Abstract
OBJECTIVES: Evaluation of accuracy and interobserver variation of diffusion-weighted magnetic resonance imaging (DW-MRI) and 18F-fluorodeoxyglucose positron emission tomography-computed tomography (18F-FDGPET-CT) to detect residual lymph node metastases after chemoradiotherapy (CRT) in advanced staged head and neck squamous cell carcinoma (HNSCC). MATERIALS AND METHODS: Retrospectively, routinely performed DW-MRI (n=73) and 18F-FDG-PET-CT (n=58) 3months after CRT in HNSCC-patients with advanced nodal disease (N2-N3) were assessed by two radiologists and two nuclear medicine physicians (individually and in consensus). Imaging was scored dichotomously and on a five-point Likert scale. We also explored different scenarios for the potential added value of DW-MRI to PET-CT using the consensus Likert scale. Histopathology and a follow-up of 9months after CRT served as reference standard. RESULTS: Five patients (7%) had residual regional disease. DW-MRI showed a sensitivity of 60% and a specificity of 93%, vs. 100% and 84% for PET-CT, respectively. DW-MRI and PET-CT observers had 'moderate' and 'substantial' interobserver agreement (kappa=0.58 and kappa=0.64, respectively) with the dichotomous system. The combination of PET-CT and DW-MRI showed a sensitivity of 100% and a specificity of 95%. CONCLUSION: The high sensitivity of PET-CT authorizes a neck dissection in all patients with a positive test result and the high specificity of DW-MRI justifies avoidance of invasive neck dissections if the test is negative. Interobserver agreement varied as a function of test positivity criteria. Adding DW-MRI to PET-CT seemed to increase the specificity of PET-CT alone, thereby ensuring that less patients are exposed to unnecessary neck dissections.
