Why the IOTA ADNEX Model Outperforms RMI for Ovarian Masses

Preoperative risk assessment of adnexal masses remains a critical challenge in gynecology. Fortunately, a landmark multicenter study has prospectively validated the IOTA ADNEX model in general gynecological settings. This advanced predictive tool significantly helps clinicians differentiate between benign and malignant ovarian tumors. Historically, many gynecologists relied on the Risk of Malignancy Index (RMI) or CA 125 alone. However, these older methods often fail to detect early-stage malignancies or unnecessarily refer benign cases. Consequently, this prospective validation marks a major leap forward in standardizing ovarian cancer risk stratification.

Why the IOTA ADNEX model Outperforms Traditional Tools

The primary advantage of the IOTA ADNEX model lies in its multi-class prediction capabilities. Specifically, the model estimates the probability of five distinct tumor types. These categories include benign, borderline, stage I, stage II-IV, and secondary metastatic cancers. Moreover, it integrates clinical and ultrasound variables. In contrast, traditional systems like the RMI rely heavily on menopausal status and absolute CA 125 values. Therefore, RMI often yields lower sensitivity for early-stage malignancies. Furthermore, including CA 125 does not significantly improve the ADNEX model’s performance. Thus, clinicians can confidently use the ultrasound-only version of the model. This makes it highly useful in low-resource settings where lab results take longer to arrive.

Key Findings of the Multicenter Danish Validation

A Danish multicenter study recently evaluated the model’s accuracy during routine use by non-expert ultrasound examiners. To achieve this, the researchers included 1,656 patients across five gynecological departments. Importantly, all participating clinicians received standardization training in IOTA terminology beforehand. The results demonstrated that the IOTA ADNEX model achieved an outstanding AUC of 0.92 without CA 125. In comparison, the RMI achieved a significantly lower AUC of 0.88. Furthermore, the model excelled in detecting early-stage malignancies, achieving an AUC of 0.90. At a risk cut-off of 5%, the model successfully identified 84.4% of early-stage ovarian cancers. Consequently, decision-curve analyses confirmed that this tool provides superior clinical utility compared to traditional indices.

Clinical Application and Recommendations for Gynecologists

Integrating this tool into routine clinical practice can greatly optimize referral pathways. Specifically, clinicians should seek formal training in standardized IOTA terminology. Furthermore, many modern ultrasound systems now integrate this software directly into their reporting packages. Although some providers worry about non-expert performance, this validation shows that brief training ensures excellent diagnostic accuracy. Therefore, adopting this model can significantly reduce unnecessary surgical interventions. Ultimately, this approach improves patient outcomes by ensuring timely oncological referral for high-risk cases.

Frequently Asked Questions

Q1: What is the primary clinical advantage of using the IOTA ADNEX model over the traditional Risk of Malignancy Index (RMI)?

The IOTA ADNEX model provides multi-class risk predictions for five different types of ovarian masses, including borderline and early-stage cancers, which traditional indices like the RMI cannot achieve.

Q2: Does the addition of serum CA 125 improve the diagnostic performance of the IOTA ADNEX model?

No, external validation studies show that adding CA 125 does not significantly improve the model’s ability to discriminate between benign and malignant masses. This allows clinicians to safely use the ultrasound-only version of the model.

Q3: Can non-expert ultrasound examiners successfully use the IOTA ADNEX model?

Yes, non-expert examiners trained in standardized IOTA terminology achieved excellent diagnostic accuracy in a prospective multicenter clinical setting, significantly outperforming the RMI.

References

1. Dueholm Hjorth IM et al. Prospective external validation of ADNEX and RMI for estimating risk of malignancy in adnexal masses in general gynecological settings: Danish multicenter study. Ultrasound Obstet Gynecol. 2026 Jul 09. doi: 10.1002/uog.70271. PMID: 42425520.
2. Moro F et al. External validation of ultrasound-based models for discrimination between benign and malignant adnexal masses in Italy: the prospective multicenter IOTA phase 6 study. medRxiv. 2024. doi: 10.1101/2024.11.29.24318146.
3. Yasmin et al. Comparison of ADNEX Model with GI-RADS Ultrasound Scoring System in Evaluation of Adnexal Mass. The Journal of Obstetrics and Gynecology of India. 2025;75(Suppl 1):86-92.