Chromogranin A as a marker of neuroendocrine neoplasia: an Italian Multicenter Study

Abstract

Elevated circulating chromogranin A (CgA) levels are found in neuroendocrine tumors (NETs), but the diagnostic usefulness of this marker is still debatable. To assess the role of CgA for the diagnosis of gastroenteropancreatic (GEP) NETs and the identification of metastatic patients, an Italian multicenter observational study has been performed. CgA was evaluated in 202 GEP NET patients by IRMA and ELISA. The cutoffs for diagnosis and presence of metastases were identified by receiver-operating characteristic (ROC) curve. We found good correlation between IRMA and ELISA. The ROC analysis identified a cutoff of 53 ng/ml for IRMA and 16 U/l for ELISA as discriminating between controls and patients with active disease (sensitivity 71.3 and 84%; specificity 71 and 85% respectively). Metastases were present in 123 patients, having significantly higher CgA levels than patients without metastases. ROC analysis identified a cutoff of 146 ng/ml for IRMA and 67.3 U/l for ELISA as discriminating between patients with and without metastases (sensitivity 57 and 63.3%; specificity 55.6 and 71.4% respectively). For pancreatic NETs positive and negative predictive values were 84 and 78% respectively (90% specificity and 68% sensitivity). We found lower CgA levels in patients with extensive metastatic spread than in those with liver metastases only. These data assess the role of CgA evaluation in GEP NETs, and demonstrate that higher CgA levels associate with metastatic disease, confirming that CgA levels can provide a helpful practical biochemical marker for the clinical management of NETs, but with low sensitivity and specificity.

Impact of calibration fitting models on the clinical value of chromogranin A

“…Increased concentrations of circulating chromogranin A (CgA) are characteristic of neuroendocrine tumors (NETs), although the diagnostic and prognostic usefulness of this marker is widely debated (1,2). Some studies reported low sensitivity and specificity of CgA concentrations for the diagnosis of NET.…”

“…Some studies reported low sensitivity and specificity of CgA concentrations for the diagnosis of NET. Concentrations may be dependent on tumor features, such as: type, secretory activity, neuroendocrine differentiation degree and total burden (1,2). However, several lines of evidence show that the diagnostic accuracy is affected primarily by the type of assay used to measure CgA (1,(3)(4)(5).…”

“…Concentrations may be dependent on tumor features, such as: type, secretory activity, neuroendocrine differentiation degree and total burden (1,2). However, several lines of evidence show that the diagnostic accuracy is affected primarily by the type of assay used to measure CgA (1,(3)(4)(5). As a matter of fact, it has been reported that in healthy subjects, there is a discordance rate of ;20% between CgA concentrations measured using the two commercially available methods: immunoradiometric assay (IRMA) and enzyme-linked immunosorbent assay (ELISA) (1).…”

“…However, several lines of evidence show that the diagnostic accuracy is affected primarily by the type of assay used to measure CgA (1,(3)(4)(5). As a matter of fact, it has been reported that in healthy subjects, there is a discordance rate of ;20% between CgA concentrations measured using the two commercially available methods: immunoradiometric assay (IRMA) and enzyme-linked immunosorbent assay (ELISA) (1). The discordance has been ascribed to the different antibodies that are used to measure CgA-derived peptides (1).…”

“…As a matter of fact, it has been reported that in healthy subjects, there is a discordance rate of ;20% between CgA concentrations measured using the two commercially available methods: immunoradiometric assay (IRMA) and enzyme-linked immunosorbent assay (ELISA) (1). The discordance has been ascribed to the different antibodies that are used to measure CgA-derived peptides (1). In addition, the Italian program for CgA external quality control described the pattern of differences between marker concentrations measured by these assays.…”