Obstetric anal sphincter injuries detection by digital rectal examination and impedance spectroscopy: post-hoc analysis of the results from a prospective, comparative, multicentre clinical study

Objective: To evaluate the clinical performance of digital rectal examination (DRE) and machine learning-assisted impedance spectroscopy (ONIRY) for detecting obstetric anal sphincter injuries (OASI) compared to three-dimensional endoanal ultrasound (EAUS). Design: A post-hoc analysis of data from a prospective, comparative, multicentre, international study ( NCT04903977 ). Setting: Poland, Czech Republic, Slovakia, and Spain. Population: 152 vaginally delivered women to 8 weeks post-partum. Methods: Participants were divided into two groups based on EAUS results: Group I (no OASI, including 1st and 2nd degree perineal tears per OASIS classification) and Group II (OASI, including 3rd and 4th-degree tears per OASIS classification). DRE and impedance measurements using the ONIRY system were performed. The machine learning (ML) model for ONIRY was trained and validated. The diagnostic outcomes of ONIRY and DRE were compared to EAUS. Main Outcome Measures: Accuracy, sensitivity, and specificity. Results: DRE's performance versus EAUS was 67.8% for accuracy, 44.3% for sensitivity, and 83.5% for specificity. After completion of ML model training, impedance spectroscopy's accuracy, sensitivity, and specificity were 87.0%, 90.6%, and 84.6%, respectively, compared with EAUS. Conclusion: DRE is insufficient for OASI detection. The ML-assisted impedance spectroscopy demonstrated higher accuracy and could facilitate rapid OASI detection post-delivery. Funding: The study was financed by the European Union's Fast Track program through the Polish National Centre for Research and Development (POIR.01.01.01-00-0726/18)