Gut neuroendocrine tumor blood qPCR fingerprint assay: characteristics and reproducibility

Abstract

We have developed a PCR-based tool that measures a 51-gene panel for identification of gastroenteropancreatic (GEP) neuroendocrine neoplasms (NENs) in peripheral blood. This manuscript assesses the robustness (performance metrics) of this tool with a specific focus on the effects of individual parameters including collection, storage, acid suppressive medication [proton pump inhibitor (PPI)], age, sex, race and food on accuracy. Performance metrics were evaluated using a gold standard (mRNA derived from three individual human neuroendocrine tumor cell lines) and clinical samples using qPCR. One hundred percent of the 51 transcripts were amplified in the gold standard (NEN cell line-derived mRNA) (CQ<35, average efficiency 1.94). The inter- and intra-assay variations were 1%-2%. In clinical samples, 50 of 51 targets (98%) were amplified. The inter- and intra-assay reproducibility ranged between 0.4% and 1.2%. The coefficient of variation (CV) was 5.3%. Expression of the reference gene, ALG9, was robust [low variation, low M-value, high (99.5%) PCR efficiency] and unaffected by sample processing. Test meals, long-term PPI use (>1 year), age, sex and ethnicity had no effect on the signature. Expression of two genes, ALP2 and CD59 correlated strongly with RNA integrity (R=0.72, p<0.001) and could be used to assess storage and processing. The 51 marker gene signature was robust and reproducible, exhibiting acceptable inter- and intra-assay metrics (<5%). Feeding, PPI intake, age, sex and ethnicity do not affect the signature. Expression levels of APLP2 and CD59 are effective surrogate markers of proper sample collection and processing.