Diagnostic performance of 18 fluorodesoxyglucose positron emission/computed tomography and magnetic resonance imaging in detecting T1-T2 head and neck squamous cell carcinoma.

ABSTRACT

OBJECTIVES/HYPOTHESIS: The aim of this study was to assess and compare the diagnostic accuracy of 18 fluorodesoxyglucose positron emission/computed tomography (FDG-PET/CT) and magnetic resonance imaging (MRI) to detect T1-T2 head and neck squamous cell carcinoma (HNSCC). STUDY DESIGN: Prospective case series. METHODS: Thirty-five consecutive patients with histologically proven T1-T2 HNSCC were prospectively included. All patients underwent pretherapeutic FDG-PET/CT and MRI. Two nuclear medicine physicians and 2 radiologists blindly reviewed all FDG-PET/CT and MRI, respectively. A five-point qualitative scale was used to estimate tumor detection ability. Sensitivity of each modality was compared together using a McNemar test. Interobserver variability was assessed by kappa index (κ) of Cohen statistics. Maximal standardized uptake value (SUVMAX ), metabolic tumor volume (MTV) in FDG-PET/CT, and gadolinium enhancement (%GE) in MRI of each tumor were recorded and compared with T stage using a Mann-Whitney test. Tumor-to-normal tissue ratios in FDG-PET/CT and MRI (TNRPET and TNRMRI ) were calculated and compared together using a Student t test. RESULTS: Among the 35 primary tumors, 29 were detected by FDG-PET/CT and 22 by MRI. MRI detected none of the six lesions incorrectly identified by FDG-PET/CT. FDG-PET/CT correctly identified seven of the 13 MRI false-negative results. Sensitivity of FDG-PET/CT to detect T1-T2 HNSCC was significantly higher than MRI (83% vs. 63%, P = .015). T stage was significantly correlated with MTV (P = .002) unlike with SUVMAX (P = .06) and %GE (P = .70). TNRPET was significantly higher than TNRMRI (3.5 ± 3.2 vs. 1.2 ± 0.3, P < .0001). CONCLUSIONS: Our study showed a higher diagnostic accuracy of FDG-PET/CT than MRI to detect T1-T2 HNSCC with a good interobserver agreement. LEVEL OF EVIDENCE 4. Laryngoscope, 128378-385, 2018.