To compare the diagnostic accuracy of transvaginal sonography (TVS) and saline infusion sonography (SIS) for detecting intracavitary abnormalities in premenopausal women with abnormal uterine bleeding. Consecutive premenopausal women who underwent hysterectomy for abnormal uterine bleeding were included. All women underwent TVS and SIS before their hysterectomy. The findings at TVS and SIS were compared with the findings of the hysterectomy specimen. Sensitivity, specificity, and likelihood ratios were calculated. The results of 50 patients with abnormal uterine bleeding were evaluated. Histological examination revealed normal endometrial histology in 27 patients, submucous myomas in 13 patients and intracavitary polyps in 10 patients. The sensitivity of TVS in directly visualizing intracavitary abnormalities was 61% for a specificity of 96%. The likelihood ratio of presence of an intracavitary abnormality was 16 and the likelihood ratio of absence of such a finding was 0.41. When defining abnormality at TVS as direct visualization of an intracavitary abnormality or an increased endometrial thickness (cut-off level 5 mm), the sensitivity of TVS was 87% for a specificity of 56%, with corresponding likelihood ratios of 2 and 0.23, respectively. The sensitivity and specificity of SIS was 100% and 85% with likelihood ratios of 6 and 0.0, respectively. No intracavitary abnormality was missed by SIS. The diagnostic accuracy of SIS is higher than the accuracy of TVS. A combined approach using endometrial thickness measurement by TVS and, reserving SIS for patients with increased (> 5 mm) endometrial thickness, or endometrium inadequately visualized on TVS, is the optimal method of reducing the hysteroscopy rate.
Read full abstract