Ultrasound-based ADNEX model for differentiating between benign, borderline, and malignant epithelial ovarian tumours

Area of Science:

• Gynecologic Oncology
• Medical Imaging
• Diagnostic Accuracy

Background:

• Distinguishing between benign, borderline, and malignant epithelial ovarian tumors (EOTs) is crucial for appropriate patient management.
• Accurate preoperative assessment of adnexal masses can guide surgical planning and improve patient outcomes.

Purpose of the Study:

• To evaluate the diagnostic performance of the International Ovarian Tumor Analysis-Assessment of Different NEoplasias in the adneXa (IOTA-ADNEX) model.
• To assess the IOTA-ADNEX model's ability to differentiate benign (BeEOTs), borderline (BEOTs), and malignant epithelial ovarian tumors (MEOTs).

Main Methods:

• Retrospective analysis of 813 patients with BeEOTs, BEOTs, and MEOTs who underwent ultrasound and pelvic operations.
• Histopathological diagnosis served as the gold standard for classification.
• Calculation of sensitivity, specificity, positive predictive value, negative predictive value, and AUC for the IOTA-ADNEX model.

Main Results:

• The IOTA-ADNEX model demonstrated high sensitivity (99.1%) and specificity (73.2%) for identifying overall ovarian cancer risk at a 10% cut-off.
• Area Under the Curve (AUC) values were high for differentiating BeEOTs from MEOTs (0.987) and BeEOTs from advanced-stage OC (0.995).
• The model also showed good performance in distinguishing between other categories, with AUCs ranging from 0.614 to 0.903 for various comparisons.

Conclusions:

• The IOTA-ADNEX model exhibits strong diagnostic performance in categorizing benign, borderline, and malignant epithelial ovarian tumors.
• The model shows potential for widespread adoption to assist radiologists in the assessment of adnexal masses.
• Further validation may support the integration of the IOTA-ADNEX model into routine clinical practice for improved ovarian tumor diagnosis.