The effect of acquisition time on visual and semi-quantitative analysis of F-18 FDG-PET studies in patients with head and neck cancer

Abstract

To evaluate the impact of acquisition time on F-18 fluorodeoxyglucose positron emission tomography (PET) image quality, lesion detection rate, standard uptake and lesion volume measures. In 17 consecutively referred head and neck cancer patients, a 10-min acquisition of the head and neck was performed in list mode after completion of a whole-body PET-computed tomography scan. For each patient, 0.5, 1, 2, 3, 5, 7 and 10-min PET images were reconstructed. Image quality was scored on a 5-point scale. Lesions were visually identified and lesion locations were compared between scans with various acquisition times. Standard uptake values (SUVs) and lesion volumes were automatically obtained and compared. Image quality was scored excellent/good for 10 and 7-min scans; moderate for 5 and 3-min scans and poor/very poor for 2, 1 and 0.5-min scans. Lesion detection rate did not differ between scans with shorter acquisition times and the reference. Similarly, SUVmax was not significantly different between images with various scan times and the reference, except for the 0.5-min PET image (Wilcoxon, P = 0.04). Overall, lesion volume increased significantly with longer scan time (Friedman, P<0.0001). For individual lesions, however, the effect of scan time on volume was quite variable. The variability was significantly larger for images with scan times below 3 min than for images with scan times of > or = 3 min. Despite poor image quality for scans with short acquisition times, the lesion detection rate was not adversely affected. Increased scan time resulted in an increase in lesion volumes. This finding is of particular interest if PET images are implemented in radiation oncology treatment planning.