Positron emission tomography: poor sensitivity for occult tonsillar cancer

Abstract

Purpose The aim of the study was to determine the sensitivity of preoperative positron emission tomography (PET) scans in the detection of primary tumors of the tonsils. Materials and methods We conducted a retrospective review of 46 patients treated at a university medical center for occult tonsillar cancer during the years 2002 to 2004. We identified patients who underwent a preoperative PET scan to locate an unknown primary tumor. Fusion computed tomography with PET (PET/CT) was used to further delineate anatomic localization to the tonsil area. A positive PET/CT scan was defined as asymmetric increased tracer uptake in the tonsil and/or tonsillar fossa ipsilateral to the tonsillar cancer site when compared with the contralateral site. A negative PET/CT scan was defined as equivocal symmetric tracer uptake bilaterally. Results Of the 46 patients, 6 (13.0%) had pretreatment PET scans. Of these 6 patients, 16.7% (n = 1) had positive PET, 66.7% (n = 4) had negative PET, and 16.7% (n = 1) demonstrated increased tracer uptake in tonsils bilaterally greater on the side contralateral to the cancer. In this group, PET scans had a sensitivity of 0.167 and false-negative ratio of 0.667 for tonsillar cancer detection. Conclusions Although the patient population in this study is small (n = 6), the findings suggest that PET/ CT scans may offer a low sensitivity in detection of primary tonsillar cancers. However, PET/CT scans still have a significant role in the detection of other unknown primary head and neck tumors. Technical reasons for this finding are discussed.