Abstract
Background
Although the Assessment of Different Neoplasias in the adnexa (ADNEX) model has been widely validated, there remains challenges in differentiating true solid components from benign mimics. We aimed to evaluate the feasibility and value of substituting conventional ultrasound-derived solid component size with CEUS-derived measurements in the ADNEX model for adnexal lesions discrimination.
Methods
This multicenter prospective study recruited consecutive patients with suspected adnexal lesions with solid components detected at ultrasound between January 2022 and June 2024. All patients underwent preoperative CEUS examination. Pathological results serve as the reference standard. Each adnexal lesion was evaluated using ADNEX, a CEUS-based modified ADNEX score (CEUS-derived substitution for solid-component measurement without updating model coefficients) and Ovarian-Adnexal Reporting and Data System (O-RADS) Ultrasound (US) score.
Results
A total of 204 patients (median age, 47.0 years [IQR: 33.0-56.0 years]) with 204 lesions were included. The area under the receiver operating characteristic curve (AUROC) for CEUS-based modified ADNEX was 0.906 (95% CI: 0.867-0.944), which was superior to that of ADNEX (AUROC:0.865; 95% CI, 0.814-0.916; P = 0.019) and O-RADS US (AUROC: 0.783; 95% CI, 0.726 -0.840; P < 0.001) in classifying adnexal lesions with solid components. CEUS-based modified ADNEX demonstrated greater specificity (0.664) compared to ADNEX (0.283; P < 0.001) and O-RADS US (0.221; P < 0.001).
Conclusions
CEUS-derived modification for solid-component measurement in the ADNEX model showed potential improvements in the diagnostic specificity and clinical benefit compared with the original ADNEX framework in the classification of adnexal lesions with solid components.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12905-026-04700-y.
Keywords: Adnexal Lesion, Contrast-enhanced Ultrasound, Malignancy
Introduction
Ovarian cancer remains a major threat to women’s health due to its high mortality rate. It still represents the deadliest malignancy in the female reproductive system, with its five-year survival rate below 50% [1, 2]. The clinical management for benign and malignant adnexal lesions differs greatly [3]. Consequently, the accurate characterization of newly detected adnexal lesions is essential for guiding optimal therapeutic decision-making for patients.
Transvaginal ultrasound is widely recommended as the primary investigation for adnexal lesions, with a high sensitivity in ruling out malignancy [3, 4]. Subjective evaluation by ultrasound experts using transvaginal ultrasound provides the highest accuracy for the assessment of adnexal malignancy [5]. To standardize evaluation of adnexal lesions for less-experienced examiners, several ultrasound-based risk prediction models have been developed for guiding treatment. Among these, the International Ovarian Tumor Analysis (IOTA) [6] Assessment of Different Neoplasias in the adnexa (ADNEX) model is the most thoroughly validated tool for assessing adnexal lesions [7–12]. In a recent nationwide prospective multicenter study, namely IOTA phase 6, it has been demonstrated that both the ADNEX model and the IOTA 2-step strategy discriminated well between benign and malignant adnexal lesions in the hands of ultrasound examiners with differing levels of skill [5].
The largest diameter of the solid components is one of the key parameters for risk stratification of the ADNEX model [7]. Nevertheless, it should be noted that in the real clinical practice, there remains diagnostic challenges in differentiating true solid components from benign mimics (e.g., organized hemorrhagic content or dense fibrosis) on grayscale ultrasound. This is more challenging and difficult for examiners with limited expertise [13]. The misclassification of such “pseudo-solid” components may result in an overestimation of malignancy risk by the ADNEX model, potentially leading to unnecessary surgical interventions.
Color Doppler evaluation serves as one of the effective tools for distinguishing between avascular “pseudo-solid” components and truly vascularized solid tissue. Nevertheless, qualitative color Doppler assessment is not included in the ADNEX model, primarily because the technique is unable to show blood flow at the perfusion level [14]. This limitation makes it suboptimal for evaluating vascular information in small nodules and solid components, which are frequent findings among adnexal lesions at ultrasound. In contrast, contrast-enhanced ultrasound (CEUS) represents a promising complementary imaging modality in evaluating microvascular blood flow at perfusion-level hemodynamics in adnexal lesions [14, 15]. Evidence shows that the integration of CEUS into the Ovarian-Adnexal Reporting and Data System (O-RADS) system greatly improved the diagnostic accuracy and specificity in characterizing adnexal lesions [16–18]. Yet, the impact of substituting conventional ultrasound-derived solid-component size with CEUS findings within the ADNEX model has not been evaluated. To overcome the diagnostic challenges regarding uncertain “pseudo-solid” components in conventional ultrasound, we proposed a CEUS-based modified ADNEX score (CEUS-derived substitution for solid-component measurement without updating model coefficients of the ADNEX model). Therefore, the aim of this study is to prospectively evaluate the feasibility and value of substituting conventional ultrasound-derived solid component size with CEUS-derived measurements in the ADNEX model for adnexal lesions discrimination, in comparisons with the original ADNEX model and O-RADS ultrasound (US) system.
Methods
Study design and participants
This multicenter prospective study was conducted at 10 tertiary hospitals in China between January 2022 and June 2024. The study recruited consecutive patients with suspected adnexal masses with solid components detected at ultrasound. The inclusion criteria were as follows: (1) patient’s age > 18 years, (2) patient underwent preoperative CEUS examinations before surgery, and (3) surgery performed within 120 days after the examinations. The exclusion criteria were as follows: (1) patients who were pregnant, (2) patients with previous bilateral adnexectomy, and (3) patients with previous history of ovarian cancer.
