Ultrasonographic assessment of pediatric adnexal lesions: validation of the O-RADS system and development of a simplified C-CAS framework

Introduction

Adnexal lesions represent the most common cause of pelvic masses in the pediatric population, with a reported incidence of approximately 2.2–2.6 per 100,000 and an increasing trend annually (1). Approximately 90% of these lesions are physiological cysts or benign neoplasms, while malignancies account for only 3–8% (2-4). Most benign lesions can be managed conservatively or with minimally invasive surgery. In contrast, timely referral to a gynecologic oncology specialist when malignancy is suspected can significantly improve patient outcomes (1,5). However, due to the lack of dedicated ultrasound (US) risk stratification tools for this age group, current clinical decision-making heavily relies on individual physician experience. This has led to unnecessary surgeries in 46–58% of children with benign adnexal lesions as a result of overdiagnosis and overtreatment (5-7). Such interventions not only increase perioperative risks but may also lead to premature ovarian insufficiency, pubertal developmental disorders, and impaired future fertility (8,9).

In 2020, the American College of Radiology released the Ovarian-Adnexal Reporting and Data System ultrasound (O-RADS US) lexicon and risk stratification guidelines (10), which were developed and validated in adults, with an updated version (O-RADS v2022) introduced in 2022 (11). In adult populations, using O-RADS category 4 as the optimal cutoff for distinguishing malignant ovarian tumors demonstrates a sensitivity of 94–96% and a negative predictive value (NPV) of up to 98%, effectively reducing unnecessary referrals and surgeries (12,13). However, notable differences between pediatric and adult adnexal lesions—including distinct pathological spectra, lower malignancy risk, and varied clinical presentations (3)—raise questions about the direct applicability of adult-derived models in children. Moreover, the relatively complex decision-tree structure of O-RADS v2022 may affect its reproducibility and ease of use in routine pediatric practice. Currently, evidence regarding the system’s performance in children remains scarce. Therefore, further validation of O-RADS in pediatric populations, together with the development of optimized and simplified assessment strategies, is of considerable clinical importance for improving the management of adnexal lesions in children and adolescents.

Through a multicenter retrospective analysis, this study aimed to evaluate the value of O-RADS v2022 in assessing adnexal lesions in patients under 18 years old, to analyze clinical and imaging factors affecting malignancy risk, and to propose a simplified imaging assessment framework accordingly. We present this article in accordance with the STARD reporting checklist (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2754/rc).

Methods

Patients

This multicenter retrospective study incorporated patient data from four tertiary medical institutions: Fujian Medical University Union Hospital (108 patients with 116 lesions), Yongkang First People’s Hospital (49 patients with 55 lesions), The First Affiliated Hospital of Xiamen University (124 patients with 130 lesions), and Zhangzhou Municipal Hospital (87 patients with 95 lesions). Patients from the first three centers formed the training cohort, which was used to validate the applicability of O-RADS in pediatric adnexal lesions and to develop a simplified reading framework. Patients from the fourth, independent center constituted the external validation cohort, which was reserved for the independent testing of the simplified framework. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study protocol was reviewed and approved by the Ethics Committee of Yongkang First People’s Hospital [approval No. EC2025-LW-036-01(K)], and the requirement for informed consent was waived due to the retrospective design. All participating institutions were informed of and agreed to the study.

The study included pediatric female patients with adnexal lesions who underwent surgical treatment at the four involved centers between January 2018 and May 2025. Inclusion criteria were: (I) female patients aged <18 years; (II) adnexal lesions identified by gynecological or pelvic ultrasonography; (III) surgical treatment and definitive pathological diagnosis obtained within 1 month after the US examination. Exclusion criteria were: (I) patients undergoing surgery for gynecological emergencies (e.g., ovarian torsion, ruptured corpus luteum hemorrhage); (II) poor-quality US images or missing key planes preventing O-RADS classification; (III) repeatedly enrolled cases, i.e., patients with multiple US examinations during the study period; in such instances, only the last examination leading to surgical intervention was retained to avoid data duplication (Figure 1). In cases of multiple adnexal lesions, each lesion was treated as an independent observation. Based on these criteria, the final study population included 281 patients with 301 lesions in the training cohort and 87 patients with 95 lesions in the validation cohort.

