Comparison of serum miR-371a-3p assay performance by digital droplet PCR and reverse transcriptase quantitative PCR in patients with malignant germ cell tumor.

Abstract

426 Background: MicroRNAs have shown promise as potential germ cell tumor (GCT) biomarkers. In particular, previous work has shown miR-371a-3p alone may serve as a biomarker for GCTs, demonstrating higher sensitivity than current biomarkers. Quantitative reverse transcriptase PCR (RT-qPCR) has been commonly used to measure circulating levels of miR-371a-3p. More recently, miR-371a-3p has also been measured using digital droplet PCR (ddPCR), which has the potential advantage of absolute quantification. Here, we compared the performance of miR-371a-3p as quantified by RT-qPCR and ddPCR. Methods: Patient samples from the University of Texas Southwestern (UTSW) Medical Center and the University of Cambridge (UoC) were evaluated using both RT-qPCR and ddPCR, as per current protocols (RT-qPCR) or as per standard manufacturer’s recommendations (ddPCR). A range of clinical scenarios (pre-orchiectomy, chemotherapy-naïve RPLND) were intentionally selected. We compared the performance of the two assays using receiver operating characteristic (ROC) curves and area under the curve (AUC) values. We also determined an optimal threshold for each procedure by maximizing the Youden Index and compared the corresponding estimated sensitivity and specificity. Results: Data were available for 69 patients. Among these patients, 35 (50.7%) had malignant GCT (MGCT) and 34 (49.3%) had either non-MGCT (n=26) or no tumor (n=8). Patients with non-MGCT or with no tumor were considered controls. Cq values were generally lower among patients with MGCT and the number of positive droplets was higher compared with controls. The AUC was 0.96 when using RT-qPCR and 0.82 when using ddPCR to classify patients based on circulating miR-371a-3p. The optimal threshold for ddPCR was determined to be 17.5 positive droplets with a corresponding estimated sensitivity of 71% and 100% specificity. For RT-qPCR, the optimal threshold was determined to be Cq=28.52, with a corresponding estimated sensitivity of 83% and specificity of 100%. Conclusions: RT-qPCR was more sensitive in identifying MGCT patients in our current cohort when compared with ddPCR. Further investigations are required to optimize PCR methodology, particularly for ddPCR, and determine whether e.g., tumor volume and/or clinical context affects ddPCR performance. [Table: see text]