Background
Epithelial ovarian cancer (EOC) has poor survival given that there are no effective early detection strategies and non-specific early disease symptoms; thus, the disease is largely diagnosed at advanced stage. MicroRNAs (miRNAs) are small (~18–24 nucleotide) non-coding RNAs that regulate gene expression post-transcriptionally, and have emerged over the past decade as promising biomarker candidates for a variety of diseases, including cancer. Our prior research shows the potential of a miRNA signature in blood for ovarian cancer earlier detection.
 
Objectives
The overarching objectives of our study are to validate serum miRNA profiles as biomarkers for early detection in samples collected from cases <=3 years prior to EOC diagnosis and matched controls. We conducted a feasibility study to optimize absolute miRNA quantification of the miRNA candidates with digital PCR (dPCR) assays to translate findings from miRNA sequencing studies to a more cost-, time, and sample volume-efficient assay method.
 
Methods
We have formed the miRPOC consortium with international partners to provide the necessary validation of a miRNA signature for earlier detection of EOC using serum samples and data from prospective cohorts, paired with clinical data, serum and tumor samples from a hospital research biobank. For the pilot study, seven serum samples from Janus Serum Bank were selected among individuals that had no known cancer diagnosis identified through registry linkage. These samples were sequenced, and results from dPCR were compared to those from sequencing data. Fourteen miRNAs were originally evaluated (miR-23b-3p, miR-29a-3p, miR-32–5p, miR-92a-3p, miR-150–5p, miR-200a-3p, miR-200c-3p, miR-203a-3p, miR-320c, miR-320d, miR-335–5p, miR-450b-5p, miR-1246 and miR-1307–5p). EOC cases and matched controls were identified within the Janus Serum Bank cohort in Norway (cases, n=91; controls, n=91) and the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort (cases, n=88; controls, n=88). The thirteen circulating miRNAs signature was measured using dPCR (QIAcuity, QIAGEN). MiRNA measurements were subjected to quality control based on expression levels and summarized using descriptive statistics. Subsequent analyses will include the development and evaluation of predictive machine learning models to assess the early detection potential of miRNA profiles for EOC.
 
Results
Normalised relative miRNA counts from sequencing and absolute measurements using dPCR were highly correlated (Spearman correlation >0.7 for 9 of 13 miRNA). Thirteen of the 14 miRNAs from the original signature were successfully quantified using dPCR, and miR-1246 was excluded due to non-specific binding in the pilot. Analyses evaluating the early detection potential of miRNA signature are ongoing, and results will be presented at the conference.
 
Conclusions  
miRNA levels can be reliably translated from measurement with sequencing to dPCR, including in long-term biobanked samples. Analyses are currently underway to evaluate the diagnostic discrimination of the miRNA signature in samples collected up to three years prior to diagnosis in cases and matched controls. Forthcoming results to be presented will clarify the role of this signature of miRNAs in ovarian cancer earlier detection.