Figure 1 Flowchart of included and excluded patients. O-RADS, Ovarian-Adnexal Reporting and Data System.

Relevant clinical information was extracted from medical records, including patient age, symptoms, number of lesions, imaging data, and final pathological diagnosis. Symptoms involving changes in menstrual patterns, premenstrual vaginal bleeding, premature breast development, or precocious puberty were considered sex hormone-related (14). Because borderline ovarian tumors necessitate similar surgical staging and oncologic referral as malignant tumors, and to maintain consistency with prior pediatric studies (15), borderline tumors were classified as malignant for the purpose of our diagnostic accuracy analyses.

US examination

All included patients underwent transabdominal and, in selected cases, transrectal ultrasonography. Examinations were performed using Color Doppler Ultrasound systems from the four medical centers, including Toshiba Aplio 400/500, GE Voluson E8/E9/E10, Philips EPIQ 5/EPIQ 7, and Samsung WS80A, equipped with 2–5 MHz abdominal probes, 5-12 MHz endocavitary probes, and 4–12 MHz linear array probes. With the bladder moderately filled, patients were placed in a supine position for systematic scanning of the pelvis and, when necessary, the abdomen, to examine the uterus, bilateral ovaries, and any ovarian or pelvic masses. Transrectal US was used adjunctively when required for differential diagnosis. Images were stored in the Picture Archiving and Communication System.

Image analysis

The imaging data were independently reviewed by two sonographers—one with 5 years of experience in gynecological imaging (Y.C.) and one senior sonographer with 18 years of experience (X.Z.)—who were blinded to clinical information and pathological results. Each lesion was described and classified according to the O-RADS V2022 lexicon and classification criteria (11), with recording of US features including lesion US type, maximum diameter, external contour, number of papillary projections, acoustic shadowing, color score (color score was assessed using color doppler imaging to quantify vascularity according to the O-RADS lexicon, with scores ranging from 1 to 4), and presence of ascites. Lesions were categorized into five US risk classes: O-RADS 1: normal ovary with follicles or corpus luteum ≤3.0 cm; O-RADS 2: almost certainly benign (<1% risk of malignancy); O-RADS 3: low risk (1–10% risk of malignancy); O-RADS 4: intermediate risk (10% to <50% risk of malignancy); O-RADS 5: high risk (≥50% risk of malignancy). Both readers were proficient in applying O-RADS V2022. In cases of uncertainty in terminology application or classification, the two readers re-evaluated the images together to reach a consensus.

Statistical analysis

Statistical analyses were performed using SPSS 25.0 (IBM Corp., Armonk, NY, USA) and R version 4.3.3. The Shapiro-Wilk test was used to assess normality of continuous variables. Non-normally distributed variables are presented as median (interquartile range) and compared using the Mann-Whitney U test. Categorical variables are presented as frequencies (percentages) and compared using the chi-square or Fisher’s exact test. To address complete separation and small-sample bias, independent clinical and sonographic predictors of malignancy were identified through Firth penalized logistic regression, including significant univariate predictors (P<0.05). Model calibration and internal validation were assessed using bootstrap resampling. Based on these key predictors, a simplified image analysis framework—designated the Cysts-Color-Ascites-Shadowing (C-CAS) framework—was developed and subsequently validated in an independent validation cohort. Diagnostic performance was evaluated by plotting receiver operating characteristic curves and calculating the area under the curve (AUC). Sensitivity, specificity, positive predictive value (PPV), and NPV for diagnosing malignant adnexal lesions were computed for both O-RADS and the C-CAS framework. Inter-reader agreement for O-RADS US and the simplified assessment framework was assessed using Cohen’s kappa statistics on the independent raw ratings obtained prior to consensus resolution, interpreted as follows: ≤0.20 poor, 0.21–0.40 fair, 0.41–0.60 moderate, 0.61–0.80 good, and 0.81–1.00 excellent. The statistical significance of the difference between kappa values for O-RADS and C-CAS was assessed using a Z-test. A two-tailed P value <0.05 was considered statistically significant.

Results

Study population

A total of 368 pediatric patients with 396 adnexal lesions were included in this study. The overall cohort had a median age of 14 years (interquartile range, 12–15 years). Of the 396 lesions, 44 (11.11%) were pathologically confirmed as malignant.

