Ovarian Malignancy Risk Factors In The Performance Of Color Doppler Ultrasound

objective: to discuss the risk factors which could distinguish the malignanttumors based on the features of color ultrasound images., which belong to193cases of clincical ovarian tumor patients. We aim to build up a three-level riskstratification from which we could use alone to tell the risk factors of themalignant tumors, and test the clinical values of the risk stratification we builtup.Methods: All the patients are chosen from the First Hospital of JilinUniversity, Bethune,in the department of gynecology, starting from January2012to December2012by the first diagnosis for ovarian tumors patients (193cases).All of them are given color Doppler ultrasound examinationpreoperatively together with pathological examination postoperatively, and arediagnosed finally by the FIGO surgical–pathological standards. Regardless ofthe pathological diagnosis, we only reviewed each patient’s color ultrasoundimages retrospectively, and concluded11kinds of features. The probability ofmalignant or borderline cases in two tests is similar. Using the binary logisticregression analyses, by which we made the malignant and borderline ovariantumors as a group,we have identified the most relevant features associated withmalignant or borderline tumors. According to the features we created athree-level risk stratification, from which we could tell the risk factors of the malignant or borderlin tumors. The risk stratification was applied in testing teststo assess its clinical value.Results: High, intermediate and low risk factors were concluded by thebinary logistic regression analyses. High risk factor (means to be malignanttumors at highest risk) is inner blood inside the tumor. Intermediate riskfactors(means the risk of malignant tumors followed by high risk factors)include the features of a thick wall, a thick septa, and lesions with irregular solidcomponents.Low risk factors（means to be malignant tumors at a lower risk）include simple cyst, thin septa, turbid echoes, and low level echoes. We appliedthe risk stratification in the testing tests, assuming that the cases which containshigh-risk or intermediate risk factors were malignant or borderline, and thosewhich contains low-risk factors only were benign, so the sensitivity, specificity,positive and negative predictive values of identifying the malignant orborderline tumors were92.4%、58.2%、53.5%and93.7%, comparing to the finalpathological results. Assuming that the cases that contains high-risk only weremalignant or borderline, and those which contains intermediate or low-riskfactors were benign, so the sensitivity, specificity, positive and negativepredictive values were63.6%、81.1%、65%and81.1%.Conclusion: The risk stratification created by the binary logistic regressionanalyses could be used in clinal to evaluate the risk of the chance to be amalignant or borderline tumor. For the ovarian tumor patients, if the colorultrasound results can exclude the high or intermediate risk factors, it is a low probability of malignancy or borderline, and we doctors can admit the patientswith a safe conservative follow-up.