What are the novel findings of this work? Computerized ultrasound image analysis using deep neural networks (DNNs) can discriminate between benign and malignant ovarian lesions with a diagnostic accuracy comparable to that of a human expert examiner. What are the clinical implications of this work? There is a shortage of expert ultrasound examiners, resulting in suboptimal diagnostic accuracy. DNN models may have a role in the triage of women with an ovarian tumor, by supporting clinical decision-making by less experienced examiners, and potentially reducing morbidity and optimizing the use of healthcare resources.

Objectives To determine how various objective two-dimensional (2D) and three-dimensional (3D) ultrasound parameters allow prediction of deep stromal tumor invasion and lymph node involvement, in comparison to