The protocol was approved by the Ethical Committee of the Third Affiliated Hospital of Sun-Yat-sen University (approval number [2021]02–113-01). The study was conducted in compliance with the Declaration of Helsinki. Written informed consent was obtained from all patients.
In the current study, only one adnexal lesion per patient was selected for analysis. In patients with bilateral adnexal lesions, the lesion showing the most complex ultrasound morphology was included. In case where bilateral lesions showing similar morphology, the larger lesion was chosen for analysis.
Data collection and reference standard
Information on patients’ age, serum cancer antigen 125 (CA125), menopausal status and pathological diagnosis were obtained from the hospital database. Postmenopausal status was defined as (1) aged over 50 years with a previous history of hysterectomy or (2) experienced at least 12 consecutive months of amenorrhea [11].
The postoperative pathologic diagnosis within 120 days after the ultrasound examination served as the reference standard to confirm all adnexal lesions. Borderline tumors were classified as malignant lesion.
Imaging procedures
All patients underwent a standardized preoperative imaging protocol including unenhanced pelvic ultrasonography and CEUS examinations. All scans were performed primarily transvaginally (or transrectally in case patient was not sexually active) by experienced radiologists with at least five years of experience in gynecological ultrasound at their institution. The performers had no access to patient’s pathological results. Transabdominal ultrasound examination was conducted if lesions exceeded 6 cm.
CEUS scans were performed for adnexal lesions with solid components, using the contrast agent SonoVue (Bracco), with 1.5 mL (for transabdominal route) or 2.4 mL (for transvaginal route) of contrast agent intravenously injected, immediately flushed with 5 mL of saline. A 120-s CEUS video clip recording was started at the moment of contrast injection.
Examinations were performed using the following ultrasound systems: the Philips EPIQ 7 (using C10-3 V transvaginal and C5-1 transabdominal transducers) and the Mindray Nuewa R9 (using V11-3HU transvaginal and SC6-1U transabdominal transducers).
Image evaluations
All images were reviewed by two independent senior radiologists (M.Z. and E.Q.), both possessing more than 10 years of experience in gynecological ultrasound. The two reviewers were blinded to patients’ clinical and pathological information, as well as other imaging results. Ultrasound images were evaluated according to IOTA terminology and the 2022 updated version of the O-RADS US lexicon. The presence of enhancement of solid components on CEUS was recorded. Solid components were defined as solid tissue ≥ 3 mm in height.
The discrepancies between two reviewers were resolved through a discussion, wherein they re-assessed the videos and images to reach a consensus.
ADNEX model with CA-125
The IOTA-ADNEX prediction model was applied to assess adnexal malignancy. Its algorithm incorporates nine input variables: age (years), oncology center referral (binary), maximal diameter of the lesion (mm), maximal diameter of the largest solid components (mm), presence of > 10 locules (binary), number of papillary projections (categorical), presence of acoustic shadows (binary), presence of ascites (binary) and CA125 (U/mL, optional) [7]. These inputs were entered into an integrated software to yield a malignancy probability score for each lesion. A computed risk score of 10% or higher was classified as malignancy.
CEUS-based modified ADNEX score
A CEUS-based modified ADNEX score was proposed in the study, which used CEUS-derived substitution for solid-component measurement without updating model coefficients of the ADNEX model. As detailed in Fig. 1, in case solid components showed no enhancement on CEUS, it was considered as “pseudo-solid” components in conventional ultrasound. Therefore, the “maximal diameter of the largest solid part” of the ADNEX model should be adjusted as 0 mm. If enhancement of solid components was observed, the measured maximum diameter of the enhanced area on CEUS was input. The adjusted malignancy risk was calculated a CEUS-based modified ADNEX score.
Fig. 1.
Illustration of the flowchart of contrast-enhanced ultrasound (CEUS)-derived substitution for solid-component measurement in the Assessment of Different Neoplasias in the adnexa (ADNEX) model
O-RADS US evaluation
Lesion risk was stratified using the American College of Radiology O-RADS US framework [19–23]. The O-RADS US system assigns a score from 1 to 5 [24–26], correlating with different malignancy risk for guiding management. The relevant risk categories were defined as follows: O-RADS 2 (almost certainly benign lesions with estimated malignancy risk < 1%.), O-RADS 3 (low-risk lesions with estimated malignancy risk 1% to less than 10%), O-RADS 4 (intermediate-risk lesions with estimated malignancy risk 10% to less than 50%), and O-RADS 5 (high-risk lesions with estimated malignancy risk ≥ 50%). O-RADS 4 and higher was considered as malignancy.
Sample size estimation and statistical analysis
The sample size evaluation was estimated for a paired comparison of two diagnostic specificities using the McNemar test. The estimation targeted 90% power (β = 0.10) with a significance level (α) of 0.05. We assumed that the specificities of ADNEX and CEUS-based modified ADNEX score were 0.70 and 0.80, respectively. The within-subject correlation between paired comparison was assumed to be 0.60. Under these assumptions, a minimum of 164 lesions was needed. The sample size calculation was assessed using PASS (version 21.0.3).
