Comparison of saline infusion sonography (SIS) versus SIS-guided endometrial sampling in the diagnosis of endometrial pathology

Purpose To compare saline infusion sonography (SIS) to tissue obtained from SIS-guided endometrial sampling (SISES) to determine sensitivity and specificity for benign and malignant conditions of the endometrium. Methods Added value of SISES after SIS was evaluated over 5 years. SIS and SISES techniques have been previously described. Women with abnormal uterine bleeding and endometrial findings underwent SIS. Criteria for SISES included atypical polypoid mass, focal irregularity or global endometrial thickening, nondiagnostic SIS, discordance between SIS and endometrial biopsy, or request from the primary team. Final surgical pathology was compared with SIS and SISES. Results One-hundred twenty-three SIS patients had SISES. SIS alone had a sensitivity of 100%, specificity of 42%, positive predictive value of 16%, and negative predictive value of 100%. A total of 120 patients had final clinical or surgical outcomes. Specificity was greatly improved with addition of SISES (95.5%, p < 0.001). Combination of SIS with SISES detected all 12 premalignant and malignant endometrial lesions. Three cases had small foci of hyperplasia (n = 2) or atypia (n = 1) in polyps on final surgical pathology, which were not prospectively identified on SISES. Conclusions SISES significantly improves the specificity in the diagnosis of endometrial lesions. False-negative findings of SISES in polyps in postmenopausal women with bleeding suggest the need for surgical removal in this setting. © 2016 Wiley Periodicals, Inc. J Clin Ultrasound, 2016