The value of CT, MRI, and PET-CT in detecting retropharyngeal lymph node metastasis of head and neck squamous cell carcinoma

BACKGROUND: The diagnostic accuracies of the imaging studies should be clearly acknowledged in managing head and neck cancer patients; however, the accuracies of preoperative imaging studies in detecting retropharyngeal lymph node (RPLN) metastasis are still not clarified. This study was to evaluate diagnostic accuracies of computed tomography (CT), magnetic resonance imaging (MRI), and positron emission tomography-computed tomography (PET-CT) in detecting RPLN metastasis of head and neck squamous cell carcinomas. METHODS: For 123 patients who had performed RPLN dissection during the surgery of their squamous cell carcinoma of the head and neck, preoperative CT, MRI, and/or PET-CT were reviewed for RPLN metastasis in a blinded fashion by one experienced radiologist. Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy of each imaging modality were assessed, by comparing with the histopathologic findings of the resected RPLNs that served as the standard of reference. RESULTS: RPLNs were pathologically positive for metastasis in 43 of the 123 patients (35%). Sensitivity, specificity, positive predictive value, negative predictive value, and overall accuracy in detecting metastasis to RPLN were 65, 94, 85, 83, and 84% for CT; 74, 94, 87, 87 and 87% for MRI; 83, 93, 89, 89 and 89% for PET-CT, respectively. When all the three imaging modalities were considered together (n = 74), they offered sensitivity of 90%, specificity of 91%, positive predictive value of 87%, negative predictive value of 93%, and accuracy of 91%. CONCLUSIONS: The preoperative imaging studies offered relatively high specificity rates, but rather low sensitivity rates. The three imaging modalities altogether increased diagnostic accuracies, which highlights the potential of the three studies when used altogether can minimize missed diagnoses of RPLN metastasis.