P-344 Feasibility and Clinical Potential of a Serum-based Noninvasive Molecular miRNA Diagnostic for Endometriosis

Abstract Study question Is a reliable blood-based noninvasive diagnostic test for mild and severe endometriosis feasible using five miRNA biomarkers and have clinical utility? Summary answer Noninvasive molecular diagnosis using serum and a novel miRNA panel is both feasible and clinically relevant in resolving discordance between visual laparoscopy and histology. What is known already Endometriosis affects approximately 10% of reproductive-age women worldwide, and for those facing infertility, this figure may climb up to 50%. Endometriosis clinical presentation varies widely, complicating efforts to timely diagnose patients with accuracy. Early detection and timely intervention are critical to improving patient outcomes. Yet, diagnosis is often delayed for 6 to 11 years given that the cause(s) of endometriosis remain unknown and that current diagnostic standard of care is invasive which leads to greater risks. The lack of noninvasive methods to identify and select endometriosis patients further hinders the development of endometriosis-targeted drug therapies Study design, size, duration This is a prospective study designed to demonstrate feasibility of a blood-based molecular miRNA test to detect endometriosis using diagnostic laparoscopy with and without histopathology as a comparator/reference method. Determination of an optimal method for analysis using both machine learning random forrest method as well as a cut-off threshold method is also assessed. Participants/materials, setting, methods Under IRB, blood samples (10 mL) were collected prospectively using standard red-top blood collection tubes from 76 participants with mild to severe symptoms of pelvic pain and/or abnormal bleeding prior to undergoing diagnostic laparoscopy. Following serum isolation and RNA extraction, five differentially expressed and one confirmed endogenous miRNAs were used to assess feasibility in detecting endometriosis. Histology and laparoscopy results for endometriosis were compared to miRNA results using NGS, quantitative PCR and digital PCR. Main results and the role of chance Of the 76 subjects, 84% (64 of 76) were concordant by laparoscopy and histology, with 33 confirmed patients and 31 controls. Using the five differentially expressed and one endogenous control miRNAs and a standard cut-off thresholding method to interpret NGS, digital PCR or quantitative PCR results, an overall accuracy of 85% was observed with each. Similarly, using a random forest model, an overall accuracy of 86% was observed. The remaining 12 subjects that were discordant, were all reported as positive stage I endometriosis by laparoscopy but negative for endometriosis by histopathology. In 9 of 12 cases, miRNA analysis confirmed a negative for endometriosis result which was consistent with histology. In the other 3 cases, miRNA analysis confirmed a positive endometriosis result consistent with laparoscopy which may represent tissue bias leading to false-negative histology result. Limitations, reasons for caution This is an initial study and the accuracy reported is a demonstration of feasibility. Formal perfomance analysis is underway with larger sample size. Wider implications of the findings The results of this study confirm that noninvasive diagnosis of endometriosis based on serum miRNAs is feasible with accuracy equivalent to gold standard. In addition, unnecessary invasive laparoscopy may be avoided with a reliable noninvasive test to diagnose endometriosis. Trial registration number No