Abstract
In this study, we determined the diagnostic performance of adding ultrasound (US) with/without fine-needle aspiration cytology (FNAC) to computed tomography (CT)/magnetic resonance imaging (MRI) in evaluating neck lymphadenopathy (LAP) in patients with head and neck cancer treated with irradiation. We included 269 patients who had neck LAP after radiotherapy (RT) or concurrent chemoradiotherapy (CCRT) resulting from cancers of the head and neck region between October 2008 and September 2018. The diagnostic methods consisted of the following: 1) CT/MRI alone, 2) CT/MRI combined with a post-RT US predictive model, and 3) CT/MRI combined with US + FNAC. We compared their diagnostic performance using receiver operating characteristic (ROC) curves. In total, 141 (52%) malignant and 128 (48%) benign LAPs were observed. Regarding the diagnostic accuracy, the area under the ROC curves was highest for the combined CT/MRI and US + FNAC (0.965), followed by the combined CT/MRI and post-RT US predictive model (0.906) and CT/MRI alone (0.836). Our data suggest that the addition of a US examination to CT/MRI resulted in higher diagnostic performance than CT/MRI alone in terms of diagnosing recurrent or persistent nodal disease during the evaluation of LAP in patients with irradiation-treated head and neck cancer.