Prospective Evaluation of the NETest as a Liquid Biopsy for Gastroenteropancreatic and Bronchopulmonary Neuroendocrine Tumors: An ENETS Center of Excellence Experience

Abstract

Background: There is a substantial unmet clinical need for an accurate and effective blood biomarker for neuroendocrine neoplasms (NEN). We therefore evaluated, under real-world conditions in an ENETS Center of Excellence (CoE), the clinical utility of the NETest as a liquid biopsy and compared its utility with chromogranin A (CgA) measurement. Methods: The cohorts were: gastroenteropancreatic NEN (GEP­NEN; n = 253), bronchopulmonary NEN (BPNEN; n = 64), thymic NEN (n = 1), colon cancer (n = 37), non-small-cell lung cancer (NSCLC; n = 63), benign lung disease (n = 59), and controls (n = 86). In the GEPNEN group, 164 (65%) had image-positive disease (IPD, n = 135) or were image-negative but resection-margin/biopsy-positive (n = 29), and were graded as G1 (n = 106), G2 (n = 49), G3 (n = 7), or no data (n = 2). The remainder (n = 71) had no evidence of disease (NED). In the BPNEN group, 43/64 (67%) had IPD. Histology revealed typical carcinoids (TC, n = 14), atypical carcinoids (AC, n = 14), small-cell lung cancer (SCLC, n = 11), and large-cell neuroendocrine carcinoma (LCNEC, n = 4). Disease status (stable or progressive) was evaluated according to RECIST v1.1. Blood sampling involved NETest (n = 563) and NETest/CgA analysis matched samples (n = 178). NETest was performed by PCR (on a scale of 0–100), with a score ≥20 reflecting a disease-positive status and >40 reflecting progressive disease. CgA positivity was determined by ELISA. Samples were deidentified and measurements blinded. The Kruskal-Wallis, Mann-Whitney U, and McNemar tests, and the area under the curve (AUC) of the receiver-operating characteristics (ROC) were used in the statistical analysis. Results: In the GEPNEN group, NETest was significantly higher (34.4 ± 1.8, p < 0.0001) in disease-positive patients than in patients with NED (10.5 ± 1, p < 0.0001), colon cancer patients (18 ± 4, p < 0.0004), and controls (7 ± 0.5, p < 0.0001). Sensitivity for detecting disease compared to controls was 89% and specificity was 94%. NETest levels were increased in G2 vs. G1 (39 ± 3 vs. 32 ± 2, p = 0.02) and correlated with stage (localized: 26 ± 2 vs. regional/distant: 40 ± 3, p = 0.0002) and progression (55 ± 5 vs. 34 ± 2 in stable disease, p = 0.0005). In the BPNEN group, diagnostic sensitivity was 100% and levels were significantly higher in patients with bronchopulmonary carcinoids (BPC; 30 ± 1.3) who had IPD than in controls (7 ± 0.5, p < 0.0001), patients with NED (24.1 ± 1.3, p < 0.005), and NSCLC patients (17 ± 3, p = 0.0001). NETest levels were higher in patients with poorly differentiated BPNEN (LCNEC + SCLC; 59 ± 7) than in those with BPC (30 ± 1.3, p = 0.0005) or progressive disease (57.8 ± 7), compared to those with stable disease (29.4 ± 1, p < 0.0001). The AUC for differentiating disease from controls was 0.87 in the GEPNEN group and 0.99 in BPC patients (p < 0.0001). Matched CgA analysis was performed in 178 patients. In the GEPNEN group (n = 135), NETest was significantly more accurate for detecting disease (99%) than CgA positivity (53%; McNemar test χ² = 87, p < 0.0001). In the BPNEN group (n = 43), NETest was significantly more accurate for disease detection (100%) than CgA positivity (26%; McNemar’s test χ² = 30, p < 0.0001). Conclusions: The NETest is an accurate diagnostic for GEPNEN and BPNEN. It exhibits tumor biology correlation with grading, staging, and progression. CgA as a biomarker is significantly less accurate than NETest. The NETest has substantial clinical utility that can facilitate patient management.