Abstract
A critical requirement in neuroendocrine tumor (NET) management is a blood biomarker test that is sensitive, specific and reproducible. We evaluated a PCR-based 51-transcript signature to detect tumors, compared it with chromogranin A (CgA) and examined the confounding effect of proton pump inhibitors (PPIs), which cause falsely elevated CgA levels. The multigene signature was evaluated in two groups. Group 1: 125 prospectively collected NETs: gastroenteropancreatic NETs (n=91, including 42 pancreatic and 40 small intestinal), carcinoids of unknown primary (n=18) and other sites (n=16). Group 2: prospectively collected non-NET patients receiving PPIs (>1 month; dyspepsia, n=19; GERD, n=6; and pancreatitis, n=4) and 50 controls. All samples were analyzed by PCR (marker genes) and ELISA (DAKO-CgA). Sensitivity comparisons included χ(2), non-parametric measurements, and receiver operating characteristic (ROC) curves. Group 1: 123 NETs were PCR-positive (98.4%) compared with 50 (40%) CgA-positive (χ(2)=97.3, P<10(-26)). Significant differences (P<0.001) were noted between pancreas: PCR 95% vs CgA 29.2% (P<10(-9)) and small intestine: 100 vs 58% (P<10(-4)). The multigene test was elevated in all grades (G1-G3), in both local and disseminated disease, and was not normalized by somatostatin analog therapy. It was also elevated in 97% of CgA normal NETs. Group 2: PPI administration increased CgA in 83% and CgA was elevated in 26% of controls. PCR values were not elevated in either group. PCR performance metrics were as follows: sensitivity 98.4%, specificity 100%, positive predictive value 100%, negative predictive value 97.8%, and the ROC-derived area under the curve (AUC) was 0.997. These were significantly better than CgA (all metrics <60%; AUC, 0.54; Z-statistic, 10.44, P<0.0001). A 51-panel multigene blood transcript analysis is significantly more sensitive than plasma CgA for NET detection and is unaffected by acid suppression therapy.