Background: The study was conducted to assess the accuracy power of Acoustic radiation force impulse imaging (ARFI) and generate a nomogram using ultrasound and ARFI to predict malignant cervical lymph nodes in head and neck squamous cell carcinoma.
Material and methods: 374 cervical lymph nodes from 67 patients were evaluated. The B-mode ultrasonography and the elastography findings were compared with the final histopathological diagnosis. Radiological variables were used to construct nomogram and clinical utility of the nomogram was cross-validated.
Results: In univariate analysis, status of the hilum, Long Axis Diameter, Short axis diameter, colour virtual touch imaging grade (VTI) and shear wave velocity were significant in predicting metastasis in the cervical lymph nodes. In multivariable analysis, it was found that predominance of red over yellow area on colour VTI was significantly associated with lymph node metastasis. A multiple logistic regression performed to ascertain the effects of on the likelihood that patients had lymph node metastasis on histopathology was statistically significant, χ2(10) = 44.96, p < 0.001. The model was able to correctly classify 93.28% of cases and the concordance index (c-index) was estimated to be 0.8773. A nomogram was thus established to predict metastasis in cervical lymph nodes.
Conclusions: ARFI increases the diagnostic accuracy of conventional USG in predicting metastatic lymph nodes in HNSCC. Adding the constructed nomogram to the conventional diagnostic pathway can provide an alternative option to frozen section and FNAC.
Keywords: Acoustic radiation force impulse imaging; Cervical lymph nodes; Elastography; HNSCC; Neck dissection; Nomogram.
Copyright © 2021 Elsevier Ltd. All rights reserved.