In the training cohort (n=301 lesions), the malignant lesions (n=36 lesions) consisted predominantly of borderline cystadenomas (16/36, 44.44%), followed by germ cell tumors (8/36, 22.22%; including 2 dysgerminomas, 2 immature teratomas, 2 mixed germ cell tumors, and 2 endodermal sinus tumor), and granulosa cell tumors (7/36, 19.44%). Benign lesions (n=265) were most frequently mature cystic teratomas (114/265, 43.02%), mucinous or serous cystadenomas (57/265, 21.51%), and simple cysts (23/265, 8.68%). No statistically significant differences were observed between the training and validation cohorts in terms of baseline clinical characteristics, key sonographic features, or pathological diagnoses (all P>0.05), confirming the comparability of the two datasets (Table 1).

Key sonographic features

Key sonographic terms from the O-RADS V2022 system, including lesion category, maximum diameter, external contour, acoustic shadowing, color score, and ascites, were all significantly associated with malignant adnexal lesions in pediatric patients (all P<0.05) (Table 2). Malignant lesions typically presented as multilocular cystic, cystic-solid, or solid masses, were larger in size, had irregular external contours, exhibited higher color scores, and were sometimes associated with ascites. In contrast, acoustic shadowing was almost exclusively observed in benign lesions. Notably, “uni- or bilocular cyst, no solid component” was found exclusively in benign lesions, while “color score 4” was observed exclusively in malignant lesions (Table 2).

O-RADS classification and histological diagnosis

Among the 301 evaluated lesions in the training cohort, all 181 categorized as O-RADS 2 were benign. Of the 56 O-RADS 3 lesions, 55 were benign, with only one case (gonadoblastoma with dysgerminoma) being malignant. Among the 49 O-RADS 4 lesions, 22 were malignant, including 15 borderline tumors. The 27 benign lesions misclassified as O-RADS 4 primarily included mature cystic teratomas (10 cases), mucinous or serous cystadenomas (7 cases), hemorrhagic cysts (5 cases), and thecomas/fibromas (2 cases). Of the 15 O-RADS 5 lesions, 13 were malignant (including 4 borderline tumors), while one was a thecoma/fibroma and another was a hemorrhagic cyst (Table 3). As expected in a surgical cohort, no lesions were classified as O-RADS 1, so the performance of O-RADS for this category could not be assessed.

Diagnostic performance of O-RADS US

The observed malignancy rates for each O-RADS category in the training cohort were: O-RADS 2: 0% (0/181), O-RADS 3: 1.79% (1/56), O-RADS 4: 44.90% (22/49), and O-RADS 5: 86.67% (13/15). These rates all fell within the expected risk ranges recommended by the O-RADS guideline. Inter-reader agreement for O-RADS categorization was good [κ=0.768, 95% confidence interval (CI): 0.721–0.815]. The ROC analysis for the O-RADS US classification yielded an AUC of 0.931 (95% CI: 0.898–0.964) (Figure 2), with the optimal cutoff point being ≥ O-RADS 4. Using O-RADS categories 4 and 5 to predict malignancy resulted in a sensitivity of 97.2% (95% CI: 91.9–100.0%), specificity of 89.1% (95% CI: 85.3–92.8%), PPV of 54.7% (95% CI: 42.5–66.9%), and NPV of 99.6% (95% CI: 98.8–100.0%).

Figure 2 Receiver operating characteristic curve analysis of O-RADS US. AUC, area under the curve; CI, confidence interval; O-RADS, Ovarian-Adnexal Reporting and Data System; US, ultrasound.

Factors influencing malignancy risk and the C-CAS framework

The Firth-corrected model confirmed that “uni- or bilocular cyst without solid component” [odds ratio (OR) ≈0.00, 95% CI: 0.00–0.04, P<0.001] and “acoustic shadowing” (OR =0.04, 95% CI: 0.00–0.45, P=0.004) were strong protective factors, while “ascites” (OR =49.13, 95% CI: 4.25–417.96, P<0.001) and “color score 4” (OR =21.78, 95% CI: 1.80–255.38, P=0.012) remained significantly associated with malignancy (Table 4).

Integrating these four key features—“uni- or bilocular cyst, no solid component”, “color score 4”, “ascites”, and “acoustic shadowing”—the original O-RADS v2022 image analysis workflow was streamlined into a novel, simplified tri-categorical model termed the C-CAS framework (Figure 3).

