Ultrasonography models effective at determining lesion type

Ultrasonography models effective at determining lesion type | Image Credit: © Наталия Кузина - © Наталия Кузина - stock.adobe.com.

According to a recent study published in JAMA Network Open, the Assessment of Different Neoplasias in the Adnexa (ADNEX) and Ovarian-Adnexal Reporting and Data System (O-RADS) ultrasonography models are effective in differentiating between malignant and benign lesions.

While adnexal masses are associated with a low incidence of ovarian cancer, the risks of malignant tumors must be addressed because of the high mortality rate linked to ovarian cancer. Adnexal masses are seen in 35% of premenopausal and 17% of postmenopausal women, with over 200,000 in the United States receiving surgery for an adnexal mass per year.

Of women receiving surgery for an adnexal mass, about 10% present with ovarian cancer. Complications of surgeries for benign tumors have been reported in 3% to 15% of cases, indicating a need for a more accurate diagnosis before surgery.

The ADNEX and O-RADS models are the two leading sonographic models for analyzing adnexal masses in Europe, but are not often used in the United States, indicating a need to show their accuracy in a US cohort. Investigators conducted a study to evaluate the diagnostic ability of these models and assess their performance by menopausal status and race.

Patients who received sonograms and were diagnosed with adnexal masses from January 2019 to October 2022 were included in the analysis. An imaging database was consulted to collect complex adnexal lesion data of patients who received sonograms from January 2017 to October 2021.

Eligibility criteria included having a surgical intervention within 180 days or adequate clinical or imaging follow-up,determined by a 10% or more decrease in adnexal mass size, no change, or a classic lesion identification in subsequent imaging. If follow-up was not adequate but an ultrasonography expert with at least 20 years’ experience defined the lesion as presumably benign, the patient was included.

Patients were excluded if they were aged under 18 years or over 89 years, pregnant, or had pathologically confirmed ovarian cancer or normal ovary findings. Self-reported information and electronic health records were used to determine patient race and ethnicity.

A standardized protocol was used for most sonographic examinations, which occurred at the University of Chicago Department of Obstetrics and Gynecology. An ultrasonography researcher and expert sonogram examiner performed systematic reviews of the scans collected. The International Ovarian Tumor Analysis (IOTA) Simple Rules were used to evaluate lesions.

There were 511 female patients included in the analysis, 15.9% of which presented with malignant tumor. Patients with benign tumors were aged a mean 44.1 years, while those with malignant tumors were aged a mean 52.5 years. Among patients who received surgical evaluation, 75.2% had benign tumors, 4.4% borderline, and 19.4% malignant.

Of patients, 39.1% were postmenopausal, 44.4% were Black, 42.1% White, 9.4% other race, and 4.1% did not report race. When evaluating ethnicity, 89.2% were not Hispanic or Latino, 6.1% were Hispanic or Latino, and 4.7% did not report ethnicity.

Malignant lesions had a median diameter of 97 mm, compared to a median diameter of 49.5 mm for benign lesions. A significant increase in median maximal solid diameter was also seen in malignant lesions compared to benign lesions, at 59.5 mm vs 20.1 mm respectively.

Solid component presence was also greater in malignant lesions compared to benign lesions, at 93.8% vs 20.9% respectively. Other outcomes seen more often in malignant lesions was the presence of more than 3 papillary projections, more than 10 locules, and increased vascular scores on Doppler assessment.

The ADNEX model showed a 2.6% median risk of malignancy for benign lesions and a 71.8% median risk for malignant lesions. The area under the receiver operating characteristic (ROC) curve (AUC) for the ADNEX model when differentiating the 2 lesion types was 0.96.

At a cutoff of 10%, the ADNEX model displayed a sensitivity of 91.3%, specificity of 86.3%, positive predicting value (PPV) of 55.6%, and negative predictive value (NPV) of 98.1%. In comparison, the ROC AUC, sensitivity, specificity, PPV, and NPV of the O-RADS model were 0.92, 98.8%, 74.4%, 42.1%, and 99.7% respectively.

The ADNEX and O-RADS models both displayed strong efficacy in differentiating between benign and malignant lesions among a US cohort. This indicated enhanced care should the use of these models be expanded in the United States.

Reference

Yoeli-Bik R, Longman RE, Wroblewski K, Weigert M, Abramowicz JS, Lengyel E. Diagnostic performance of ultrasonography-based risk models in differentiating between benign and malignant ovarian tumors in a US cohort. JAMA Netw Open. 2023;6(7):e2323289. doi:10.1001/jamanetworkopen.2023.23289