Related Experiment Video
Updated: Oct 21, 2026
High-frequency Ultrasound Imaging of Mouse Cervical Lymph Nodes
Published on: July 25, 2015
Ultrasound Localization Microscopy Combined With Regional Division for Differentiating Benign and Metastatic Lymph
Qing Xie1, Ruixue Cui1, Zhendan Zhao1
1Department of Ultrasound, Shandong Provincial Hospital affiliated to Shandong First Medical University, Jinan, Shandong, 250021, China.
Objective:
Ultrasound Localization Microscopy (ULM) enables high-resolution visualization of lymph node microvasculature. However, most previous studies rely on whole-node analysis, which may overlook spatial heterogeneity and be affected by confounding factors. This prospective single-center study aimed to investigate the diagnostic value of ULM combined with regional analysis for differentiating benign and metastatic lymph nodes.
Methods:
A total of 104 lymph nodes from patients with sonographically suspicious lymph nodes were included. All lymph nodes underwent ultrasound-guided core needle biopsy, which served as the reference standard for classification (benign or metastatic). Each lymph node was divided into central, peripheral, and global regions, and ULM quantitative parameters were extracted. Regional ratio features were calculated. Feature selection was performed using univariate analysis, LASSO regression, and backward stepwise logistic regression. Model performance was evaluated using 10-fold stratified cross-validation, 2000 bootstrap resamples, decision curve analysis (DCA), and clinical impact curve (CIC).
Results:
Three diagnostic models were constructed: a conventional ultrasound (US) model, a ULM model, and a combined US-ULM model. The combined model demonstrated the best discriminative performance with an AUC of 0.866 (95% CI: 0.7875-0.9320), followed by the ULM model (AUC = 0.833, 95% CI: 0.7464-0.9108) and the US model (AUC = 0.730, 95% CI: 0.6219-0.8285). In multivariable analysis, SD.Density_C_P (OR =0.0578, p = 0.012), FD_C_P (OR = 0.8847 per 0.01 increase, p = 0.003), and peripheral vascular pattern (OR = 10.9501, p < 0.001) were identified as significant predictors of lymph node metastasis. At optimal cut-off values, the combined model achieved a sensitivity of 0.733 and specificity of 0.915.
Conclusion:
The proposed region-based ULM model combined with conventional ultrasound provides an improved quantitative approach for differentiating benign and metastatic lymph nodes in patients with sonographically suspicious lymphadenopathy.