Figure 3 Image analysis assessment flowchart of the C-CAS framework. C-CAS, Cysts Color Ascites Shadowing.

Performance and external validation of the C-CAS framework

Application of the C-CAS framework to the training cohort classified 16/301 (5.31%) lesions as high-risk, 56/301 (18.60%) as intermediate-risk, and 229/301 (76.08%) as low-risk, with corresponding malignancy rates of 100%, 35.71%, and 0%, respectively. Receiver operating characteristic analysis demonstrated an AUC of 0.932 (95% CI: 0.911–0.953) (Figure 4). Using the intermediate-risk and high-risk categories to predict malignancy yielded a sensitivity of 100.0% (95% CI: 90.3–100.0%), specificity of 85.7% (95% CI: 80.9–89.6%), PPV of 50.0% (95% CI: 38.5–61.5%), and NPV of 100.0% (95% CI: 87.1–100.0%). Calibration analysis demonstrated excellent agreement between predicted and observed probabilities (mean absolute error 0.014; Figure S1). Given the risk of overfitting inherent in small-sample predictive modeling, we further assessed model stability through bootstrap internal validation. This yielded an optimism-corrected C-index of 0.865 (original C-index 0.895, optimism 0.030), confirming that the C-CAS framework maintains acceptable discriminatory performance even after adjustment for sampling variability. To assess the potential impact of intra-patient correlation due to patients with multiple lesions, we performed a sensitivity analysis restricted to solitary lesions only (n=261, representing 86.7% of the training cohort). The diagnostic performance of C-CAS in this subset was nearly identical to that of the full cohort, with an AUC of 0.928 (95% CI: 0.905–0.951) vs. 0.932 (95% CI: 0.911–0.953), sensitivity [100.0% (95% CI: 89.1–100.0%) vs. 100.0% (95% CI: 90.3–100.0%)], and specificity [85.2% (95% CI: 80.1–89.5%) vs. 85.7% (95% CI: 80.9–89.6%)], indicating that the inclusion of patients with multiple lesions had a negligible effect on the overall estimates. Inter-reader agreement for the framework was excellent (κ=0.811, 95% CI: 0.766–0.856). The difference between kappa values for O-RADS and C-CAS was not statistically significant (P=0.19), indicating that the simplified C-CAS framework maintains comparable inter-observer reliability to the standard O-RADS system and may even slightly improve it by reducing subjective interpretation variability.

Figure 4 Receiver operating characteristic curve analysis of the C-CAS framework in the training (A) and validation (B) cohorts. AUC, area under the curve; CI, confidence interval; C-CAS, Cysts Color Ascites Shadowing.

In the independent validation cohort, the C-CAS framework maintained high diagnostic performance, with an AUC of 0.948 (95% CI: 0.916–0.980) (Figure 4). At the same diagnostic cutoff (intermediate-risk and high-risk as positive), the sensitivity was 100.0% (95% CI: 63.1–100.0%), specificity was 89.7% (95% CI: 81.3–95.2%), PPV was 47.1% (95% CI: 23.0–72.2%), and NPV was 100.0% (95% CI: 95.4–100.0%).

Comparison of C-CAS and O-RADS classifications

Management recommendations differed between the two systems in 32 cases (10.6%) in the training cohort and 10 cases (10.5%) in the validation cohort. In the training cohort, 11 cases (22.45%, 11/49) were downgraded from surgery (O-RADS 4) to surveillance (C-CAS low-risk), all correctly identified as benign; 21 cases were upgraded from O-RADS 3 to C-CAS intermediate-risk, including 20 benign lesions and one malignant case (gonadoblastoma with dysgerminoma). In the validation cohort, 4 cases were downgraded (all benign) and 6 cases were upgraded (all benign).

Decision curve analysis was performed to compare the net clinical benefit of C-CAS and O-RADS (Figure 5). In the training cohort, the area under the net benefit curve was 0.0340 for C-CAS and 0.0341 for O-RADS across clinically relevant threshold probabilities (5–50%), indicating comparable clinical utility. In the validation cohort, the corresponding values were 0.0240 and 0.0244, respectively. Both systems demonstrated superior net benefit compared to ‘treat all’ or ‘treat none’ strategies across the clinically relevant range.

