Within-patient Variability of 18F-FDG Uptake in Normal Tissues

1588 Objectives: To evaluate the effect of blood glucose level on the variability of SUVmax in normal tissues within the same patient in two separate occasions. Methods: We retrospectively analyzed 334 18F-FDG PET/CT scans from 167 cancer patients. All patients had 2 studies, on average 152 ± 68 days apart. Ten matched volumes of interest (VOIs) were placed on the brain, tonsil, blood pool (over aortic arch), heart (left ventricular wall), lung, liver, spleen, bone marrow, fat and iliopsoas muscle opposite third lumber vertebra away from any pathological FDG uptake or artifact to calculate SUVmax. Results: 95 females and 72 males were retrospectively recruited with mean age; 47.7±15 years. Mean values for body weight, height, and injected FDG activity nearly the same in the two scans. Mean blood glucose in the first study was 103.7±27.4 mg/dl versus 105.3±26.2 mg/dl in the second study. SUVmax measured on the lungs and heart were significantly different in the 2 studies (P-value = 0.003 and 0.024 respectively); while that of other normal tissues showed insignificant difference between the two studies. SUVmax of the liver and spleen showed significant positive correlation with blood glucose level in the two studies (r= 0.3 & 0.2, with P-value= <0.01 & 0.005 versus r= 0.2 & 0.1, P-value=0.007 & 0.029 respectively); while in the brain it showed significant negative correlation (r= -0.3, P-value= <0.01 versus r= -0.2, P-value= <0.01). Conclusions: Within the same patient, liver, and spleen SUVmax maintains a positive correlation with blood glucose level across baseline and subsequent 18F-FDG PET/CT studies while SUVmax of lung and heart are significantly different between the two studies.