Synthetic MRI in differentiating benign from metastatic retropharyngeal lymph node: combination with diffusion-weighted imaging

ObjectivesThis study aimed to evaluate the synthetic MRI (syMRI), its combination with diffusion-weighted imaging (DWI), and morphological features for discriminating benign from metastatic retropharyngeal lymph nodes (RLNs).MethodsFifty-eight patients with a total of 63 RLNs (21 benign and 42 metastatic) were enrolled. The mean and standard deviation of syMRI-derived relaxometry parameters (T1, T2, PD; T1SD, T2SD, PDSD) were obtained from two different regions of interest (namely, partial-lesion and full-lesion ROI). The parameters derived from benign and metastatic RLNs were compared using Student’s t or chi-square tests. Logistic regression analysis was used to construct a multi-parameter model of syMRI, syMRI + DWI, and syMRI + DWI + morphological features. Areas under the curve (AUC) were compared using the DeLong test to determine the best diagnostic approach.ResultsBenign RLNs had significantly higher T1, T2, PD, and T1SD values compared with metastatic RLNs in both partial-lesion and full-lesion ROI (all p < 0.05). The T1SD obtained from full-lesion ROI showed the best diagnostic performance among all syMRI-derived single parameters. The AUC of combined syMRI multiple parameters (T1, T2, PD, T1SD) were higher than those of any single parameter from syMRI. The combination of synthetic MRI and DWI can improve the AUC regardless of ROI delineation. Furthermore, the combination of synthetic MRI, DWI-derived quantitative parameters, and morphological features can significantly improve the overall diagnostic performance.ConclusionsThe value of syMRI has been validated in differential diagnosis of benign and metastatic RLNs, and syMRI + DWI + morphological features can further improve the diagnostic efficiency for discriminating these two entities.Key Points • Synthetic MRI was useful in differential diagnosis of benign and metastatic RLNs. • The combination of syMRI, DWI, and morphological features can significantly improve the diagnostic efficiency.