Physiologic 18F-FDG activity in head and neck imaging: scope and effect on image interpretation

Objectives Mucosal-associated FDG uptake on PET imaging causes difficulty in interpretation as physiologic uptake can mimic pathology (e.g., carcinoma). An initial retrospective pilot study (unpublished) showed maximum mucosal-associated standardized uptake value (SUVmax) of up to 6.0 in non-head & neck (non-H&N) cancer patients, without significantly greater SUVmax in regions of interest (ROI) between head and neck (H&N) (2.8±2.0 (mean±standard deviation)) vs. non-H&N (2.7±0.9) cancer patients (p>0.05, one-way ANOVA). Therefore, the purpose of this study was to determine if FDG activity in the oropharynx could be substantially reduced using convenient, innocuous procedures.

Methods Prospective study of 59 patients undergoing FDG PET/CT of the head and neck, randomly assigned to 1 of 3 groups, with additional FDG PET/CT of the head and neck after intervention. Group 1: No intervention. Group 2: Aggressively rinsed the mouth and spit into a provided container. Group 3: Wiped the floor of the mouth with an oral sponge. Baseline collected data included age, sex, admitting diagnosis and presence of misregistration and beam hardening artifact. Using a 12 mm spherical ROI, SUVmax was recorded on both scans for: most intense mucosal-associated uptake; right, left and midline floor of mouth; right, left and midline vestibulum oris; piriform sinuses; vallecula; base of tongue; nasopharynx; specimen container and background air. Average SUV was also obtained for the cerebellum using a 20 mm spherical ROI. Repeated measures ANOVA was performed to compare pre- and post-intervention.

Results No significant difference was seen between mucosal-associated SUVmax of any group before or after intervention. Pre- and post-intervention scans were not different in any ROI measured for group 1 and 3 (p>0.05). There was a statistically significant difference pre- vs. post- intervention in group 2 (p=0.019), however, subgroup analysis for individual regions reveals no significant difference, with variable direction of change per ROI. No significant difference in specimen container activity was seen between control and intervention groups.

Conclusions While FDG activity along oropharyngeal mucosal surfaces can mimic pathology (e.g., carcinoma), the majority of this activity appears fixed and not readily removable within saliva. Such physiologic uptake could relate to mucosal-associated inflammation, e.g., in gingiva or buccal mucosa, or deeper structures such as muscles, salivary glands and lymphoid tissue.