Article Text

Download PDFPDF
EP452 Ultrasound diagnosis of ovarian tumours by pattern recognition
  1. J Knez,
  2. G Gruškovnjak and
  3. M Pakiž
  1. University Medical Centre Maribor, Maribor, Slovenia

Abstract

Introduction/Background Correct diagnosis is crucial for appropriate treatment of ovarian tumours. Ultrasound is the first diagnostic tool used in this process. Several diagnostic models have been developed for classification of adnexal masses as benign or malignant, but pattern recognition by experienced observer remains one of the most reliable approaches. Our aim was to compare the diagnostic accuracy of transvaginal ultrasound in ovarian tumours classification between operators specifically focused on gynaecological oncology and those practicing general gynaecology.

Methodology We conducted a retrospective analysis of all ovarian tumors treated surgically in the period between January 2016 and December 2017. All patients were seen in two university based units - Department for gynaecologic oncology (GO) and Department for general gynaecology (GG). All women had transvaginal ultrasound examination prior to surgery and all tumours were classified into three categories - benign, malignant or likely malignant by the use of pattern recognition.

Results Ultimately, 149 women were included in the analysis, 76 were seen by GO and 73 by GG. After histological examination of specimens, 37 women (25%) had various types of malignant tumours, 19 (13%) had borderline tumours and 93 (62%) had benign tumours. Sensitivity of ultrasound analysis for malignant tumour categorisation was 100% when performed by GO and 58% when performed by GG. The specificity was 41% when performed by GO and 87% when performed by GG. This translates to negative predictive value of 100% by GO and 91% by GG.

Conclusion Ultrasound diagnosis of ovarian tumours by pattern recognition is a reliable diagnostic method that rarely results in false negative diagnosis leading to additional surgical procedures. When performed by operators specifically focused on gynecologic oncology, the false negative rates seem to be lower.

Disclosure Nothing to disclose.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.