Baseline statistics were presented as median (IQR) or n (%). Categorical variables were compared with the χ2 test and continuous variables were analyzed using the Mann–Whitney U test. The discriminatory ability of different models to discriminate between benign and malignant adnexal lesions was evaluating by calculating the area under the receiver operating characteristic curve (AUROC). Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV), along with their 95% confidence intervals (CIs), were also reported. Pairwise comparisons of sensitivity and specificity were evaluated with McNemar’s test, while differences in AUROCs were compared with DeLong’s statistics. The interobserver agreement between the two radiologists were assessed using weighted kappa coefficient for categorical variables and intraclass correlation coefficient (ICC) for continuous measurements. Post-hoc power analysis was performed to assess whether our sample size provided sufficient statistical power to detect a clinically meaningful difference in specificity. Decision curve analysis (DCA) was conducted to assess the clinical utility of different ultrasound models. A P value below 0.05 was considered statistical significance. All statistical analyses were conducted using R software (version 4.1.2).
Results
During the period between January 2022 and June 2024, a total of 228 patients with 228 adnexal lesions containing solid components were initially recruited (Fig. 2). Five patients were younger than 18 years, four patients had previous history of ovarian cancer, and 15 patients underwent surgery more than 120 days after their ultrasound examinations. After excluding 24 patients, the remaining 204 patients (median age, 47.0 years [IQR: 33.0—56.0 years]) with 204 lesions were finally included for statistical analysis. Baseline characteristics of the study population and lesions are illustrated in Table 1. Most of the lesions were classified as O-RADS 4 or 5 (86.8%, 177/204).
Fig. 2.
Flow chart for study inclusion
Table 1.
Baseline characteristics of the study population (n = 204)
| Parameters | Total (n = 204) | Benign (n = 113) | Malignant (n = 91) | P value* |
|---|---|---|---|---|
| Patient age (y) | 47.0 (33.0, 56.0) | 37.0 (29.0, 54.0) | 51.0 (42.0, 59.0) | < 0.001 |
| Menopausal status | < 0.001 | |||
| Premenopausal | 124 (61) | 81 (72) | 43 (47) | |
| Postmenopausal | 80 (39) | 32 (28) | 48 (53) | |
| CA125 (U/mL) | 34.0 (20.0, 137.7) | 21.0 (13.3, 34.0) | 118.3 (34.0, 589.0) | < 0.001 |
| Maximum diameter of lesion (mm) | 78.5 (55.0, 115.0) | 68.8 (49.9, 97.0) | 98.0 (67.0, 130.0) | < 0.001 |
| Largest diameter of largest solid component (mm) | 39.5 (25.0, 69.5) | 32.0 (21.0, 51.0) | 53.0 (30.0, 89.0) | < 0.001 |
| More than 10 cyst locules | 15 (7) | 4 (4) | 11 (12) | 0.020 |
| Ascites | 40 (20) | 7 (6) | 33 (36) | < 0.001 |
| Acoustic shadows | 25 (12) | 23 (20) | 2 (2) | < 0.001 |
| Peritoneal thickening/nodules | 12 (6) | 0 (0) | 12 (13) | < 0.001 |
| Number of papillary projections | 0.085 | |||
| 0 | 157 (77) | 90 (80) | 67 (74) | |
| 1–3 | 28 (14) | 17 (15) | 11 (12) | |
| > 3 | 19 (9) | 6 (5) | 13 (14) | |
| Color score | < 0.001 | |||
| 1 | 46 (23) | 42 (37) | 4 (4) | |
| 2 | 90 (44) | 56 (50) | 34 (37) | |
| 3 | 51 (25) | 14 (12) | 37 (41) | |
| 4 | 17 (8) | 1 (1) | 16 (18) | |
| O-RADS US score | < 0.001 | |||
| 3 | 27 (13) | 25 (22) | 2 (2) | |
| 4 | 98 (48) | 70 (62) | 28 (31) | |
| 5 | 79 (39) | 18 (16) | 61 (67) |
Data are given as median (interquartile range) or n (%)
*P values were compared between benign and malignant adnexal lesions. Categorical variables were compared with the χ2 test and continuous variables were analyzed using the Mann–Whitney U test
Based on pathology, 113 lesions were benign (55.4%, 113/204), and 91 were malignant (44.6%, 91/204). Table 2 summarizes the specific histological diagnoses of the included lesions. Among the malignant lesions, there were 24 borderline tumors, 15 stage I tumors, 48 stage II-IV tumors and 4 metastatic tumors.
Table 2.
Specific histological diagnoses of the included lesions (n = 204)
| Histological diagnoses | n |
|---|---|
| Benign | 113 |
| Mature teratoma | 24 |
| Ovarian theca-fibroma | 22 |
| Cystadenoma (serous, mucinous, seromucinous) | 20 |
| Endometrioma | 16 |
| Abscess | 10 |
| Cystadenofibroma | 5 |
| Pelvic inflammatory disease | 3 |
| Simple cyst | 3 |
| Follicular cyst | 2 |
| Corpus luteum | 1 |
| Brenner tumor | 1 |
| Other benign tumors | 6 |
| Malignant | 91 |
| Serous borderline tumor | 12 |
| Mucinous borderline tumor | 9 |
| Seromucinous borderline tumor | 1 |
| Other borderline tumors | 2 |
| Serous cystadenocarcinoma | 36 |
| Adenocarcinoma | 6 |
| Clear cell tumor | 6 |
| Metastatic cancer | 4 |
| Germ cell tumor | 3 |
| Endometrioid carcinoma | 2 |
| Mucicous cystadenocarcinoma | 2 |
| Other malignant tumors | 8 |
Data are given as n
No adverse events regarding CEUS were observed in the current study.
