Clinical value of artificial intelligence computer-aided diagnosis system in lymph node ultrasonography: A STARD-compliant article.

This article explores the clinical value of the artificial intelligence computer-aided diagnosis system in the diagnosis of superficial lymph nodes. In this study, characteristic ultrasonic parameters of 185 cases of lymph nodes (including 110 benign and 75 malignant) confirmed by pathology were quantified by the superficial lymphadenopathy computer-aided diagnosis system, including margin, appearance, medulla proportion, medulla distribution, echogenicity, echogeneity, vascular density, and vascular pattern. A nonparametric rank sum test was used between the parameters. The sensitivity, specificity, and accuracy of the artificial intelligence computer-aided diagnosis system in the diagnosis of superficial lymph nodes were 80.0%, 80.9%, and 80.5%. The characteristics of margin, appearance, medulla proportion, medulla distribution, echogenicity, echogeneity, vascular density, and vascular pattern could identify benign and malignant lymph nodes. The artificial intelligence superficial lymphadenopathy computer-aided diagnosis system can screen out the characteristic ultrasonic parameters of benign and malignant lymph nodes, which has important value in lymph node clinical diagnosis.