 3.4% and 7.7%, and inter-laboratory reproducibility between 3.6% and 4.0%. Linearity upon dilution showed a mean recovery of 97.4% (SD+7.2%). Endogenous interferents had no influence on GI Monitor levels (mean recoveries: hemoglobin 103%, bilirubin 106%, triglycerides 106%). There was no high-dose hook effect up to 115,000 kU/L. Clinical performance investigated in sera from 1811 individuals showed a good correlation between the Access^® GI Monitor and Elecsys CA19-9 (R=0.959, slope=1.004, intercept=+0.17). GI Monitor serum levels were low in healthy individuals (n=267, median=6.0 kU/L, 95th percentile=23.1 kU/L), higher in individuals with various benign diseases (n=550, medians=5.8–13.4 kU/L, 95th percentiles=30.1–195.5kU/L) and even higher in individuals suffering from various cancers (n=995, medians=8.4–233.8 kU/L, 95th percentiles=53.7–13,902 kU/L). Optimal diagnostic accuracy for cancer detection against the relevant benign control group by the GI Monitor was found for pancreatic cancer [area under the curve (AUC) 0.83]. Results for the reference CA19-9 assay were comparable (AUC 0.85).

Conclusions: The Access^® GI Monitor provides very good methodological characteristics and demonstrates an excellent analytical and clinical correlation with the Elecsys CA19-9. The GI Monitor shows the best diagnostic accuracy in pancreatic cancer. Our results also suggest a clinical value of the GI Monitor in other cancers.

Clin Chem Lab Med 2008;46:600–11.