Reproducibility analysis
There was good inter-observer agreement between the two radiologists for O-RADS US category, with a weighted kappa value of 0.693 (95% CI: 0.599−0.788). In addition, the inter-observer reliability between the two radiologists for continuous measurements (the largest part of solid tissue) was excellent, with intraclass correlation coefficients (ICCs) of 0.992 (95% CI: 0.990−0.994) and 0.991 (95% CI: 0.988−0.993) for gray-scale ultrasound and CEUS, respectively.
Performance of ADNEX, CEUS-based modified ADNEX score and O-RADS US
Figure 3 presents the receiver operating characteristic (ROC) curves of different models for malignancy prediction. As presented in Table 3, The overall AUROC for CEUS-based modified ADNEX score was 0.906 (95% CI: 0.867−0.944), which was noticeably higher than that of ADNEX (AUROC:0.865; 95% CI, 0.814−0.916; P = 0.019) and O-RADS US (AUROC: 0.783; 95% CI, 0.726−0.840; P < 0.001). ADNEX showed significant higher AUROC than O-RADS US (P = 0.003).
Fig. 3.
Receiver operating characteristic (ROC) curves of different models for malignancy prediction. CEUS, contrast-enhanced ultrasound; ADNEX, Assessment of Different Neoplasias in the adnexa; O-RADS, Ovarian-Adnexal Reporting and Data System; US, ultrasound
Table 3.
Diagnostic performance of Assessment of Different Neoplasias in the adnexa (ADNEX), contrast-enhanced ultrasound (CEUS)-based modified ADNEX score and Ovarian-Adnexal Reporting and Data System (O-RADS) ultrasound (US) for adnexal malignancy assessment in lesion with solid components
| Model | ADNEX | CEUS-based modified ADNEX score | O-RADS US |
|---|---|---|---|
| Cut-off | 10% | 10% | Category 4 |
| AUROC (95% CI)* | 0.865 (0.814−0.916) | 0.906 (0.867−0.944) | 0.783 (0.726−0.840) |
| Sensitivity (95% CI)† | 0.967 (0.908−0.989) | 0.956 (0.891−0.988) | 0.978 (0.923−0.994) |
| Specificity (95% CI)‡ | 0.283 (0.208−0.372) | 0.664 (0.569−0.750) | 0.221 (0.155−0.306) |
| PPV (95% CI) | 0.521 (0.446−0.595) | 0.696 (0.607−0.775) | 0.503 (0.430−0.576) |
| NPV (95% CI) | 0.914 (0.776−0.970) | 0.949 (0.875−0.986) | 0.926 (0.766−0.979) |
*CEUS-ADNEX presented significant higher AUC compared with ADNEX (P = 0.019) and O-RADS US (P < 0.001) for adnexal malignancy assessment in lesion with solid components. ADNEX showed significant higher AUC than O-RADS US (P = 0.003)
†P > 0.05 for all sensitivity comparisons
‡CEUS-ADNEX presented significant higher specificity than ADNEX (P < 0.001) and O-RADS US (P < 0.001) for adnexal malignancy assessment in lesion with solid components. There was no statistical difference of specificity between ADNEX and O-RADS US (P = 0.248). AUROC, area under the receiver-operating-characteristic curve; PPV, positive predictive value; NPV negative positive value
ADNEX at a cut-off of 10%, CEUS-based modified ADNEX score at a cut-off of 10% and O-RADS US at a cut-off of category 4 all yielded high sensitivities (0.967 vs. 0.956 vs. 0.978) and NPVs (0.914 vs. 0.949 vs. 0.926) in assessing adnexal malignancy in lesions with solid components. In the contrary, CEUS-based modified ADNEX score at a cut-off of 10% showed greater specificity (0.664) compared to ADNEX at a cut-off of 10% (0.283; P < 0.001) and O-RADS US at a cut-off of category 4 (0.221; P < 0.001). There was no statistical difference of specificity between ADNEX and O-RADS US (P = 0.248).
Figure 4 and 5 demonstrate how CEUS-based modified ADNEX score helps radiologists in reclassifying adnexal lesions with solid components. As illustrated in Fig. 4, routine ultrasonography examination in a 31-year-old woman revealed a unilocular cyst with ground-glass echogenicity containing one papillary projection. The maximum diameter of the lesion was 54 mm and the largest diameter of the papillary projection was 11 mm. There was minimal Color Doppler signal within the papillary projection using microvascular imaging. On CEUS, however, the papillary projection demonstrated no internal enhancement, which indicated it might be a typical endometrioma. The CEUS-based modified ADNEX risk score was calculated as 2.5%, and pathology later confirmed an endometrioma. Conversely, Fig. 5 describes a radiographically similar case in a 47-year-old woman with a 56 mm unilocular cyst containing one 16 mm papillary projection on routine ultrasonography examination. While Color Doppler shows minimal signal within the papillary projection, CEUS revealed internal enhancement within the projection. The CEUS-based modified ADNEX risk score was 12.8%, correlating with the final pathological diagnosis of serous borderline tumor.
Fig. 4.
Example of how contrast-enhanced ultrasound (CEUS) refine adnexal lesion with non-enhancing solid components assessing by Assessment of Different Neoplasias in the adnexa (ADNEX) model. A 31-year-old woman revealed a unilocular cyst with ground-glass echogenicity containing one papillary projection. The maximum diameter of the lesion was 54 mm and the largest diameter of the papillary projection was 11 mm. There was minimal Color Doppler signal within the papillary projection using microvascular imaging. On CEUS, however, the papillary projection demonstrated no internal enhancement, which indicated it might be a typical endometrioma. The adjusted ADNEX model risk was calculated as 2.5%, and pathology later confirmed an endometrioma
Fig. 5.
