ENDOVAGINAL ULTRASONOGRAPHY IN THE DIAGNOSIS OF ADENOMYOSIS UTERI - IDENTIFYING THE PREDICTIVE CHARACTERISTICS

Objective To evaluate prospectively the role of endovaginal ultrasonog raphy in the diagnosis of adenomyosis and to identify predictive chara cteristics. Setting In Vitro Fertilisation Unit, Hammersmith Hospital. Subjects Fifty-six women with menorrhagia and dysmenorrhea. Design En dovaginal sonography was performed and uterine body morphometry and my ometrial echogenicity were assessed. The sonographic suspicion of aden omyosis was scored high or low depending on the degree of uterine enla rgement, uterine asymmetry not due to fibroids and heterogenicity of m yometrial echoes. The sonographic diagnosis was compared either with t he histological findings after hysterectomy (n=34) or to the appearanc es on magnetic resonance imaging (n=22). Results Adenomyosis was diagn osed in 28 patients: 15 by histology and 13 by magnetic resonance imag ing. Endovaginal ultrasound demonstrated a sensitivity of 86%, a speci ficity of 50%, a positive predictive value of 86%, and a negative pred ictive value of 77%. Uterine morphometry alone did not predict adenomy osis: although the mean length of the longitudinal, anteroposterior an d transverse axis was larger in uteri with, compared with those withou t, adenomyosis, this did not reach statistical significance. The uteri ne asymmetry ratio was 1.43 (SD 0.6) and 1.34 (SD 0.4) (P=0.26) in ute ri with and without adenomyosis, respectively, but in the presence of adenomyosis the mean posterior wall was significantly thicker than the mean anterior wall: 25.6 (SD 6.6) mm compared with 21.8 (SD 5.0) mm, P=0.02. Therefore, adenomyosis was best predicted on the basis of ill- defined myometrial heterogeneity. However, leiomyomas and various echo genic shadows and artefacts often complicate subjective assessment of the myometrial echogenicity. Conclusion Endovaginal sonography in symp tomatic patients can be a sensitive but not a specific procedure for t he diagnosis of adenomyosis.