Diagnostic Formula for the Differentiation of Adnexal Tumors by Transvaginal Sonography

OBJECTIVETo create a strategy for sonographic differentiation of benign and malignant adnexal tumors in premenopausal and postmenopausal patients. METHODSMultiple sonomorphologic criteria were analyzed prospectively in 754 tumors. Four hundred were found in premenopausal and 354 in postmenopausal wo...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Obstetrical & gynecological survey 1997-03, Vol.89 (3), p.428-433
Hauptverfasser: PRÖMPELER, H J, MADJAR, H, SAUERBREI, W, LATTERMANN, U, PFLEIDERER, A
Format: Artikel
Sprache:eng
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:OBJECTIVETo create a strategy for sonographic differentiation of benign and malignant adnexal tumors in premenopausal and postmenopausal patients. METHODSMultiple sonomorphologic criteria were analyzed prospectively in 754 tumors. Four hundred were found in premenopausal and 354 in postmenopausal women. In a logistic regression model, relevant criteria were selected, and a diagnostic formula for tumor differentiation was derived. RESULTSThere were 165 malignant tumors, of which 37 (9.2%) were found in premenopausal and 128 (36.2%) in postmenopausal women. In both groups, the criteria of solid phase and ascites were the most significant. Further important diagnostic criteria were structure and tumor size in premenopausal women and cyst architecture and tumor surface in postmenopausal women. These results allowed an estimation of the probability of malignancy. Using a cutoff point of 10% for the probability to classify tumors as malignant, the sensitivity and specificity in premenopausal patients were 86.5% and 92.6%, respectively, with an accuracy of 92%. In postmenopausal women, the sensitivity, specificity, and accuracy were 93%, 82.7%, and 86.6%, respectively. Assuming a prevalence as given in the study, the positive and negative predictive values were 54.4% and 98.5% in premenopausal and 75.3% and 95.4% in postmenopausal women. CONCLUSIONSWith four binary criteria, a useful diagnostic formula for tumor differentiation was obtained. However, estimates for sensitivity, specificity, and accuracy may be too optimistic because they were derived from the same data that were already used for model selection.
ISSN:0029-7828
1533-9866
DOI:10.1097/00006250-199703000-00022