Evaluation of MIR-371A-3P for clinical implementation in a CLIA-certified laboratory.

Abstract

648 Background: MicroRNAs (mi-RNAs) have emerged as promising biomarkers for the detection of viable germ cell tumors (GCT). Our previous work in a research lab demonstrated optimal performance with positive results that had Cq values less than 28, an indeterminate range between 28-35 and negative results greater than 35. In this study, we assessed performance characteristics of miR-371a-3p in a CLIA certified lab at UCSD. Methods: Due to workflows, equipment and scaling, several modifications were required from our original research protocol. Ultimately, serum samples from 53 cases and 139 controls were analyzed using qPCR for miR-371a-3p in the CLIA lab. Initial runs and re-runs were conducted for the indeterminate samples. Metrics used for calculation of performance were sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), area under the curve (AUC) and Youden’s index. Results: We used two analytic pipelines for classifying positive and negative Cq values: 1) a binary cutoff of 26.9 and, 2) an indeterminate range of 25.6 to 31.5. Using the binary cutoff, a sensitivity of 86%, specificity of 98%, PPV of 94%, NPV of 95%, AUC of 0.94 and Youden index of 0.85 was observed. In contrast, initial runs on the indeterminate ranges yielded a sensitivity of 95%, specificity of 99%, PPV of 97%, NPV of 98%, AUC of 0.97 and Youden index of 0.94. However, 36% (51 of 139) controls fell into the indeterminate range. Re-runs on these samples demonstrated similar findings. Conclusions: Our data demonstrated that there are likely lab specific, protocol specific and disease-state specific analytics and thresholding for circulating miR-371a-3p testing. These factors will need to be carefully considered while moving towards clinical implementation.