Figure 5 Decision curve analysis comparing O-RADS and C-CAS frameworks for pediatric adnexal lesions. C-CAS, Cysts Color Ascites Shadowing; O-RADS, Ovarian-Adnexal Reporting and Data System.

Discussion

The clinical management of pediatric adnexal lesions is hampered by the lack of a standardized, age-specific sonographic tool, creating uncertainty in differentiating between conditions warranting conservative management and those requiring surgery. Although O-RADS US is well-validated in adults, its applicability to children remains underexplored, and its relatively complex decision tree may hinder clinical convenience. Therefore, this study not only validates the diagnostic performance of O-RADS US in a pediatric cohort, but, more importantly, translates its core principles into a novel, simplified C-CAS framework. Anchored on the key sonographic discriminators, the framework aims to provide a rapid and reproducible tool, thereby increasing diagnostic confidence and reducing inter-operator variability.

The validity of applying the O-RADS system to this population is first supported by the applicability of its standardized lexicon. Our analysis confirmed that several key terms within the O-RADS US lexicon—including lesion category, maximum diameter, external contour, acoustic shadowing, color score, and ascites—possess significant discriminatory value in pediatric patients, which aligns with the findings reported by Wang et al. (16).

Regarding diagnostic performance, our study demonstrated that O-RADS US achieves excellent diagnostic efficacy (AUC 0.931) in children, which is highly consistent with results reported in adult populations (12,17,18). This indicates that despite differences in the disease spectrum of adnexal lesions between children and adults, the morphology-based O-RADS classification criteria remain applicable to the pediatric population. Recent years have seen growing research exploring the value of O-RADS US in children. Wang et al. (16) reported good diagnostic performance of O-RADS in this population (AUC 0.944); Wu et al. (19) further compared O-RADS, IOTA Simple Rules, and the PRMI model, finding that O-RADS achieved an AUC of 0.989, demonstrating excellent discriminatory ability. These studies collectively provide robust empirical support for the application and promotion of the O-RADS system in pediatric adnexal lesions.

The observed malignancy rates for each O-RADS category in our cohort (O-RADS 2: 0%; O-RADS 3: 1.79%; O-RADS 4: 44.90%; O-RADS 5: 86.67%) all fell precisely within the expected risk ranges recommended by the O-RADS US guidelines for adults, demonstrating a clear malignancy risk gradient. Using O-RADS category 4 as the cutoff for predicting malignancy yielded a sensitivity of 97.2% and an NPV of 99.6%, comparable to results reported by Timmerman (18) and Jha (20) in adult studies. This suggests that the system also possesses excellent risk stratification capability in the pediatric population, effectively distinguishing lesions requiring different clinical intervention intensities. Its high sensitivity and NPV provide a reliable basis for opting for imaging surveillance or more conservative surgical management.

As determinants of lesion nature, we observed that ascites is a strong independent predictor of malignancy, often indicating peritoneal dissemination of tumor cells (21). Although rarely, benign conditions like severe pelvic inflammatory disease or Meigs’ syndrome can also present with ascites (22,23), its presence in the context of a cystic-solid adnexal mass significantly increases the likelihood of malignancy. Furthermore, uni- or bilocular cysts without solid components were consistently benign in our study. These lesions are typically functional ovarian cysts common in children, making this feature a reliable indicator of benignity (24). Similarly, the presence of acoustic shadowing was significantly associated with benign lesions, being a common sonographic finding in mature teratomas and fibromas. Wu et al. (25) confirmed that incorporating acoustic shadowing as a benign feature significantly improves the specificity of O-RADS assessment. In our cohort, the vast majority of lesions with acoustic shadowing were confirmed benign. The single lesion misclassified as malignant, which also exhibited a color score of 4, was histologically confirmed as a malignant teratoma. Therefore, recognizing acoustic shadowing is crucial to avoid over-classifying such benign lesions. Conversely, a color score of 4 was observed exclusively in malignant lesions, making it a highly specific indicator for predicting malignancy.

