[Correlation of FDG-PET and MRI/CT with histopathology in primary diagnosis, lymph node staging and diagnosis of recurrency of head and neck cancer]

Abstract

Aim: Correct staging of head and neck cancer is important for the patient's prognosis and further therapeutic strategies. Aim of the present study was to investigate the diagnostic value of FDG-PET regarding the pre-surgical diagnosis of primary tumor and cervical lymph node metastases, the diagnosis of tumour recurrence, and the localisation of unknown primary, further to compare the results to those of morphological imaging modalities (CT/MRI) and to correlate the results of both methods with histopathological findings.

Patients/methods: 115 patients (pts) (72 x primary diagnosis, 37 x recurrence, and 6 x unknown primary) underwent FDG-PET (ECAT EXACT HR+) and CT or MRI. Results were correlated with histopathological findings in terms of detection of primary and recurrent tumors as well as lymph node metastases.

Results: Regarding the pre-surgical diagnosis, sensitivity and specificity for identifying primary tumors were 85 % and 100 % for PET and 88 % and 75 % for CT/MRI, respectively. Accuracy was 86 % for PET and 87 % for CT/MRI. Sensitivity and specificity for detecting primary lymph node involvement were 71 %/86 % for PET and 74 %/57 % for CT/MRI, resulting in an accuracy of 77 % with PET and 68 % with morphological imaging. In 23 pts histopathology revealed pT1 stages with tumor diameters < 12 mm. In 8 pts CT/MRI and in 10 pts PET failed to identify these small primary lesions. Detecting tumor recurrence (n = 37) PET showed a higher sensitivity (83 %), specificity (76 %) and accuracy (78 %) compared to CT/MRI (sensitivity: 67 %; specificity: 52 %; accuracy: 57 %). In 4/6 pts with unknown primary, imaging was able to identify a primary lesion (3/4 in FDG-PET, 2/4 in CT/MRI), in 2/6 patients even in the follow-up no primary tumor was found.

Conclusions: FDG-PET provides only minor additional information to morphological imaging concerning diagnosis of primary tumors. At a similar level of sensitivity, however, it seems to be more specific regarding the lymph node involvement. PET seems to be superior to CT/MRI in detecting tumor recurrence as well as occult primary tumors in pts with known cervical lymph node metastases.