Example of how contrast-enhanced ultrasound (CEUS) refine adnexal lesion with enhancing solid components assessing by Assessment of Different Neoplasias in the adnexa (ADNEX) model. A 47-year-old woman with a 56 mm unilocular cyst containing one 16 mm papillary projection on routine ultrasonography examination. While Color Doppler shows minimal signal within the papillary projection, CEUS revealed internal enhancement within the projection. The largest diameter of the papillary projection remained unchanged. The corresponding ADNEX model risk was 12.8%, correlating with the final pathological diagnosis of serous borderline tumor
Performance of ADNEX and CEUS-based modified ADNEX score for discriminating borderline tumors from invasive malignant lesions
As detailed in Supplementary Table 1, both ADNEX and CEUS-based modified ADNEX score showed similar diagnostic performance in discriminating borderline tumors from invasive malignant lesions (AUC: 0.765 vs. 0.782; P = 0.521). CEUS-based modified ADNEX score showed higher specificity (0.785) in discriminating borderline tumors from invasive malignant lesions compared with ADNEX model (0.723, P = 0.046).
Sensitivity analysis excluding CA125
Supplementary Table 2 compares the diagnostic performance of ADNEX without CA125, CEUS-based modified ADNEX score without CA125 for adnexal malignancy assessment in lesion with solid components. When excluding CA125, CEUS-based modified ADNEX score significantly improved the diagnostic performance and specificity of the original ADNEX model, with the AUC rising from 0.830 to 0.889 and the specificity increasing from 0.292 to 0.628.
Histological diagnosis of lesions with “pseudo-solid” components showing no internal enhancement on CEUS
Of the 204 adnexal lesions included in analysis, there were 50 lesions showing no internal enhancement within the solid components on CEUS examinations. Table 4 details the specific histological diagnosis of lesions with “pseudo-solid” components. The most prevalent diagnoses were cystadenomas (28.0%, 14/50), mature teratomas (26.0%, 13/50), and endometriomas (22.0%, 11/50). The remaining diagnosis included various other benign lesions, such as inflammatory lesions and functional cysts.
Table 4.
Histological diagnosis of lesions with “pseudo-solid” components showing no internal enhancement on contrast-enhanced ultrasound (n = 50)
| Histological diagnoses | n |
|---|---|
| Benign | 50 |
| Mature teratoma | 13 |
| Cystadenoma (serous, mucinous, seromucinous) | 14 |
| Endometrioma | 11 |
| Abscess | 3 |
| Cystadenofibroma | 2 |
| Pelvic inflammatory disease | 1 |
| Simple cyst | 3 |
| Follicular cyst | 2 |
| Corpus luteum | 1 |
Data are given as n
Clinical utility
Figure 6 shows the DCA analysis comparing the clinical net benefit of different ultrasound models. Across the malignancy risk thresholds from 10 to 50%, the CEUS-based modified ADNEX score achieved the highest clinical net benefit, superior to that of the ADNEX model and O-RADS US, which indicates its advantage in clinical decision-making for ovarian tumor classification. The O-RADS US achieved the lowest net benefit of the three models.
Fig. 6.
Decision-curve analysis showing net benefits of three ultrasound-based risk models in characterizing adnexal lesions with solid components. CEUS, contrast-enhanced ultrasound; ADNEX, Assessment of Different Neoplasias in the adnexa; O-RADS, Ovarian-Adnexal Reporting and Data System
Post-hoc power analysis
A post-hoc power analysis was performed using PASS (version 21.0.3). The target significance level is 0.05. The actual significance level achieved by the specificity test is 0.0368. The prevalence of the disease is 44.6%. A total sample size of 204 (which includes 91 malignant lesions) achieves 100% power to detect a change in specificity from 0.283 to 0.664 using a two-sided binomial test.
Discussion
In this study, we demonstrated that CEUS-derived modification for solid-component measurement in the ADNEX model showed potential improvements in diagnostic performance and specificity for characterizing adnexal lesions with solid components, and helped triage adnexal lesions to appropriate workup, avoiding unnecessary surgery for benign lesions. The established CEUS-based modified ADNEX score was superior to IOTA-ADNEX and O-RADS US in classifying adnexal lesions with higher clinical net benefit. CEUS-derived modification for solid-component measurement in the ADNEX framework holds potential value in effectively reclassifying those lesions with non-enhancing solid appearing components.
MRI has been recommended as a secondary modality for classifying sonographically indeterminate adnexal lesions [27], particularly those lesions containing solid components, offering superior diagnostic specificity than conventional ultrasound [28–30]. However, in Asian areas where clinical resources are limited, MRI is not universally accessible. CEUS provides an optional tool to allow ultrasound to characterize perfusion-level vascularity of adnexal lesions. Previous studies have confirmed that integrating CEUS into O-RADS US markedly improved diagnostic specificity and overall accuracy for adnexal lesions [31, 32]. The results of our study are in line with previous research. Compared to MRI, CEUS is more cost-effective, allows real-time evaluation and has rare contrast-related adverse events. Therefore, incorporating CEUS into routine ultrasound practice holds considerable potential for improving the classification of adnexal lesions, especially in settings where MRI is not readily accessible.
