Normal, abnormal, and inconclusive: has the ultrasound pattern of healthy cervical lymph nodes been defined?

Magnetic resonance imaging: dynamic contrast enhancement and diffusion-weighted imaging to identify malignant cervical lymph nodes

Objective

To examine the potential of two magnetic resonance imaging (MRI) techniques-dynamic contrast enhancement (DCE) and diffusion-weighted imaging (DWI)-for the detection of malignant cervical lymph nodes.

Materials and Methods

Using DCE and DWI, we evaluated 33 cervical lymph nodes. For the DCE technique, the maximum relative enhancement, relative enhancement, time to peak enhancement, wash-in rate, wash-out rate, brevity of enhancement, and area under the curve were calculated from a semi-quantitative analysis. For the DWI technique, apparent diffusion coefficients (ADCs) were acquired in the region of interest of each lymph node. Cystic or necrotic parts were excluded. All patients underwent neck dissection or node biopsy. Imaging results were correlated with the histopathological findings. None of the patients underwent neoadjuvant treatment before neck dissection.

Results

Relative enhancement, maximum relative enhancement, and the wash-in rate were significantly higher in malignant lymph nodes than in benign lymph nodes (p < 0.009; p < 0.05; and p < 0.03, respectively). The time to peak enhancement was significantly shorter in the malignant lymph nodes (p < 0.02). In the multivariate analysis, the variables identified as being the most capable of distinguishing between benign and malignant lymph nodes were time to peak enhancement (sensitivity, 73.7%; specificity, 69.2%) and relative enhancement (sensitivity, 89.2%; specificity, 69.2%).

Conclusion

Although DCE was able to differentiate between benign and malignant lymph nodes, there is still no consensus regarding the use of a semi-quantitative analysis, which is difficult to apply in a clinical setting. Low ADCs can predict metastatic disease, although inflammatory processes might lead to false-positive results.