The aim of this work was to reduce the SUV variability related to the time delay between 18F-FDG injection and the static PET acquisition, by means of a normalization to a 1-h time delay. Two static PET acquisitions separated by approximately 1 h were performed on each of 14 cancer patients, with SUVs on 22 hypermetabolic lesions calculated for both scans. The pairs of SUVs were normalized to each other using the parameterized input function with one free parameter (alpha3). This optimized parameter was found by computing the value which yielded equal normalized SUV pairs, on average, over the whole series. Without normalization, SUVs measured at later scans were found to be significantly greater than the earlier ones: mean (+/- SD) ratio of 0.84 (+/-0.08; range 0.69-0.97). After normalization, with an alpha3 value of 0.0257 min(-1), as expected, the mean (+/- SD) ratio was 1.00 (+/-0.07; range 0.88-1.10).