The IOTA-ADNEX model has been widely validated as an effective tool in discriminating benign from malignant adnexal lesions, with AUROCs ranging from 0.91 to 0.96 [5, 8]. However, as observed in the current study, the IOTA-ADNEX model achieved an AUROC of 0.865 in characterizing lesions with solid components, which is somewhat lower than previous findings. This difference might be attributed to the following reasons. First, our cohort specifically focused on adnexal lesions with solid components, which are associated with a higher malignant potential and ultrasonographic complexity. Secondly, a considerable proportion of included lesions of our study were atypical endometriomas or mature teratomas. At ultrasound, these lesions sometimes can be challenging because of irregular walls or suspected solid components, which can mimic adnexal malignancy. This might lead to the overestimation of risk of the IOTA-ADNEX model, resulting in an increased false positive rate and thus reducing its overall diagnostic accuracy.
The EFSUMB guidelines for the clinical practice of CEUS in non-hepatic applications suggest that adnexal lesions without internal enhancement are invariably benign, whereas the presence of enhancement is not predictive of malignancy [33]. Yet malignant lesions may occasionally present extensive necrosis, mucinous degeneration or intratumoral hemorrhage. However, most of necrosis, mucinous degeneration or intratumoral hemorrhage in malignant lesions are incomplete, which suggests that enhancement may be detectable in the peripheral areas of the malignant lesions. In our study, 50 non-enhancing lesions were confirmed benign on histopathology. Although our results are in line with established guidelines, the possibility of false-negative results cannot be entirely excluded due to the limited sample size. Importantly, radiologists should be cautious that a non-enhancing lesion may represent a malignant tumor with extensive necrosis mucinous degeneration or intratumoral hemorrhage in clinical practice.
In our study, there were four malignant cases that were missed by the CEUS-based ADNEX score. Two of these cases were borderline tumors. The primary reason for misclassification was that the largest part of the enhancing solid tissue of theses lesions was relatively small, resulting in a low ADNEX risk score.
While CEUS provides superior diagnostic accuracy compared to conventional ultrasound, it is not considered as a first-line tool for adnexal malignancy, mainly due to its higher cost and limited availability. Presently, there is no international consensus on the scope of application for ultrasound contrast agents in adnexal lesions. Inappropriate application of CEUS for typical benign lesions may lead to overtreatment, increased costs, and potential patient risk. Similar to MRI, CEUS is most advantageous in further evaluating sonographically indeterminate adnexal lesions, especially those with solid components [14]. Future prospective studies are warranted to define the standardized referral criteria of the use of CEUS in clinical settings.
This study has several limitations that should be acknowledged. First, the current study showed a relatively high malignancy rate of 44.6%. This may be driven by the fact that the study relied on surgical pathology as the sole reference standard for diagnosis and excluded adnexal lesions with adequate follow-up referred for conservative management, which might introduce selection bias in the study. The reported specificity and positive predictive value may not accurately reflect real-world clinical performance in a broader screening or routine diagnostic population where conservative management is common. Second, we failed to perform further analysis for discrimination between subgroups of tumor outcomes (benign vs borderline vs stage Ⅰ vs stage Ⅱ-Ⅳ vs metastasis to the adnexa) due to limited sample size. Third, the current study did not incorporate contrast enhanced features, such as the timing or pattern of enhancement into the ADNEX model, which is valuable for malignancy discrimination. Finally, we did not compare the discriminatory ability of the proposed CEUS-based modified O-RADS score with the established MRI-based models, such as O-RADS MRI or ADNEX-MRI. Future larger, multicenter studies are needed to validate our findings, and perform comparative analysis between CEUS and MRI-based frameworks.
Conclusions
In conclusion, our findings revealed that CEUS-derived modification for solid-component measurement in the ADNEX model markedly improved the diagnostic specificity and clinical net benefit of conventional-US based models in the classification of adnexal lesions with solid components. CEUS-derived modification for solid-component measurement in the ADNEX model might serve as a valuable secondary method in further evaluating ultrasound-indeterminate adnexal lesions, which offers a practical and cost-effective alternative to MRI in resource-limited settings. Further work is warranted to define the referral criteria of the use of CEUS in gynecology and conduct head-to-head comparisons between CEUS and established MRI risk stratification systems.
Supplementary Information
Acknowledgements
We would like to acknowledge the patients and staff in the involved hospitals.
Patient consent statement
All participants provided written informed consent.
Abbreviations
- ADNEX
Assessment of Different Neoplasias in the adnexa
- CEUS
Contrast-enhanced ultrasound
- IOTA
International Ovarian Tumor Analysis
- CA125
Serum cancer antigen 125
- O-RADS
Ovarian-Adnexal Reporting and Data System
- US
Ultrasound
- AUROC
Area under the receiver operating characteristic curve
- PPV
Positive predictive value
- NPV
Negative predictive value
- CI
Confidence interval
- DCA
Decision curve analysis
Authors’ contributions
Manli Wu, Ying Chen and ShuangFVyu Wu conceptualized the study, performed the statistical analysis and drafted the manuscript. Ruili Wang, Xiaotao Meng, Hong Wen, Rui Zhang, Liang Mu, Li Xiao, Tingting Liu, Sijia Chen, Zeping Huang, Junlin Zhong, Yatong Chen, Ying Wang, Man Zhang, Licong Huang and Jingjing Liu collected and curated the clinical data. Xiaofeng Sun and Xinling Zhang revised the manuscript. All authors contributed to manuscript revision, read, and approved the final version.
