Two- and three-dimensional transvaginal ultrasonography for diagnosis of adenomyosis of the inner myometrium

Research question: How diagnostically accurate is two-dimensional (2D-TVS) compared with three-dimensional transvaginal ultrasonography (3D-TVS) in diagnosing adenomyosis of the inner myometrium. What is the most accurate combination of ultrasonographic features? Design: Premenopausal women (n = 110) scheduled for hysterectomy or transcervical resection of the endomyometrium owing to abnormal uterine bleeding were consecutively enrolled. All participants had real-time 2D-TVS and, later, blinded off-line 3D-TVS to diagnose adenomyosis. Results were compared with a detailed histopathological examination of the inner myometrium as gold standard. Results: Prevalence of adenomyosis of the inner myometrium was 29%. For 2D-TVS and 3D-TVS, respectively, the diagnostic accuracy was sensitivity 72% (95% CI 53 to 86) and 69% (95% CI 50 to 84); specificity 76% (95% CI 65 to 85) and 86% (95% CI 76-93); and area under the curve (AUC) 0.74 (95% CI 0.7 to 0.8) and 0.77 (95% CI 0.7 to 0.9). Specificity of 3D-TVS was not statistically significantly better than 2D-TVS; the difference between them almost reached statistical significance (P = 0.06). The most accurate three-dimensional feature was junctional zone irregularity (J(max)-JZ(min) >= 5mm) (AUC: 0.78). A combination of two or more two-dimensional and two or more three-dimensional features was highly accurate (AUC: 0.77). Conclusions: For diagnosing adenomyosis of the inner myometrium, 3D-TVS offers a high accuracy similar to 2D-TVS. Identification of junctional zone irregularity with 3D-TVS may be beneficial to diagnosis. Two or more two-dimensional features and two or more three-dimensional features combined may give a more objective diagnosis, and may be useful for clinical practice and future research.