Based on the four pivotal sonographic features identified, we developed a tri-categorized simplified reading framework, termed C-CAS, to offer a more operable tool for assessing pediatric adnexal lesions. Its core logic is to leverage a few highly discriminative features: “uni- or bilocular cysts, no solid component” and “acoustic shadowing” serve as strong indicators of benignity, whereas “ascites” and “color score 4” are robust predictors of malignancy. The diagnostic performance of this simplified framework (AUC =0.932) was comparable to the full O-RADS V2022 system (AUC =0.931) in the training set and was well maintained in the independent validation set (AUC =0.948). It achieved high sensitivity and demonstrated excellent inter-observer agreement (κ=0.811), supporting its good clinical operability and reproducibility.

The C-CAS framework simplifies a complex, multi-feature decision tree into a more intuitive and user-friendly workflow (Table 5), which is particularly valuable in primary care settings or for less experienced sonographers, as it may reduce subjective variability and improve diagnostic consistency. Key clinical advantages include: (I) high sensitivity with no missed malignancies (detecting one case missed by O-RADS); (II) successful downgrade of 22.45% of O-RADS 4 benign lesions, potentially avoiding over-treatment; and (III) significant workflow simplification to a four-feature assessment. Lesions classified as low-risk support conservative management or imaging follow-up, while high-risk lesions prompt timely referral to a gynecologic oncology specialist, and intermediate-risk lesions signal the need for careful evaluation or a second opinion. This risk-stratified pathway may reduce unnecessary surgical interventions and help preserve future fertility.

Despite these advantages, the specificity (85.7%) and PPV (50.0%) of C-CAS were slightly lower than those of O-RADS using a cutoff ≥4, meaning that approximately half of lesions flagged as high- or intermediate-risk are ultimately benign, which could lead to unnecessary surgical referrals. Notably, the majority of these benign false-positive cases were classified as intermediate-risk rather than high-risk, reflecting the cautious design of the framework to prioritize sensitivity. To mitigate overtreatment—particularly the risk of unwarranted oophorectomy—we propose that intermediate-risk lesions undergo additional evaluation (e.g., second-expert US, magnetic resonance imaging, or multidisciplinary review) before final surgical decisions. When surgery is indicated, ovarian-sparing techniques should be prioritized for benign lesions to preserve future fertility. Besides, future research should explore combining imaging features from the C-CAS framework with serum biomarkers (e.g., AFP, β-hCG, CA-125), which are frequently used in clinical practice to aid diagnosis—particularly for germ cell tumors—to potentially improve specificity and PPV.

This study has several limitations. First, the retrospective design and inclusion of only surgically treated lesions introduce selection and verification bias. The observed malignancy rate (11.1%) exceeds the 3–8% reported in population-based studies, reflecting enrichment for higher-risk lesions. Thus, diagnostic performance estimates—particularly PPV and NPV—may not directly generalize to a general pediatric population, where many simple cysts are managed conservatively. In a surveillance cohort with lower pretest probability, PPV would likely be lower, while NPV might remain high if no malignancies are missed. Future prospective studies enrolling consecutive patients regardless of surgical status are needed to validate our findings in a more representative setting. Second, the events-per-variable ratio for the logistic regression was 8 (36 events, 4 predictors), which is below the commonly recommended threshold of 10, indicating a risk of overfitting and wide CIs. Although the Firth penalized regression and bootstrap internal validation support the robustness of the C-CAS framework, the wide CIs for certain predictors (e.g., ascites and color score 4) reflect the limited number of malignant events and sparse data. The external validation cohort contained only eight malignant lesions, resulting in wide CIs for sensitivity and NPV; thus, the performance estimates—particularly the ‘perfect’ sensitivity—should be interpreted with caution, as even a single misclassified malignancy would reduce sensitivity from 100% to 87.5%; this underscores the need for validation in larger, multi-institutional cohorts. Third, a potential limitation is the inclusion of patients with multiple lesions, which could introduce intra-patient correlation. However, a sensitivity analysis restricted to patients with solitary lesions yielded nearly identical results,indicating that the clustering effect had minimal impact on our overall estimates. Finally, we classified borderline tumors as malignant, which may have influenced the sonographic feature analysis, as borderline lesions often exhibit imaging characteristics intermediate between benign and malignant tumors. This grouping could affect the generalizability of our findings to populations where borderline tumors are managed differently.

