COMMENTARY

Ultrasound Markers Can Differentiate Benign, Malignant Masses

Peter Kovacs, MD, PhD

Disclosures

May 04, 2016

Predicting the Risk of Malignancy in Adnexal Masses Based on the Simple Rules from the International Ovarian Tumor Analysis Group

Timmerman D, Van Calster B, Testa A, et al.
Am J Obstet Gynecol. 2016;214:424-437.

Background

Ovarian cancer is the eighth most common cancer among women.[1] Owing to the slow progression and initial vague symptoms of the disease, many cases are diagnosed at an advanced stage, when treatments have limited therapeutic benefits.

Certain tumor markers and imaging studies could be used to detect cancer at an earlier point. Ultrasound is a simple test with good access to the ovaries. It can differentiate cystic vs solid lesions, provide accurate assessment of size, follow changes in appearance, and assess vascular supply and flow.

The problem is that the accuracy of ultrasound is operator- and machine-dependent and is not specific enough. Certain benign and malignant lesion cannot be differentiated sufficiently well. This is why various models using several ultrasound markers have been developed.[2]

One of these models—the "simple rules" developed by the International Ovarian Tumor Analysis group—is based on five markers indicative of benign lesions and five markers indicative of malignant lesions.[3] The aim of the current study was to provide risk estimates for malignancy based on these 10 ultrasound markers.

The Study

Data on 4848 patients with ovarian masses who subsequently were operated on and had a histologic diagnosis were collected from 22 centers (some oncologic, some general practice) in several phases (development and validation of the model). During the scans, the following parameters were recorded: cystic or solid nature of the mass, cyst wall irregularity, size of the solid component, size of the papillary projection, acoustic shadows, ascites, and Doppler flow.

Sensitivity, specificity, and positive and negative predictive values were calculated. Risk cutoffs were provided to classify a mass as malignant. It was found that 34% of the lesions were malignant; the rate was higher in oncologic centers. The median age was 42 years for patients with benign lesions and 57 for those with malignant lesions.

The predictive value for malignancy of all 10 ultrasound markers was quantified and then used to obtain an overall estimated risk. The presence of benign ultrasound markers lowered the overall estimated risk, whereas markers indicating malignancy increased it.

The presence of a unilocular cyst was the most predictive for a benign lesion, and the presence of ascites was the most predictive for a malignant lesion. The area under the curve risk estimate to predict malignancy was 0.917 (95% confidence interval, 0.902-0.930) in the validation phase of the study and was similar in oncologic centers and general hospitals.

Without any sonographic findings, the risk for malignancy preoperatively was 48.7% in oncologic centers and 27.5% in other centers. The presence of any of the five benign or five malignant ultrasound features increased or decreased this risk. For example, when more than two benign features and no malignant features were present, the risk for malignancy dropped to 0.29%. Conflicting results (one or more benign and malignant feature) were seen in only 3.3% of all cases. On the basis of the number of benign and malignant features present in the mass, an individual risk estimate can be obtained for a given patient that could help her undergo individualized treatment.

When this model was used in the validation phase of the study, 22.8% of the patients considered at low (<1%) risk (sensitivity, 99.7%; specificity, 33.7%; positive predictive value, 44.8%; negative predictive value, 98.9%). In 48.5% of the patients, the risk estimate for malignancy was > 30% (sensitivity, 89%; specificity, 84.7%; positive predictive value, 75.4%; negative predictive value, 93.9%).

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....