Logistic Regression Analysis of Conventional Ultrasound, and Contrast-Enhanced Ultrasound Characteristics Is It Helpful in Differentiating Benign and Malignant Superficial Lymph Nodes?

Objectives This study aimed to screen the significant sonographic features for differentiation of benign and malignant superficial lymph nodes (LNs) by logistic regression analysis and fit a model to diagnose LNs. Methods A total of 204 pathological LNs were analyzed retrospectively. All the LNs underwent conventional ultrasound (US) and contrast-enhanced ultrasound (CEUS) examinations. A total of 16 suspicious sonographic features were used to assess LNs. All variables that were statistically related to the diagnosis of LNs were included in the logistic regression analysis in order to ascertain the significant features of diagnosing LNs, and to establish a logistic regression analysis model. Results The significant features in the logistic regression analysis model of diagnosing malignant LNs were absence of echogenic hilus, age, and absence of hilum after enhancement. According to the results of logistic regression analysis, the formula to predict whether LNs were malignant was established. The area under the receiver operating curve (ROC) was 0.908 and the accuracy, sensitivity, and specificity were 85.0%, 92.9%, and 85.3%, respectively. Conclusion The logistic regression model for the significant sonographic features of conventional US and CEUS is an effective and accurate diagnostic tool for differentiating malignant and benign LNs.

Automated summary

The study used logistic regression analysis to identify significant sonographic features for distinguishing between malignant and benign superficial lymph nodes, developing an accurate diagnostic model.