Funding
The study was supported by grants from the Guangzhou Basic and Applied Basic Research Foundation (no. 2024A04J4786), Science and Technology Projects of Guangzhou (no.2023A03J0217), GuangDong Basic and Applied Basic Research Foundation (no.2023A1515220008), Guangzhou Medical Research Fund Project (2026KL-A1001), the National Natural Science Foundation of China (no.82402296); Guangdong Province Hospital Association Ultrasound Medicine Research Special Fund (no.2023CSM004); Guangzhou Basic and Applied Basic Research Scheme (no.2025A03J3213). The funders had no influence on study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Declarations
Ethics approval and consent to participate
The protocol was approved by the Ethical Committee of the Third Affiliated Hospital of Sun-Yat-sen University (approval number [2021]02-113-01).The study was conducted in compliance with the Declaration of Helsinki. Written informed consent was obtained from all patients.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Manli Wu, Ying Chen and Shuangyu Wu contributed equally to this work.
Xiaofeng Sun and Xinling Zhang are co–senior authors.
Contributor Information
Xiaofeng Sun, Email: sxf7251@zxhospital.com.
Xinling Zhang, Email: zhxinl@mail.sysu.edu.cn.
References
- 1.Siegel RL, Giaquinto AN, Jemal A. Cancer statistics, 2024. CA Cancer J Clin. 2024;74(1):12–49. [DOI] [PubMed] [Google Scholar]
- 2.Caruso G, Weroha SJ, Cliby W. Ovarian cancer: a review. JAMA. 2025;334(14):1278–91. [DOI] [PubMed] [Google Scholar]
- 3.Timmerman D, Planchamp F, Bourne T, Landolfo C, du Bois A, Chiva L, et al. ESGO/ISUOG/IOTA/ESGE consensus statement on preoperative diagnosis of ovarian tumors. Ultrasound Obstet Gynecol. 2021;58(1):148–68. [DOI] [PubMed] [Google Scholar]
- 4.Wu M, Wu J, Wang Y, Zhang W, Zhu J, You L, et al. Ultrasound-trained artificial intelligence for postmenopausal adnexal lesion characterization: a multicenter diagnostic accuracy study. BIO Integration. 2025;6(1):1–12. [Google Scholar]
- 5.Moro F, Momi M, Ledger A, Barreñada L, Ceusters J, Sturla D, et al. External validation of ultrasound-based models for differentiating between benign and malignant adnexal masses: a nationwide prospective multicenter study (IOTA phase 6). Am J Obstet Gynecol. 2025;233(6):629.e621-629.e613. [DOI] [PubMed] [Google Scholar]
- 6.Tian C, Wen SB, Zhao CY, Yan XN, Du JX. Comparative diagnostic accuracy of the IOTA SRR and LR2 scoring systems for discriminating between malignant and benign adnexal masses by junior physicians in Chinese patients: a retrospective observational study. BMC Womens Health. 2023;23(1):585. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Van Calster B, Van Hoorde K, Valentin L, Testa AC, Fischerova D, Van Holsbeke C, et al. Evaluating the risk of ovarian cancer before surgery using the ADNEX model to differentiate between benign, borderline, early and advanced stage invasive, and secondary metastatic tumours: prospective multicentre diagnostic study. BMJ. 2014;349:g5920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Barreñada L, Ledger A, Dhiman P, Collins G, Wynants L, Verbakel JY, et al. ADNEX risk prediction model for diagnosis of ovarian cancer: systematic review and meta-analysis of external validation studies. BMJ Med. 2024;3(1):e000817. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Geysels A, Garofalo G, Timmerman S, Ceusters J, Buonomo F, Fischerová D, et al. ADNEX-AI: automated extraction of ultrasound predictors for interpretable ovarian cancer risk stratification. NPJ Precis Oncol. 2025;10(1):18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Landolfo C, Bourne T, Froyman W, Van Calster B, Ceusters J, Testa AC, et al. Benign descriptors and ADNEX in two-step strategy to estimate risk of malignancy in ovarian tumors: retrospective validation in IOTA5 multicenter cohort. Ultrasound Obstet Gynecol. 2023;61(2):231–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Meys EMJ, Jeelof LS, Achten NMJ, Slangen BFM, Lambrechts S, Kruitwagen R, et al. Estimating risk of malignancy in adnexal masses: external validation of the ADNEX model and comparison with other frequently used ultrasound methods. Ultrasound Obstet Gynecol. 2017;49(6):784–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rodríguez-Rubio C, Vegas-Viedma S, Del Olmo-Reillo M, Quintana-Zapata P, Sancho-Sauco J, Pablos-Antona MJ, et al. ECO-SCORE: development of a new ultrasound score for the study of cystic and solid-cystic adnexal masses based on imaging characteristics. Biomedicines. 2025;13(2):317. [DOI] [PMC free article] [PubMed]
- 13.Patel-Lippmann KK, Gupta A, Martin MF, Phillips CH, Maturen KE, Jha P, et al. The roles of Ovarian-Adnexal Reporting and Data System US and Ovarian-Adnexal Reporting and Data System MRI in the evaluation of adnexal lesions. Radiology. 2024;312(2):e233332. [DOI] [PubMed] [Google Scholar]
- 14.Reid M, Brisebois K, Merrill C, Wilson SR. Contrast-enhanced ultrasound of the ovary technique and lexicon recommendations: technique and lexicon recommendations. J Ultrasound Med. 2026;45:653–67. [DOI] [PubMed]
