Ultrasound,endoscopic ultrasound elastography,and the strain ratio in differentiating benign from malignant lymph nodes

Background and study aims

Endoscopic ultrasound elastography (EUS-elastography), or sono-elastography, has emerged in the past 2 decades as a qualitative method of estimating tissue stiffness. Strain elastography allows for semi-quantitative measurements of the average elasticity of a lesion, and previous studies have proposed the strain ratio (SR) for overcoming the limitations of the elasticity score. The main objective of this study is to assess the specificity, sensitivity and predictive values of the SR measured by EUS-elastography in differentiating benign from malignant lymph nodes (LNs). This study also aims to find significant ultrasonographic features other than the SR which could help in predicting LN malignancy.

Patients and methods

This prospective study included 126 Egyptian patients with lymphadenopathy. US and EUS-elastography and the SR were assessed, in addition to detailed sonographic features, including size, longest diameter, shortest diameter, ratio of shortest/longest diameter, echotexture (echogenic or echo-poor) and hilum (lost or preserved).

Results

The SR cut-off value of 4.61 showed a sensitivity and specificity of 89.8% and 83.3%, respectively. This parameter had high positive and negative predictive values of 82.5% and 90.2%, respectively, for predicting malignant LNs. Univariate regression analysis showed that echogenicity, hilum preservation, elastography, the shortest dimension, the ratio of the shortest/longest dimension, ultrasound diagnosis and SR could be potential predictors of the final lymph node diagnosis. Sono-diagnosis depending on echogenicity, the shortest/longest diameter ratio and a preserved hilum in combination was the only predictive parameter in multivariate regression analysis.

Conclusion

EUS-elastography and the SR could be excellent prognostic indices in differentiating benign from malignant lymph nodes if combined with other US features.