Ultrasonography ( IOTA ‐ ADNEX ) vs magnetic resonance imaging ( ADNEX MR ) for diagnosing ovarian cancer in high‐risk patients: a prospective study

ABSTRACT Objectives To systematically validate and compare ultrasonography, using the ultrasound‐based International Ovarian Tumor Analysis – Assessment of Different NEoplasias in the adneXa (IOTA‐ADNEX) model, with magnetic resonance imaging (MRI), using the MRI‐based ADNEX MR scoring system, as second‐line imaging modalities for estimating the risk of adnexal malignancy in patients identified as being at risk of malignancy on first‐line ultrasound examination. Methods This was a prospective, single‐center diagnostic accuracy study designed as a head‐to‐head comparison. All patients referred to a tertiary referral center at Aarhus University Hospital, Aarhus, Denmark, for suspected adnexal malignancy on first‐line imaging between May 2021 and December 2022 were offered second‐line diagnostic testing. This comprised both a systematic expert ultrasound examination, including IOTA‐ADNEX risk estimation, and a systematic MRI scan with ADNEX MR scores assigned by expert radiologists. The imaging assessments were performed prospectively, with ultrasound examiners and radiologists blinded to the results of other imaging modalities and to the reference standard (tumor outcome). The reference standard was a histopathological diagnosis of the primary adnexal mass or, in a small number of cases, clinical follow‐up. Borderline tumors, invasive adnexal malignancy and metastases to the adnexa were grouped as malignancies. Diagnostic performance of the IOTA‐ADNEX model, ADNEX MR scoring system and CA 125 for predicting malignancy was evaluated using the area under the receiver‐operating‐characteristics curve (AUC), and AUCs were compared using DeLong's test. Diagnostic metrics were estimated at clinically relevant cut‐offs and scores. Agreement between modalities was assessed for the identification of morphological features and overall risk estimation using Cohen's kappa statistics, intraclass correlation coefficients (ICC) and Bland–Altman plots. For the IOTA‐ADNEX model, calibration and the ability to discriminate between tumor subtypes were evaluated. Results The included cohort comprised 342 patients (342 lesions analyzed). In the final cohort, 78.1% of patients were postmenopausal and the overall prevalence of malignancy was 58.8%. The AUC for distinguishing patients with benign tumors from those with malignant tumors was 0.96 (95% CI, 0.94–0.98) for the IOTA‐ADNEX model (with or without CA 125) compared with an AUC of 0.93 (95% CI, 0.91–0.96) for ADNEX MR. Agreement between ultrasound and MRI ranged from moderate to substantial for key morphological features (κ = 0.46–0.75) and from good to excellent for lesion‐size measurements (ICC = 0.78–0.92). The IOTA‐ADNEX model performed similarly to the ADNEX MR scoring system regardless of lesion morphology or size. The IOTA‐ADNEX model had good calibration. Furthermore, its multinomial discriminative ability between tumor subtypes was generally high, although performance was reduced when distinguishing borderline tumors from Stage‐I invasive malignancy (AUC, 0.71 (95% CI, 0.56–0.82)). Conclusions The IOTA‐ADNEX model and ADNEX MR scoring system demonstrated comparable diagnostic accuracy for differentiating benign from malignant adnexal masses in this study. Results appeared to be consistent when stratified by lesion morphology. Both modalities may be used for second‐line assessment of masses deemed at risk of malignancy in the first‐line ultrasound examination, depending on resources in specific centers. © 2026 The Author(s). Ultrasound in Obstetrics & Gynecology published by John Wiley & Sons Ltd on behalf of International Society of Ultrasound in Obstetrics and Gynecology.