Bo Wang1, Meng Ke Zhang1, Mao Ping Zhou2, Yuan Liu3, Nan Li1, Gang Liu4, Zhi Li Wang1. 1. Department of Ultrasound, First Medical Center, Chinese People's Liberation Army General Hospital, Beijing, China. 2. Department of Ultrasound, Affiliated Hospital of Qingdao University, Qingdao City, China. 3. Department of Ultrasound, Fifth Medical Center, Chinese People's Liberation Army General Hospital, Beijing, China. 4. Department of Radiology, First Medical Center, Chinese People's Liberation Army General Hospital, Beijing, China.
Abstract
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.
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.