P119: Transvaginal ultrasound based soft markers: inter-observer variability and experience

To evaluate the accuracy, user variability and impact of experience on the subjective assessment of transvaginal based soft markers in the evaluation of women with chronic pelvic pain undergoing laparoscopy. Prospective study. 38 consecutive women with CPP had TVS 1–2 weeks before laparoscopy. All patients had a negative pelvic infection screen. Each observer performed a transvaginal scan (TVS), and assessed the pelvis for hard markers and soft markers (reduced ovarian mobility and site-specific pelvic tenderness) of pelvic pathology. The laparoscopic surgeon was blind to the soft marker findings. The average age of the patients was 28 years (range 18–43 years) and the average duration of pain was 31 months (range 6–96 months). 25/38 (66%) patients had pathology while 13/38 (34%) patients had a normal pelvis at laparoscopy. The diagnostic accuracy was 44.7% and 65.8% for observer 1 and 2 respectively. The observers agreed in 22/38 (58%) cases. The simple kappa agreement between observer 1 and 2 was 0.17 (poor agreement). When the data from the last 20 patients were analysed alone, the accuracy of observer 1 increased to 50% (observer 2, 65%), and the observers agreed in 15/20 (75%) cases. The kappa value between the observers increased to 0.47 (moderate agreement). Pelvic pathology or likely normality can be predicted by transvaginal ultrasound. This study shows that experience and technique are important factors in the assessment of women with CPP, and diagnostic accuracy improves with experience.