Purpose: Previous studies of FDG-PET performance in detecting head and neck cancer have focused on squamous cell carcinoma. There is scarce data describing FDG-PET performance in detection of non-squamous cell head and neck tumors. With the advent of PET/CT and image fusion, it is now possible to accurately localize abnormal FDG activity to specific anatomical structures, allowing for direct correlation with pathology. The aim of this study is to assess the accuracy of PET/CT in detecting non-squamous cell (NSC) tumors of the head and neck with direct histopathological correlation of suspected lesions found to have increased FDG uptake.
Methods and Materials: From 8/02 to 1/03, forty-five patients with abnormal head and neck findings by either physical exam or imaging underwent PET/CT fusion-imaging at UCLA Medical Center. These patients were suspected or known to have an underlying neoplasm. Of these patients, we excluded cases of squamous cell carcinoma and those without histopathology specimens. This resulted in a total of 22 patients, whose PET/CT findings were correlated with biopsy results. These studies were reviewed in collaboration by the nuclear medicine, radiology, and pathology faculty at our institution.
Results: Nineteen studies demonstrated abnormal uptake within the neck on PET/CT: of these, 14 (73.7%) were confirmed to be positive for tumor within the abnormal foci at biopsy; three (15.8%) correctly identified some of the lesions, but missed other lesions later found at surgery; and two (10.5%) patients with positive PET/CT findings demonstrated no evidence of tumor involvement by histopathology. In this group, the histopathological subtypes included Hodgkin's lymphoma (4 cases), non-Hodgkin's lymphoma (3), papillary thyroid carcinoma (3), medullary thyroid carcinoma (3), non-small cell lung cancer (1), Warthin's tumor (1), neuroendocrine tumor (1), and melanoma (1). Three patients had a negative PET/CT exam: two (66.7 %) were confirmed to be a true negative by biopsy; and one (33.3%) was a false negative shown to be low-grade lymphoma by histopathologic analysis. By Bayesian analysis, PET/CT demonstrated sensitivity of 94 %, specificity of 50%, positive predictive value of 89%, and negative predictive value of 67%. In 73.7% of the positive cases, PET/CT was able to precisely localize the disease.
Conclusion: Our study indicate that FDG PET/CT is a sensitive tool in evaluating non-squamous cell head and neck tumors. Additional studies that compare FDG PET/CT to traditional FDG PET and contrast enhanced CT may be helpful.