OBJECTIVES: The role of positron emission tomography (PET) scans in the staging of head and neck cancer (HNC) is unclear. The National Comprehensive Cancer Network guidelines do not recommend routine metastatic workup beyond physical examination, chest x-ray, and laboratory tests. The purpose of this report is to examine the accuracy of staging PET scans for detecting distant metastatic disease in patients with HNC. METHODS: Retrospective review of 182 consecutive newly diagnosed HNC patients who had a staging PET scan at Thomas Jefferson University Hospital between 2003 and 2007. RESULTS: The overall incidence of confirmed distant metastatic disease in this population was 5.0%. About 26 of the staging PET scans had areas suspicious for a metastatic lesion(s). Of total, 23 patients were further evaluated with imaging and/or biopsy, revealing 9 (39%) true positives, and 14 (60%) false positives. Of the 156 negative PET scans, there was 1 false negative and 155 true negatives. Thus, the sensitivity of PET was 90% and specificity was 92%. Positive predictive value was 39% and negative predictive value was 99.4%. No patients with pre-PET clinical stage I or II cancer had confirmed distant metastases. The only statistically significant predictor for metastatic disease was clinical stage IV versus all other stages (P=0.03). CONCLUSIONS: Given the marked differences in the treatment of locally advanced/nonmetastatic HNC versus metastatic HNC, we recommend PET for clinical stage IV disease. Although the sensitivity, specificity, and negative predictive value rates were acceptable, the positive predictive value was suboptimal. Patients found to have a PET scan "positive" for metastatic disease require confirmatory imaging or ideally biopsy.
OBJECTIVES: The role of positron emission tomography (PET) scans in the staging of head and neck cancer (HNC) is unclear. The National Comprehensive Cancer Network guidelines do not recommend routine metastatic workup beyond physical examination, chest x-ray, and laboratory tests. The purpose of this report is to examine the accuracy of staging PET scans for detecting distant metastatic disease in patients with HNC. METHODS: Retrospective review of 182 consecutive newly diagnosed HNC patients who had a staging PET scan at Thomas Jefferson University Hospital between 2003 and 2007. RESULTS: The overall incidence of confirmed distant metastatic disease in this population was 5.0%. About 26 of the staging PET scans had areas suspicious for a metastatic lesion(s). Of total, 23 patients were further evaluated with imaging and/or biopsy, revealing 9 (39%) true positives, and 14 (60%) false positives. Of the 156 negative PET scans, there was 1 false negative and 155 true negatives. Thus, the sensitivity of PET was 90% and specificity was 92%. Positive predictive value was 39% and negative predictive value was 99.4%. No patients with pre-PET clinical stage I or II cancer had confirmed distant metastases. The only statistically significant predictor for metastatic disease was clinical stage IV versus all other stages (P=0.03). CONCLUSIONS: Given the marked differences in the treatment of locally advanced/nonmetastatic HNC versus metastatic HNC, we recommend PET for clinical stage IV disease. Although the sensitivity, specificity, and negative predictive value rates were acceptable, the positive predictive value was suboptimal. Patients found to have a PET scan "positive" for metastatic disease require confirmatory imaging or ideally biopsy.