Purpose: The aim of the study is to explore the diagnostic value of an original scoring system for the differentiation between benign and malignant superficial cervical lymph nodes.
Subjects and Method : There were 63 patients in this study, with 120 lymph nodes (benign 50 and malignant 70). In the malignant category were 20 lymphomas and 50 metastastic lymph nodes. The images obtained at sonoelastographic examination were assessed acording to a new elasticity score proposed by our group, with the following patterns:
· 1: the whole nodule is soft (similar to the surrounding tissues);
· 2: <50% of the nodule surface is blue with no individualized hypoechoic nodules or with soft hypoechoic nodules in the lymph node structure;
· 3 : <50% of the nodule surface is blue and individualized hard hypoechoic nodules are seen in the lymph node structure;
· 4: 50% - 100% of the nodule surface is blue with or without individualized hard hypoechoic nodules in the lymph node structure;
· 5: the blue area covers the whole nodule with or without extension in the neighboring soft tissues
· 6: blue (hard) nodule containing fluid areas (necrosis).
Results: The analysis of the interobserver agreement for the investigated score provided a weighted Kappa = 0.879, 95%CI [0.7787-0.9026] and a standard error = 0.031. For the differentiation benign – malignant lymph nodes, area under ROC curve (AUROC) was 0.852, with a sensibility of 64.29% and a specificity of 94% for the score >3. For the differentiation benign – metastatic lymph nodes, the same criteria provided an AUROC of 0.856, with a sensibility of 66% and a specificity of 94%.
Conclusions: Our study suggest that applying the proposed score provides a very good interobserver agreement, also very high specificity and resonable sensitivity in differentiation malignant from benign cervical lymph nodes.
