Mark Kidd, Srinivas V. Koduru, Abdel Halim
Background/Objectives: Blood-based RNA diagnostics requires rigorous analytical validation prior to clinical implementation. NETest2.0® is a machine learning-enhanced 55-gene mRNA assay designed to detect neuroendocrine tumor (NET)-associated gene expression signatures in peripheral blood. Methods: Analytical performance was evaluated in accordance with CLSI EP05, EP06, EP07, and EP17 guidelines and ISTA 7D 2007 standards. Accuracy was assessed in 973 samples (568 NETs, 219 non-NET cancers, 186 controls). Precision, sensitivity, linearity, and specificity were evaluated using cell line-spiked blood, patient samples, and interference testing. Stability and robustness were assessed under varying storage and transport conditions. Results: NETest2.0® demonstrated > 90% diagnostic accuracy. All transcripts were consistently amplified (Ct < 35; mean efficiency 1.94). Intra-assay coefficients of variation (CVs) were 0.56% (Ct) and 2.07% (score), while inter-assay CVs were 4.12% and 6.85%, respectively. High concordance across operators and instruments was observed (r = 0.82–0.94, p < 0.0001). The limit of detection was 2.2 cells/mL RNA-equivalent, with 95% detection at 2.29 cells/mL; the limit of quantification was 8.6 cells/mL. Assay output was linear across defined ranges and unaffected by endogenous or exogenous interferents. No associations were identified with demographic, hematologic, renal, or hepatic variables. Samples remained stable at ambient temperature for up to 10 days and at −80 °C for up to 5 years, with no impact from shipping conditions. Conclusions: NETest2.0® demonstrates high analytical sensitivity, precision, and robustness, supporting its validity for clinical application in NET management.