Purpose: The use of positron emission tomography with 18F-fluorodeoxyglucose (FDG-PET) to detect residual/recurrent squamous cell carcinoma of the head and neck has been tested only in small groups of patients. Our purpose, therefore, was to evaluate the ability of this technique to detect the presence of tumor at both primary and nodal sites in a large cohort of patients.
Methods: All patients referred for PET scanning over a 2.5-year period with a question of residual or recurrent squamous cell carcinoma of the head and neck were identified. Thirty-five of 44 patients had sufficient follow-up to be meaningful to our analysis (range, 6-33 months). PET scans were interpreted visually with knowledge of the clinical history and correlative anatomic imaging findings. Detection of disease involving primary and nodal sites was assessed independently. Additionally, because each patient had been referred in an attempt to resolve a specific clinical problem, the usefulness of PET in accurately addressing these questions was assessed.
Results: At the primary site, sensitivity and specificity for residual/recurrent disease were 100% and 64%, respectively; for nodal disease, sensitivity and specificity were 93% and 77%, respectively. In helping to resolve the clinical question being asked, the positive predictive value of the test result was 65% and the negative predictive value was 91%.
Conclusion: The high sensitivity and negative predictive value of PET scanning in our cohort of patients suggest an important role for this technique in the care of patients with suspected residual/recurrent head and neck carcinoma. The lower figures obtained for specificity and positive predictive value reflect the fact that increased FDG uptake may be due to either tumor or inflammation.