- 15.Shi Y, Li H, Wu X, Li X, Yang M. O-RADS combined with contrast-enhanced ultrasound in risk stratification of adnexal masses. J Ovarian Res. 2023;16(1):153. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wu M, Yang H, Chen Y, Wu S, Liang T, Zhang M, et al. Adnexal lesion discrimination using deep learning analysis of dynamic contrast-enhanced US images. Radiol Artif Intell. 2026;8(1):e240786. [DOI] [PubMed] [Google Scholar]
- 17.Lu B, Liu C, Qi J, He W, Shi T, Zhu Y, et al. Comparison of contrast-enhanced ultrasound, IOTA simple rules and O-RADS for assessing the malignant risk of sonographically appearing solid ovarian masses. J Gynecol Obstet Hum Reprod. 2023;52(4):102564. [DOI] [PubMed] [Google Scholar]
- 18.Shentu W, Zhang Y, Gu J, Wang F, Zhao W, Liu C, et al. Contrast-enhanced ultrasonography for differential diagnosis of adnexal masses. Front Oncol. 2022;12:968759. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Andreotti RF, Timmerman D, Strachowski LM, Froyman W, Benacerraf BR, Bennett GL, et al. O-RADS US risk stratification and management system: a consensus guideline from the ACR Ovarian-Adnexal Reporting and Data System Committee. Radiology. 2020;294(1):168–85. [DOI] [PubMed] [Google Scholar]
- 20.Strachowski LM, Jha P, Phillips CH, Blanchette Porter MM, Froyman W, Glanc P, et al. O-RADS US v2022: an update from the American College of Radiology’s Ovarian-Adnexal Reporting and Data System US Committee. Radiology. 2023;308(3):e230685. [DOI] [PubMed] [Google Scholar]
- 21.Hiett AK, Sonek JD, Guy M, Reid TJ. Performance of IOTA simple rules, simple rules risk assessment, ADNEX model and O-RADS in differentiating between benign and malignant adnexal lesions in North American women. Ultrasound Obstet Gynecol. 2022;59(5):668–76. [DOI] [PubMed] [Google Scholar]
- 22.Xie WT, Wang YQ, Xiang ZS, Du ZS, Huang SX, Chen YJ, et al. Efficacy of IOTA simple rules, O-RADS, and CA125 to distinguish benign and malignant adnexal masses. J Ovarian Res. 2022;15(1):15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhou A, Hu R, Liu X, Kuyeyibieergan G. Analysis of the efficacy of adnexal mass diagnosis by senior and junior sonologists based on subjective diagnosis and O-RADS classification. BMC Womens Health. 2026. [DOI] [PMC free article] [PubMed]
- 24.Campos A, Villermain-Lécolier C, Sadowski EA, Bazot M, Touboul C, Razakamanantsoa L, et al. O-RADS scoring system for adnexal lesions: diagnostic performance on TVUS performed by an expert sonographer and MRI. Eur J Radiol. 2023;169:111172. [DOI] [PubMed] [Google Scholar]
- 25.Liu Y, Cao L, Chen S, Zhou J. Diagnostic accuracy of ultrasound classifications - O-RADS US v2022, O-RADS US v2020, and IOTA SR - in distinguishing benign and malignant adnexal masses: enhanced by combining O-RADS US v2022 with tumor marker HE4. Eur J Radiol. 2024;181:111824. [DOI] [PubMed] [Google Scholar]
- 26.Lu B, He W, Liu C, Wang P, Yang P, Zhao Z, et al. Differentiating benign from malignant ovarian masses with solid components: diagnostic performance of CEUS combined with IOTA simple rules and O-RADS. Ultrasound Med Biol. 2024;50(9):1449–58. [DOI] [PubMed] [Google Scholar]
- 27.Wang T, Cui W, Nie F, Huang X, Huang L, Liu L, et al. Comparative study of the efficacy of the Ovarian-Adnexa reporting and data system ultrasound combined with contrast-enhanced ultrasound and the ADNEX MR scoring system in the diagnosis of adnexal masses. Ultrasound Med Biol. 2023;49(9):2072–80. [DOI] [PubMed] [Google Scholar]
- 28.Sadowski EA, Thomassin-Naggara I, Rockall A, Maturen KE, Forstner R, Jha P, et al. O-RADS MRI risk stratification system: guide for assessing adnexal lesions from the ACR O-RADS committee. Radiology. 2022;303(1):35–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Guo Y, Phillips CH, Suarez-Weiss K, Roller LA, Frates MC, Benson CB, et al. Interreader agreement and intermodality concordance of O-RADS US and MRI for assessing large, complex ovarian-adnexal cysts. Radiol Imaging Cancer. 2022;4(5):e220064. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Forstner R, Thomassin-Naggara I, Cunha TM, Kinkel K, Masselli G, Kubik-Huch R, et al. ESUR recommendations for MR imaging of the sonographically indeterminate adnexal mass: an update. Eur Radiol. 2017;27(6):2248–57. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Wang R, Liu H, Tang J, Geng J. The application value of two-dimensional ultrasound combined with contrast-enhanced ultrasound in the differential diagnosis of benign, borderline, and malignant ovarian epithelial tumors. J Ovarian Res. 2024;17(1):191. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Liu C, Zhu Y, Dai K, Tan B, Dong H, Lin J, et al. Accurate prediction of benign and malignant adnexal tumors in surgical resection and conservative treatment: construction and external validation of a diagnostic model based on CEUS, HE4, and O-RADS US v2022 evaluation. J Ovarian Res. 2025;18(1):123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Sidhu PS, Cantisani V, Dietrich CF, Gilja OH, Saftoiu A, Bartels E, et al. The EFSUMB guidelines and recommendations for the clinical practice of contrast-enhanced ultrasound (CEUS) in non-hepatic applications: update 2017 (long version). Ultraschall Med. 2018;39(2):e2–44. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.