Ovarian-Adnexal Reporting and Data System
O-RADS is an ACR-maintained framework for describing ovarian and adnexal masses and assigning each a probability of malignancy, so reports are consistent and management is risk-appropriate. It is a family with two independent tracks: an ultrasound track (usually first-line, tuned for sensitivity) and an MRI track (typically a problem-solver for sonographically indeterminate masses, adding specificity). Each track pairs a standardized lexicon of imaging descriptors with a numeric risk category and management guidance. The tracks version differently — ultrasound carries explicit year labels, while MRI is anchored to its 2021–2022 founding papers rather than a year-version number.
An ACR-convened multidisciplinary update (with IOTA and SRU input) to improve specificity, reduce false positives, and make the system applicable to all pelvic ultrasound reports; added benign-favoring features and sharpened management wording. Validation reports better risk stratification and fewer unnecessary surgeries versus the 2019 system.
The consensus guideline that first turned the US lexicon into a 0–5 risk-and-management system; superseded by v2022, against which it is now the comparison baseline.
The original standardized ultrasound descriptor lexicon for ovarian/adnexal findings; foundational vocabulary later folded into and superseded by the risk-stratification systems.
The ACR O-RADS Committee's MRI risk-stratification guide, turning lesion composition, signal characteristics, and enhancement of any solid tissue into a 1–5 malignancy-probability score. With the 2021 lexicon it constitutes the operative MRI track; it does not use a 'v2019/v2022' numbering scheme.
Established the standardized MRI descriptor lexicon for ovarian/adnexal lesions — the vocabulary foundation the 2022 risk system depends on. Foundational and still in force.
Both O-RADS tracks place a mass on a risk ladder. The ultrasound track uses categories from 0 (assessment incomplete) and 1 (normal/physiologic) up through 5 (high probability of malignancy), with the middle categories spanning almost-certainly-benign to intermediate risk. The MRI track uses an analogous 0–5 scheme, with higher numbers reflecting progressively greater malignancy probability based largely on enhancing solid tissue. Each category links to a suggested management direction. Consult the official ACR O-RADS documents for the precise descriptor definitions, category thresholds, and management recommendations.
These are our plain-language summaries. For the exact criteria, thresholds, and management rules, see the official source.
Going Beyond O-RADS US to Substratify Ovarian Lesions Classified as Intermediate or High-risk of Malignancy.
Invited Commentary: Optimizing O-RADS MRI: Sharper Images, Smarter Decisions.
Improving diagnostic performance of O-RADS MRI: semi-quantitative intensity curves vs visual assessment of enhancement
For ultrasound-based ovarian mass risk stratification, a vision transformer (ViT16-384) achieved AUC 0.941 and accuracy 87.4%, significantly outperforming radiologists using O-RADS v2022 (AUC 0.683, accuracy 68.0%). Hybrid human-AI models further boosted performance.
A modified O-RADS model subclassifying Category 4 adnexal masses by >10 locules and color score yielded a low-risk subgroup (4a) with 2.9% malignancy and a high-risk subgroup (4b) with 45.3% malignancy, achieving an AUC of 0.956.
O-RADS ultrasound v2022 had AUC 0.935, sensitivity 98.3%, specificity 70.8% for adnexal mass malignancy (threshold ≥ category 4). Performance comparable to v2019 and ADNEX model, with numerically higher specificity and greater reader agreement.
In 674 patients with adnexal masses, an interpretable model integrating O-RADS ultrasound with CA125, HE4, CEA, and maximum tumor diameter detected early ovarian cancer with validation AUC 0.973 (95% CI 0.952–0.994), sensitivity 0.897, specificity 0.980.
O-RADS MRI validation studies show high specificity (90–95%) for adnexal lesion characterization. The 2022 O-RADS US update refined descriptors to improve specificity while maintaining high sensitivity and reproducibility.
The June 2026 AJR issue highlights an AI model predicting future breast cancer from normal mammograms, an AI tool showing high accuracy for fracture detection on radiographs, and new data supporting contrast-enhanced ultrasound (CEUS) for characterizing O-RADS indeterminate ovar…
Adding ADC histogram parameters and CA-125 to O-RADS MRI improved AUC from 0.915 to 0.956 for classifying adnexal lesions in a retrospective study of 235 lesions.
Non-DCE MRI O-RADS score has excellent inter-rater agreement (κ=0.90) and high PPV for adnexal masses; ADCmean of solid (≤1.3×10⁻³ mm²/s) and cystic (>2.07×10⁻³ mm²/s) components perfectly discriminate malignancy (AUC 1.00, 0.95) in a retrospective study.
Retrospective study validates O-RADS US for pediatric adnexal lesions (AUC 0.93); simplified C-CAS (Cysts-Color-Ascites-Shadowing) framework maintained accuracy in external validation (AUC 0.95).
O-RADS US grew out of, and shares vocabulary with, the IOTA group's work, so it sits alongside IOTA tools such as the Simple Rules and the ADNEX risk model, as well as the older GI-RADS. The key difference is that O-RADS bundles a standardized lexicon, an ordinal risk category, and explicit management guidance into one ACR-endorsed package, whereas ADNEX outputs a calculated percentage risk and IOTA Simple Rules give a benign/malignant/inconclusive triage. The O-RADS MRI track is conceptually descended from the earlier ADNEX-MR scoring approach and is positioned as the specificity-raising second step after an indeterminate ultrasound. Head-to-head studies generally find O-RADS, ADNEX, and IOTA Simple Rules broadly comparable in discrimination, with O-RADS favored for standardized reporting.
External validation of O-RADS US reports strong discrimination between benign and malignant adnexal masses, and the v2022 update is associated with improved specificity and fewer false positives than the 2019 version, translating into fewer avoidable operations. O-RADS MRI has shown high sensitivity and specificity in multicenter cohorts, supporting its role for indeterminate masses. Comparative meta-analyses place O-RADS in the same performance tier as the IOTA ADNEX model and Simple Rules, with substantial-to-near-perfect interobserver agreement among trained readers. The most cited limitation is the intermediate category (especially O-RADS MRI 4), where reported malignancy prevalence varies widely, and external validation by less-experienced readers remains comparatively limited.
RadPigeon is an independent radiology news digest and is not affiliated with or endorsed by American College of Radiology (ACR). “O-RADS” is a trademark of its owner and is named here only to refer to the system. Always consult the official source for the exact, current criteria.