Background Despite widespread use, the utility of pretreatment positron emission tomography/computed tomography (PET/CT) remains undefined. In this study, we aim to determine its accuracy in nodal disease. Methods In all, 111 patients managed between 2003 and 2007 were analyzed. Results For squamous cell cancers (SCCs) with any increased nodal uptake (mild to high), sensitivity, specificity, positive predictive values (PPVs), and negative predictive values (NPVs) were 75%, 95%, 85%, and 90%. However, if PET was deemed positive for only moderate to high uptake, sensitivity, specificity, PPV, and NPV were 71%, 95%, 85%, and 89%. In patients undergoing elective neck dissection, the accuracy was lower: for any increased uptake, sensitivity, specificity, PPV, and NPV were 66%, 96%, 83%, and 92%, whereas for moderate to high uptake, these were 52%, 97%, 79%, and 89%. Conclusions Despite advances in PET/CT technology, these statistics suggest that PET/CT is not sufficiently accurate to direct superselective neck dissections, or to guide therapeutic decisions in the N0 neck.
- level-by-level analysis