Somatostatin receptor scintigraphy versus chromogranin A assay in the management of patients with neuroendocrine tumors of different types: clinical role

Abstract

Current diagnosis and staging of neuroendocrine tumors (NETs) are significantly improved by the introduction of the chromogranin A (CgA) assay in plasma or serum as a tumor marker, and by the use of somatostatin receptor scintigraphy (SRS) for tumor localization. However, the clinical role of CgA assay compared with SRS in the management of NETs has not been well elucidated. Sixty-three consecutive patients with a histological diagnosis of NET underwent plasma CgA assay and SRS for tumor staging (23 cases), evaluation of tumor response (18 cases) and evaluation of tumor recurrence on follow-up (22 cases). Twenty-one patients had well-differentiated neuroendocrine tumors (WDNETs: 18 gastroenteropancreatic tumors and three lung NETs); 22 patients had well-differentiated neuroendocrine carcinomas (WDNECs: 17 gastroenteropancreatic carcinomas, two lung neuroendocrine carcinomas and three neuroendocrine carcinomas of unknown origin) and 20 patients had poorly differentiated neuroendocrine carcinomas (PDNECs: 14 extra-pulmonary small-cell carcinomas and six Merkel cell carcinomas). Almost all (58 of 63) NETs were non-functioning. The quantitative determination of CgA was performed in plasma using an enzyme immunoassay with a cut-off value fixed at 34 U/l. Scintigraphies with indium 111-DTPA-octreotide ((111)In-pentetreotide) included whole-body images and single photon emission computed tomography (SPECT) scans of the chest and abdomen. SRS results were compared with CgA findings and final clinical data. The overall sensitivity of SRS and CgA, based on the final clinical data, was 77% and 55%, respectively, whereas the specificity of both SRS and CgA was 94%. Concerning tumor type, SRS accuracy was 95% for WDNETs, 86% for WDNECs and 60% for PDNECs; CgA accuracy was 76% for WDNETs, 68% for WDNECs and 50% for PDNECs. With regard to disease extent, SRS sensitivity was 100% for limited disease and 72% for advanced disease; CgA sensitivity was 43% for limited disease and 57% for advanced disease. In our NET series, SRS proved to be more sensitive than CgA, with equivalent specificity. Tumor differentiation influences the sensitivity of SRS and CgA analysis. In addition, the plasma CgA level is related to tumor secretory activity. Nevertheless both SRS and CgA should be considered useful tools in the diagnostic work-up of NET patients.