Contrast-Enhanced Ultrasound Sharpens Diagnosis of Ovarian Tumors That Leave Experts Unsure

Some ovarian masses refuse to be sorted. When ultrasound images of an adnexal mass are fed through the International Ovarian Tumour Analysis simple rules, a widely used set of sonographic criteria for classifying ovarian tumors, a substantial fraction of lesions end up neither clearly benign nor clearly malignant but indeterminate, leaving clinicians and patients facing an anxious gray zone. A new exploratory pilot study published in BMC Women’s Health reports that contrast-enhanced ultrasound, a technique that uses injected microbubble contrast agents to visualize blood flow within a lesion in real time, may help resolve exactly this uncertainty.

The study, led by Ru-Yu Yan, Wei Chen, Xing-Pan Niu, Hong Xu and Xiao-Yu Fu, was a single-center retrospective pilot investigation of 90 patients whose adnexal masses had been classified as indeterminate by the IOTA simple rules and who were subsequently evaluated with contrast-enhanced ultrasound between April 2019 and May 2026. Because it was a retrospective analysis at one center, the findings are explicitly framed as exploratory, and the authors call for validation in large multicenter prospective trials before the approach becomes routine practice.

To establish ground truth, the researchers used postoperative histopathology as the reference standard. Among the 90 patients, 44 masses proved benign, 26 were borderline ovarian tumors, and 20 were malignant. This distribution highlights a familiar clinical problem: borderline ovarian tumors, which sit histologically between benign cysts and frank carcinomas, are notoriously difficult to distinguish from both neighbors on conventional ultrasound, yet their management differs substantially from that of either category.

The study focused on a specific sonographic clue called the microcystic pattern, a feature that has been associated with borderline tumors. On standard two-dimensional ultrasound, recognizing this pattern depends on subtle grayscale appearances that different observers may interpret differently. The investigators contrasted the diagnostic performance of the two-dimensional microcystic pattern with a contrast-enhanced ultrasound-derived microcystic pattern, in which the microbubble signal reveals the vascular architecture of the lesion’s small cystic spaces. Using the DeLong test, a statistical method for comparing the accuracy of two diagnostic markers measured on the same patients, they found that the contrast-enhanced version performed significantly better for detecting borderline ovarian tumors, with a P value of 0.024.

The second analytical strategy asked a broader question: does adding contrast-enhanced ultrasound features to conventional two-dimensional features improve the classification of these indeterminate masses across the board? For each of three binary comparisons, benign versus non-benign, borderline versus non-borderline, and malignant versus non-malignant, the team built nested Firth penalized logistic regression models, a form of regression suited to small samples where ordinary maximum-likelihood estimates can be unstable. The models were nested in the sense that the smaller model contained only two-dimensional sonographic features that showed significant differences between groups, while the larger model added the contrast-enhanced ultrasound features, allowing a direct comparison of the two on the same patients.

The results were consistent across all three outcomes. For distinguishing benign from non-benign masses, the area under the receiver operating characteristic curve, or AUC, rose from 0.664 with two-dimensional features alone to 0.849 when contrast-enhanced features were added, an increase of 0.185 that reached statistical significance at P = 0.0003. For borderline versus non-borderline classification, the AUC climbed from 0.727 to 0.846, a gain of 0.119 with P = 0.016. For malignant versus non-malignant classification, the AUC increased from 0.672 to 0.815, a gain of 0.143 with P = 0.015. In all three comparisons, the improvement in discrimination was statistically significant.

Two additional analyses reinforced the pattern. Net reclassification improvement and integrated discrimination improvement, both metrics that quantify how much a new marker reassigns patients into more accurate risk categories beyond what the area under the curve captures, were significantly positive and consistent across the three outcomes. Interobserver agreement, measured with the kappa statistic, was also higher for the contrast-enhanced microcystic pattern than for its two-dimensional counterpart, at 0.450 versus 0.345. Both values indicate only moderate agreement, but the contrast-enhanced approach was demonstrably more reproducible between observers, a meaningful advantage for a finding whose recognition on grayscale imaging has historically been subjective.

The clinical logic behind these results is straightforward. Conventional ultrasound depicts structure; contrast-enhanced ultrasound depicts perfusion. Borderline tumors often contain small cystic spaces within solid-appearing tissue, and the microbubble signal can delineate vascular patterns within those spaces that grayscale imaging cannot resolve. When the IOTA simple rules cannot classify a mass, that residual uncertainty frequently triggers either surgical intervention for lesions that might have been safely monitored or delayed intervention for lesions that warranted prompt treatment. A supplementary tool that measurably improves discrimination in exactly this indeterminate subgroup addresses a well-recognized gap in preoperative assessment.

The authors are careful about the limits of their evidence. With 90 patients from a single center and a retrospective design, the study is a proof of concept rather than a practice-changing trial. The Firth penalized models were used precisely because the sample size is modest, and the reported AUCs, while significantly improved, leave room for error in both directions. The authors conclude that contrast-enhanced ultrasound may act as a supplementary tool for IOTA-indeterminate cases, pending further validation in large multicenter prospective trials. Until such trials are completed, the study stands as an exploratory but statistically coherent signal that perfusion imaging adds real information to structural imaging in one of gynecologic ultrasound’s most stubborn diagnostic gray zones.

Subject of Research: Diagnostic accuracy of contrast-enhanced ultrasound for adnexal masses indeterminate on IOTA simple rules

Article Title: Contrast-enhanced ultrasound for indeterminate adnexal masses on IOTA simple rules: an exploratory single-center pilot study

Article References: Yan, R.-Y., Chen, W., Niu, X.-P., Xu, H., & Fu, X.-Y. (2026). Contrast-enhanced ultrasound for indeterminate adnexal masses on IOTA simple rules: an exploratory single-center pilot study. BMC Women's Health. https://doi.org/10.1186/s12905-026-05035-4

Image Credits: AI Generated

DOI: 10.1186/s12905-026-05035-4

Keywords: contrast-enhanced ultrasound, adnexal masses, IOTA simple rules, borderline ovarian tumors, microcystic pattern, diagnostic accuracy, ovarian tumors, retrospective pilot study, interobserver reproducibility, AUC, gynecologic imaging, Contrast-enhanced