Head and Neck Cancer: Clinical Usefulness and Accuracy of PET/CT Image Fusion

Abstract

PURPOSE: To compare diagnostic accuracy of attenuation-corrected positron emission tomography (PET) with fused PET and computed tomography (CT) in patients with head and neck cancer and to evaluate the effect of PET/CT findings on patient care. MATERIALS AND METHODS: Studies of 68 patients were reviewed by two physicians in consensus. Focal fluorodeoxyglucose (FDG) uptake in the head and neck on attenuation-corrected PET images was graded as benign, equivocal, or malignant. CT and PET/CT images were then reviewed, and initial findings were amended if necessary. Comparison was performed on a lesion-by-lesion basis. Accuracy was evaluated on the basis of follow-up and histopathologic findings. Potential effects on patient care were assessed by a head and neck surgeon. PET and PET/CT accuracy was compared with a McNemar test adjusted for clustering. RESULTS: A total of 157 foci with abnormal FDG uptake were noted, two of which were seen only on PET/CT images. PET/CT images were essential in determining the exact anatomic location for 100 lesions (74% better localization in regions previously treated surgically or with irradiation vs 58% in untreated areas; P = .06). On the basis of PET findings alone, 45 lesions were considered benign; 39, equivocal; and 71, malignant. With PET/CT, the fraction of equivocal lesions decreased by 53%, from 39 of 155 to 18 of 157 (P < .01). PET/CT had a higher accuracy of depicting cancer than did PET (96% vs 90%, P = .03). Six proved malignancies were missed with PET, but only one was missed with PET/CT. PET/CT findings altered the care for 12 (18%) of 68 patients. CONCLUSION: PET/CT is more accurate than PET alone in the detection and anatomic localization of head and neck cancer and has the clear potential to affect patient care. © RSNA, 2004