Sonographic diagnosis of ovarian torsion: Accuracy and predictive factors

Reuven Mashiach*, Nir Melamed, Noa Gilad, Gadi Ben-Shitrit, Israel Meizner

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

109 Scopus citations

Abstract

Objectives-The purpose of this study was to determine the accuracy of sonographic diagnosis of ovarian torsion and the predictive value of typical sonographic signs. Methods-The study included 63 women attending an ultrasound unit of a tertiary obstetrics and gynecology department in 2002 through 2008 who had suspected ovarian torsion on sonography and subsequently underwent laparoscopy. Results-Sonography had diagnostic accuracy of 74.6% for ovarian torsion. Abnormal ovarian blood flow and the presence of free fluid were the most diagnostically accurate isolated sonographic signs (positive predictive values, 80.0% and 89.2%, respectively; negative predictive values, 46.2% and 46.2%). Using combinations of sonographic signs yielded higher specificity and positive predictive values and lower sensitivity and negative predictive values for ovarian torsion. The diagnostic accuracy was largely affected by the ultrasound operator (mean ± SD, 78.8% ± 16.0%; range, 60.0%-100%). Conclusions-In the setting of a specialized ultrasound unit, sonographic diagnosis of ovarian torsion had high (74.6%) accuracy compared with previous reports. The absence of typical sonographic signs does not rule out ovarian torsion, especially when the clinical presentation is suggestive. Basing assessments on multiple sonographic signs, including Doppler evaluation, increases the diagnostic specificity.

Original languageEnglish
Pages (from-to)1205-1210
Number of pages6
JournalJournal of Ultrasound in Medicine
Volume30
Issue number9
DOIs
StatePublished - 1 Sep 2011

Keywords

  • Accuracy
  • Diagnosis
  • Ovarian torsion

Fingerprint

Dive into the research topics of 'Sonographic diagnosis of ovarian torsion: Accuracy and predictive factors'. Together they form a unique fingerprint.

Cite this