Dual time point fluorine-18 fluorodeoxyglucose positron emission tomography: a potential method to differentiate malignancy from inflammation and normal tissue in the head and neck.

Fluorine-18 fluorodeoxyglucose (FDG) positron emission tomography (PET) studies imaging FDG PET imaging is used to detect and stage head and neck cancers. However, the variable physiologic uptake of FDG in different normal structures as well as at inflammatory sites may either obscure a tumor focus or be falsely interpreted to represent tumor activity. Twenty-one patients (9 men, 12 women, median age 59) were scanned serially at two time points, one at 70 min (range 47-112) and the second at 98 min (77-142) after the intravenous injection of 4.3 MBq/kg of FDG. The mean interval between emission scans was 28 min (13-49). Transmission scans were performed and regions of interest (ROIs) were overlayed on the fully corrected images. Standardiued uptake values (SUVs) were generated for the cerebellum, tongue, larynx, every lesion, and a matched contralateral site. Follow-up and pathologic studies revealed 18 squamous cell carcinomas and nine inflammatory or infectious lesions. Tumor SUVs were 4.0+/-1.6 (mean +/- SD) for the first scan and 4. 5+/-2.2 for the second scan. Contralateral SUVs were 1.2+/-0.5 and 1. 1+/-0.5 for the two scans. Tumor SUVs increased by 12%+/-12% as compared with a 5%+/-17% decrease for contralateral sites (P<0.05). SUVs for inflammatory sites (2.0+/-0.7 and 2.0+/-0.9), cerebellum (4. 2+/-1.3 and 4.3+/-1.4), tongue (1.8+/-0.4 and 1.9+/-0.5) and larynx (1.5+/-0.6 and 1.5+/-0.6) remained constant over time (+0.6%, +2.8%, +1.4%, and -2.4%; P<0.05 when compared with tumor SUV changes). The ratio tumor/contralateral SUV increased by 23%+/-29% over time while this ratio for inflamed sites increased by only 5%+/-15% (P=0.07). The time interval between scans correlated with increase in SUV for tumors (r=0.55, P<0.05) but not for any of the other ROIs. Separation was superior when studies were performed more than 30 min apart (P<0.05). These preliminary data suggest that dual time point imaging compatible with a clinical study protocol is helpful in differentiating malignant lesions from inflammation and normal tissues, especially when separated by a sufficient time interval.