Positron-emission tomography for surveillance of head and neck cancer

OBJECTIVES/HYPOTHESIS: To determine the diagnostic accuracy and the ideal timing of fluoro-fluorodeoxyglucose positron-emission tomography (PET) in the posttreatment surveillance of head and neck mucosal squamous cell carcinoma (HNSCC). STUDY DESIGN: Retrospective chart review. METHODS: Our sample includes 103 adult patients with 118 posttreatment PET scans who had undergone treatment for HNSCC. We correlated PET results with surgical pathology and clinical outcome in the subsequent 6 months. RESULTS: For the detection of locoregional persistent or recurrent HNSCC, PET scans had a sensitivity of 82%, specificity of 92%, positive predictive value (PPV) of 64%, negative predictive value (NPV) of 97%, and overall accuracy of 90%. For the detection of distant metastases, PET scans had a sensitivity of 89%, specificity of 97%, PPV of 85%, NPV of 98%, and overall accuracy of 96%. PET scans of the head and neck region performed greater than 1 month after the completion of radiation compared with scans performed within 1 month had a significantly higher sensitivity of 95% versus 55% (P < .01) and NPV of 99% versus 90% (P < .01). CONCLUSION: PET is effective in detecting distant metastases in the posttreatment surveillance for HNSCC patients. A negative PET is highly reliable for all sites. However, a positive PET in the head and neck region is unreliable because of a high false-positivity rate. PET of the head and neck region has a statistically significant risk of a false-negative reading when performed within 1 month of radiation.