External validation of gynecological imaging and reporting data system for sonographic evaluation of adnexal masses

Abstract

Objective: To perform an external validation of Gynecological Imaging and Reporting Data System (GI-RADS) and to assess how referring clinicians value this reporting system in their daily practice. Materials and methods: Prospective observational study comprising 257 women (mean age 40.3 years) and 281 adnexal masses, referred by 20 clinicians to an ultrasound referral center. All women underwent transvaginal or transrectal ultrasound. Presumed diagnosis of the adnexal mass was based on examiner’s subjective impression according to patter recognition analysis. Reporting was performed according to GI-RADS classification. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR+), and negative likelihood ratio (LR−) of the GI-RADS were calculated. The gold standard was histologic diagnosis (benign or malignant) or spontaneous resolution of the cyst during follow-up (benign). Referring clinicians were asked for completing a survey in order to assess how useful they considered this reporting system. Results: In this study, 56 masses were classified as GI-RADS 2, 174 masses were classified as GI-RADS 3, 19 masses were classified as GI-RADS 4, and 32 masses were classified as GI-RADS 5. Among them, 230 masses were removed surgically and 51 masses resolved spontaneously. There were 35 malignant lesions. Sensitivity, specificity, LR+, and LR− were 97.1% (95% confidence interval (CI): 85.5–99.5%), 93.1% (95%CI: 89.2– 95.6%), 14.1 (95%CI: 8.8–22.3), and 0.03 (95%CI: 0.004–0.21) respectively. All 20 referring clinicians reported that GI-RADS was very useful for their clinical decision-making process. Conclusion: GI-RADS shows a high diagnostic performance and is helpful for referring clinicians for taking clinical decisions.