Introduction: Endometriosis (END) is a debilitating gynecological disorder. Clinical examination, imaging, and laparoscopy can provide a definitive diagnosis of END. The discovery of non-invasive biomarkers is necessary to overcome the disadvantages of surgical practice. MicroRNAs (miRNAs) are a family of small non-coding RNAs that, thanks to their high stability in biologic fluids, can represent excellent biomarkers for END. The purpose of this study was to develop a diagnostic model based on serum miRNA to identify patients affected by END. Methods: Serum samples were collected before surgery and total RNA was extracted from 400 uL of serum of 67 patients with END and 60 patients with benign gynecological non endometriotic pathology, used as controls (CNT). miRNA expression profiling was performed via TaqMan OpenArray technology. For the development of the diagnostic algorithm to discriminate between END and CNT, the ‘Random Forest’ method was used, along with “Recursive Forward Elimination.” Results: To explore the discrimination ability between END and CNT, a Random Forest algorithm based on 18 miRNAs was developed. The diagnostic performance of this model was characterized by Area Under the Curve (AUC)=0.863, False Positve Rate (FPR)=0.227, False Negative Rate (FNR)=0.196, Specificity=0.773 and Sensitivity=0.804. Discussion: Our study identified a diagnostic algorithm that shows good performance in discriminating between END and CNT, supporting the potential role of circulating miRNA as non-invasive biomarkers of END.
Development of a diagnostic model based on serum microRNAs for the diagnosis of endometriosis
Bergamaschi C.;Ravaggi A.;Zanotti L.;Conforti J.;Epis M.;Gennarelli M.;Ciravolo G.;Calza S.;Bignotti E.;Odicino F. E.
2025-01-01
Abstract
Introduction: Endometriosis (END) is a debilitating gynecological disorder. Clinical examination, imaging, and laparoscopy can provide a definitive diagnosis of END. The discovery of non-invasive biomarkers is necessary to overcome the disadvantages of surgical practice. MicroRNAs (miRNAs) are a family of small non-coding RNAs that, thanks to their high stability in biologic fluids, can represent excellent biomarkers for END. The purpose of this study was to develop a diagnostic model based on serum miRNA to identify patients affected by END. Methods: Serum samples were collected before surgery and total RNA was extracted from 400 uL of serum of 67 patients with END and 60 patients with benign gynecological non endometriotic pathology, used as controls (CNT). miRNA expression profiling was performed via TaqMan OpenArray technology. For the development of the diagnostic algorithm to discriminate between END and CNT, the ‘Random Forest’ method was used, along with “Recursive Forward Elimination.” Results: To explore the discrimination ability between END and CNT, a Random Forest algorithm based on 18 miRNAs was developed. The diagnostic performance of this model was characterized by Area Under the Curve (AUC)=0.863, False Positve Rate (FPR)=0.227, False Negative Rate (FNR)=0.196, Specificity=0.773 and Sensitivity=0.804. Discussion: Our study identified a diagnostic algorithm that shows good performance in discriminating between END and CNT, supporting the potential role of circulating miRNA as non-invasive biomarkers of END.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