Conclusions

This study validated the diagnostic performance of the O-RADS V2022 system in children and proposed a simplified C-CAS framework. The C-CAS framework maintains diagnostic accuracy comparable to the standard system while improving assessment simplicity and reproducibility, demonstrating its feasibility for clinical use.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the STARD reporting checklist. Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2754/rc

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2754/dss

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2754/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This retrospective study was approved by the Ethics Committee of Yongkang First People’s Hospital [approval No. EC2025-LW-036-01(K)], and individual consent for this retrospective analysis was waived. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. All participating institutions were informed of and agreed to the study.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Brookfield KF, Cheung MC, Koniaris LG, Sola JE, Fischer AC. A population-based analysis of 1037 malignant ovarian tumors in the pediatric population. J Surg Res 2009;156:45-9. [Crossref] [PubMed]
2. Heo SH, Kim JW, Shin SS, Jeong SI, Lim HS, Choi YD, Lee KH, Kang WD, Jeong YY, Kang HK. Review of ovarian tumors in children and adolescents: radiologic-pathologic correlation. Radiographics 2014;34:2039-55. [Crossref] [PubMed]
3. Birbas E, Kanavos T, Gkrozou F, Skentou C, Daniilidis A, Vatopoulou A. Ovarian Masses in Children and Adolescents: A Review of the Literature with Emphasis on the Diagnostic Approach. Children (Basel) 2023;10:1114. [Crossref] [PubMed]
4. Lawrence AE, Gonzalez DO, Fallat ME, Aldrink JH, Hewitt GD, Hertweck SP, et al. Factors Associated With Management of Pediatric Ovarian Neoplasms. Pediatrics 2019;144:e20182537. [Crossref] [PubMed]
5. Gonzalez DO, Cooper JN, Aldrink JH, Hewitt GD, Fallat ME, Minneci PC, Deans KJ. Variability in surgical management of benign ovarian neoplasms in children. J Pediatr Surg 2017;52:944-50. [Crossref] [PubMed]
6. Hernon M, McKenna J, Busby G, Sanders C, Garden A. The histology and management of ovarian cysts found in children and adolescents presenting to a children’s hospital from 1991 to 2007: a call for more paediatric gynaecologists. BJOG 2010;117:181-4. [Crossref] [PubMed]
7. Hermans AJ, Kluivers KB, Wijnen MH, Bulten J, Massuger LF, Coppus SF. Diagnosis and treatment of adnexal masses in children and adolescents. Obstet Gynecol 2015;125:611-5. [Crossref] [PubMed]
8. Minneci PC, Bergus KC, Lutz C, Aldrink J, Bence C, Breech L, et al. Reducing Unnecessary Oophorectomies for Benign Ovarian Neoplasms in Pediatric Patients. JAMA 2023;330:1247-54. [Crossref] [PubMed]
9. Kodaman PH. Early menopause: primary ovarian insufficiency and surgical menopause. Semin Reprod Med 2010;28:360-9. [Crossref] [PubMed]
10. Andreotti RF, Timmerman D, Strachowski LM, Froyman W, Benacerraf BR, Bennett GL, Bourne T, Brown DL, Coleman BG, Frates MC, Goldstein SR, Hamper UM, Horrow MM, Hernanz-Schulman M, Reinhold C, Rose SL, Whitcomb BP, Wolfman WL, Glanc P. O-RADS US Risk Stratification and Management System: A Consensus Guideline from the ACR Ovarian-Adnexal Reporting and Data System Committee. Radiology 2020;294:168-85. [Crossref] [PubMed]
11. Strachowski LM, Jha P, Phillips CH, Blanchette Porter MM, Froyman W, Glanc P, Guo Y, Patel MD, Reinhold C, Suh-Burgmann EJ, Timmerman D, Andreotti RF. O-RADS US v2022: An Update from the American College of Radiology’s Ovarian-Adnexal Reporting and Data System US Committee. Radiology 2023;308:e230685. [Crossref] [PubMed]
