Objective: The incidence of enteropancreatic neuroendocrine tumours (NET) is increasing. Chromogranin A (CgA) in plasma is a marker in patients suspected of NET tumours. CgA, however, is a precursor protein subjected to cellular processing that challenges quantitation and hence the use of CgA in diagnostics.Materials and methods: CgA concentrations in plasma sampled from 130 well-characterized patients with small intestinal NETs and from 30 healthy subjects were measured with eight commercial CgA kits, an in-house radioimmunoassay (RIA) and a processing-independent assay (PIA). For the evaluation of diagnostic accuracy, we performed regression analyses and plotted receiver-operating characteristic curves (ROC). The specificity was further assessed by size chromatography.Results: Five commercial assays (Thermo-Fisher, DRG Diagnostics, Eurodiagnostica (RIA and ELISA), and Phoenix), displayed a diagnostic accuracy with area under the curve (AUC) values >0.90, whereas three immunoassays (Yanaihara, CisBio RIA, and CisBio ELISA) discriminated poorly between disease stages (AUC: 0.60-0.78). Compared with the in-house assays, however, even the most accurate commercial immunoassay still missed patients with metastatic disease. Chromatography showed non-uniform patterns of large and small CgA fragments in plasma.Conclusion: Available commercial immunoassays measure CgA in plasma with gross variability. Three commercial CgA immunoassays discriminate so poorly between health and disease that they should not be used. The highest diagnostic accuracy was obtained with processing-independent measurement of total CgA concentrations in plasma.