Three-dimensional transvaginal ultrasound versus MRI in the diagnosis and classification of congenital uterine anomalies: A systematic review and meta-analysis

Introduction: Accurate diagnosis and classification of congenital uterine abnormalities is paramount for effective decision-making during infertility workup. The present review aims to ascertain the diagnostic performance of three-dimensional transvaginal ultrasound (3D-TVUS) in the assessment of congenital uterine abnormalities and compare it to MRI. Method: The Scopus, Web of Science, Ovid, CINAHL and PubMed databases were systematically searched in accordance with the PRISMA 2020 guidelines. Relevant articles were screened for relevance and evaluated for risk of bias before their inclusion into the analysis. Results: Nine studies were included. With regard to diagnostic accuracy data, 3D-TVUS had higher specificity compared to MRI in diagnosing bicornuate (99 % vs 92 %, p = 0.047) and didelphys (100 % vs 87 %, p = 0.03) and higher sensitivity for septate uterus (99 % vs 81 %, p < 0.001). Inter-method agreement was substantial for hypoplastic, unicornuate, didelphys, bicornuate and septate (Cohen's kappa: 0.65, 0.75. 0.74, 0.61, 0.77), almost perfect for arcuate and iatrogenic (0.82 and 1.0) and substantial (0.7) cumulatively with the ASRM classification system in the absence of a reference standard. Inter-method agreement for the studies using the ESHRE/ESGE classification was perfect for U1, U4 and U5 anomalies (1.0), almost perfect for U2 and U3 anomalies (0.91 and 0.94) and almost perfect (0.91) cumulatively. Conclusions: These findings support the application of 3D-TVUS in the diagnosis and classification of uterine abnormalities, as it combines diagnostic efficacy with practical advantages over MRI, such reduced cost and better tolerability by the patients. Further studies need to validate these findings in the less common subtypes of congenital uterine abnormalities.