12. Cao L, Wei M, Liu Y, Fu J, Zhang H, Huang J, Pei X, Zhou J. Validation of American College of Radiology Ovarian-Adnexal Reporting and Data System Ultrasound (O-RADS US): Analysis on 1054 adnexal masses. Gynecol Oncol 2021;162:107-12. [Crossref] [PubMed]
13. Lee S, Lee JE, Hwang JA, Shin H. O-RADS US: A Systematic Review and Meta-Analysis of Category-specific Malignancy Rates. Radiology 2023;308:e223269. [Crossref] [PubMed]
14. Hermans AJ, Kluivers KB, Massuger LF, Coppus SF. External validation of the paediatric risk of malignancy index. BJOG 2016;123:448-52. [Crossref] [PubMed]
15. Phippen NT, Barnett JC, Lowery WJ, Miller CR, Leath CA 3rd. Surgical outcomes and national comprehensive cancer network compliance in advanced ovarian cancer surgery in a low volume military treatment facility. Gynecol Oncol 2013;131:158-62. [Crossref] [PubMed]
16. Wang H, Wang L, An S, Ma Q, Tu Y, Shang N, Pan Y. American college of radiology ovarian-adnexal reporting and data system ultrasound (O-RADS): Diagnostic performance and inter-reviewer agreement for ovarian masses in children. Front Pediatr 2023;11:1091735. [Crossref] [PubMed]
17. Basha MAA, Metwally MI, Gamil SA, Khater HM, Aly SA, El Sammak AA, Zaitoun MMA, Khattab EM, Azmy TM, Alayouty NA, Mohey N, Almassry HN, Yousef HY, Ibrahim SA, Mohamed EA, Mohamed AEM, Afifi AHM, Harb OA, Algazzar HY. Comparison of O-RADS, GI-RADS, and IOTA simple rules regarding malignancy rate, validity, and reliability for diagnosis of adnexal masses. Eur Radiol 2021;31:674-84. [Crossref] [PubMed]
18. Timmerman S, Valentin L, Ceusters J, Testa AC, Landolfo C, Sladkevicius P, et al. External Validation of the Ovarian-Adnexal Reporting and Data System (O-RADS) Lexicon and the International Ovarian Tumor Analysis 2-Step Strategy to Stratify Ovarian Tumors Into O-RADS Risk Groups. JAMA Oncol 2023;9:225-33. [Crossref] [PubMed]
19. Wu M, Huang L, Chen Y, Wang Y, Zhang M, Cao J, Huang Z, Zhang X. Diagnostic accuracy of Ovarian-Adnexal Reporting and Data System, IOTA Simple Rules and Pediatric Risk of Malignancy Index for pediatric adnexal lesions: comparative study. Ultrasound Obstet Gynecol 2025;66:361-7. [Crossref] [PubMed]
20. Jha P, Gupta A, Baran TM, Maturen KE, Patel-Lippmann K, Zafar HM, Kamaya A, Antil N, Barroilhet L, Sadowski EA. Diagnostic Performance of the Ovarian-Adnexal Reporting and Data System (O-RADS) Ultrasound Risk Score in Women in the United States. JAMA Netw Open 2022;5:e2216370. [Crossref] [PubMed]
21. Nougaret S, Razakamanantsoa L, Sadowski EA, Stein EB, Lakhman Y, Hindman NM, Jalaguier-Coudray A, Rockall AG, Thomassin-Naggara I. O-RADS MRI risk stratification system: pearls and pitfalls. Insights Imaging 2024;15:45. [Crossref] [PubMed]
22. Brun R, Hutmacher J, Fink D, Imesch P. Erroneously Suspected Ovarian Cancer in a 38-Year-Old Woman with Pelvic Inflammatory Disease and Chlamydia. Case Rep Obstet Gynecol 2017;2017:2514613. [Crossref] [PubMed]
23. Fernandez Diaz JJ, Navarro Desentre L. Meigs’ Syndrome. N Engl J Med 2024;390:2107. [Crossref] [PubMed]
24. Emeksiz HC, Derinöz O, Akkoyun EB, Güçlü Pınarlı F, Bideci A. Age-Specific Frequencies and Characteristics of Ovarian Cysts in Children and Adolescents. J Clin Res Pediatr Endocrinol 2017;9:58-62. [Crossref] [PubMed]
25. Wu C, Wu L, Shang J, Xie H, Peng R. Diagnostic Performance of O-RADS US (Version 2019 and Version 2022) Incorporating Acoustic Shadowing by Junior Radiologists: Analyzing 1061 Adnexal Masses. J Ultrasound Med 2025;44:845-55. [Crossref] [PubMed]