Incidental discovery of small adnexal tumors: how well do the conventional diagnostic models work?

Background:  The increasing use of diagnostic ultrasound in gynecology has led to a growing number of small adnexal tumors being diagnosed incidentally in asymptotic women. Although numerous diagnostic models have been designed to estimate the preoperative risk of malignancy in symptomatic adnexal masses, it is not known whether these methods are also applicable to small tumors. The aim of this study was to investigate the accuracy of a number of diagnostic models in the preoperative diagnosis of clinically undetectable adnexal masses. Methods:  All patients undergoing surgery for known adnexal masses were assessed preoperatively using clinical examination, transvaginal ultrasound, color Doppler imaging and serum CA125 measurements. The presenting complaint and demographic details were also recorded. The risk of malignancy was then estimated using a variety of previously published morphological scoring systems, Doppler indices of impedance to flow, tumor markers and logistic regression models. The receiver operating characteristic curves were calculated for each diagnostic approach to examine significant differences in their performance for the detection of ovarian cancer. Results:  A total of 115 patients had surgery for adnexal tumors between August 1997 and June 2000. Of these, 90 had benign masses, 6 borderline and 19 malignant. Of the 115 patients, 69 had a clinically detectable cyst and 46 small tumors were found incidentally. In these 46 cases, the best characterization of ovarian tumors was achieved by the use of a logistic regression model, which combined menopausal status, presence of tumor papillarities, blood supply and serum CA125 levels. This model achieved 100% sensitivity with a specificity of 78% in the diagnosis of ovarian malignancy. Conclusion:  A combination of various diagnostic parameters in the form of a logistic regression model is superior to other methods for non-invasive diagnosis of ovarian cancer diagnosed incidentally on ultrasound scan.