Aim: Evaluation of F-18-FDG PET in comparison to CT/MRI as diagnostic tool in primary and recurrent head and neck cancer.
Methods: 78 F-18-FDG PET studies were performed in 71 patients with known or suspected primary (n = 48) or recurrent (n = 30) head and neck cancer and compared to CT (n = 75) or MRI investigations (n = 3) concerning detection of the primary or recurrent tumor and detection of regional lymph node metastases in the ipsilateral and contralateral neck sides. Glucose uptake (SUV) of PET findings was correlated to tumor location, grading and dignity of the lesion.
Results: Sensitivity and specificity for PET in detection of primary tumors were 87%* and 67%, respectively (CT/MRI 67%* and 44%) (*p < 0.05), in detection of local recurrence 86% and 75%, respectively (CT/MRI 57% and 92%), in detection of necks affected by lymph node metastases 80% and 92%, respectively (CT/MRI 80% and 84%). Laryngeal, buccal (cheek) and salivary gland tumors had significant lower glucose uptake (SUV) when compared to tumors of the hypopharynx (p < 0.05). G1-tumors (mean SUV 4.26) had significant (p < 0.05) lower glucose uptake when compared to G2- and G3-tumors (mean SUV 7.73 and 8.19, respectively). Mean SUV of malignant PET findings (7.88) was significant (p < 0.05) higher than mean SUV of benign PET findings (5.70). However, a SUV threshold to improve diagnostic accuracy could not be defined.
Conclusions: F-18-FDG PET is significantly more accurate than CT/MRI for detection of head and neck cancer. Both methods are valuable for detection of cervical lymph node metastases. Glucose uptake shows correlation to histological grading. A quantitative SUV analysis does not improve diagnostic accuracy.