Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights. ORIGINAL RESEARCH Jonathan T. Abele, BSc, MD, FRCPC Christopher I. Fung, BSc Purpose: To evaluate the association between diffuse fatty infiltration of the liver and average fluorine 18 fluorodeoxyglucose (FDG) uptake in the liver. Materials and Methods: Institutional review board approval was obtained for this study; the requirement for informed patient consent was waived. Consecutive nonenhanced whole-body hybrid FDG positron emission tomographic (PET)-computed tomographic (CT) scans obtained in 142 patients (mean age, 63.6 years; age range, 19–94 years) from October 1, 2008, to November 28, 2008, were retrospectively reviewed. Mean attenuation (in Hounsfield units) and standardized uptake value (SUV) measurements for the liver and spleen were obtained, with identical regions of interest used for the CT and PET examinations. The patients were assigned to three study groups: a control group—119 patients with a mean liver attenuation value greater than or equal to the mean spleen attenuation value, a diffuse fatty liver disease group—23 patients in whom the mean liver attenuation value was less than the mean spleen attenuation value, and a more strictly defined fatty liver disease group—a subset of 10 patients from the diffuse fatty liver disease group with a mean liver attenuation value minus mean spleen attenuation value difference of less than or equal to 210 HU. Mean SUV (SUVm) values were compared between the groups by using a two-sample t test for means. The association between mean liver attenuation and average FDG uptake was assessed with linear regression analysis. Results: The average SUVmfor the control group was 2.18 (standard deviation [SD], 0.36; 95% confidence interval [CI]: 2.12, 2.24). No significant difference was identified when the average SUVmfor the control group was compared with those for the fatty liver disease (SUVm, 2.03; SD, 0.36; 95% CI: 1.90, 2.16) (P ..05) and more strictly defined fatty liver disease (SUVm, 2.07; SD, 0.24; 95% CI: 1.92, 2.22) groups (P ..05). Linear regression analysis of liver SUVmas a function of mean liver attenuation revealed a mean slope of 0.005 (SD, 0.04; 95% CI: 20.005, 0.015) and a correlation coefficient of 0.02. Conclusion: No association between liver attenuation and FDG uptake measured in terms of SUVmwas observed. On the basis of these data, it is acceptable to use the liver as a comparator for extrahepatic foci of equivocal increased FDG activity in patients with fatty liver disease. 1 From the Department of Radiology and Diagnostic Imaging, University of Alberta, 8440-112 St, 2A2.41 WMC, Edmonton, AB, Canada T6G 2B7. Received May 6, 2009; revision requested June 3; revision received July 14; accepted August 26; final version accepted September 9. Address correspondence to J.T.A. (e-mail: jabele@ ualberta.ca). q q RSNA, 2010 Radiology: Volume 254: Number 3—March 2010 n radiology.rsna.org RSNA, 2010 917 n NUCLEAR MEDICINE Effect of Hepatic Steatosis on Liver FDG Uptake Measured in Mean Standard Uptake Values1 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake D iffuse fatty infiltration of the liver reflects a spectrum of diseases, from benign and treatable conditions (simple steatosis) to clinically important abnormalities (steatohepatitis) to irreversible (fibrosis) disease entities. Diffuse fatty infiltration of the liver is generally categorized as alcohol-related fatty liver disease or non–alcohol-related fatty liver disease (NAFLD). NAFLD is the most common chronic liver condition in the Western world, with an estimated prevalence of 20%–35% and of up to 95% in populations of obese individuals (1,2). Liver biopsy is the reference standard for the diagnosis of diffuse fatty infiltration (2). Imaging studies can also enable the detection of diffuse fatty infiltration of the liver, with ultrasonography, computed tomography (CT), and magnetic resonance imaging having been evaluated for this indication (3–5). CT specifically has been shown to have 93% sensitivity and a 76% positive predictive value for the detection of greater than 33% fatty infiltration (3). In general, at nonenhanced CT, diffuse fatty infiltration is considered when the attenuation of the liver (in Hounsfield units) is less than that of the spleen (HUL-S , 0) (1,6). This has been more strictly defined as a mean liver attenuation value minus mean spleen attenuation value difference of less than or equal to 210 HU (HUL-S ⱕ 210) (7). Advances in Knowledge n No clinically important association existed between fluorine 18 fluorodeoxyglucose (FDG) uptake measured in terms of the mean standardized uptake value (SUVm) and mean liver attenuation over a range of Hounsfield units in diffuse fatty liver disease. n On the basis of values measured in a control population of 119 patients by using modern iterative time-of-flight reconstruction techniques, the liver SUVm was 2.18, with a narrow range of variability (95% confidence interval: 2.12, 2.24). 918 Abele and Fung In many reports in the fluorine 18 fluorodeoxyglucose (FDG) positron emission tomography (PET) literature, the liver has been described as a comparator for foci of equivocal FDG accumulation, particularly in the abdomen (8–10). With this method, if the degree of FDG uptake is substantially increased compared with the degree of uptake in the liver, the focus should be considered abnormal. Previous studies have revealed that the standardized uptake value (SUV) for a normal liver remains stable over time and is essentially constant between patients with a mean SUV (SUVm) of 2.17 6 0.33 (standard deviation [SD]) when iterative reconstruction techniques are used (11,12). While literature on the evaluation of the relationship between hepatic steatosis and maximal SUV (SUVmax) is limited (13), to our knowledge there have been no reports on the assessment of the effect of diffuse fatty infiltration of the liver on mean FDG accumulation. While the SUVmax is the most common clinical parameter used to assess FDG accumulation in tumors, with respect to the clinical application of using the liver as a background comparator for other abdominal foci, the SUVm is a more robust parameter because there is less statistical variability and in general, foci are compared with liver uptake as a whole rather than as specific voxels. The SUV is calculated as the tissue concentration of FDG in a structure defined by a region of interest (ROI), divided by the activity injected per gram of body weight (12). Thus, increasing cellular size due to fat accumulation could conceivably result in less cellular density and in turn a decrease in the SUVm. Alternatively, the histologic evidence of inflammation in a number of these Implications for Patient Care n The liver is a stable organ to use as a visual comparator for equivocal foci of extrahepatic FDG uptake, including that in patients with diffuse fatty liver disease. n The liver had a consistent SUVm of 2.18, independent of liver attenuation. patients suggests possible increased regional metabolic activity related to the inflammatory response, which could result in an increase in the SUVm (1). The purpose of our study was to evaluate the association between diffuse fatty infiltration of the liver and average FDG uptake. Hybrid PET/CT technology represents an ideal method for this evaluation because each examination is performed with both PET scanning and immediate nonenhanced CT scanning, without the patient being moved. As such, the association between average FDG activity in the liver, measured in SUVm values, and mean liver attenuation, measured in Hounsfield units, can be directly assessed. Materials and Methods Patient Groups This study was approved by our institutional review board. The requirement for informed patient consent was waived. The images from 173 consecutive nonenhanced whole-body (base of skull to mid thigh) FDG PET/CT examinations performed in 173 patients at our institution from October 1, 2008, to November 28, 2008, were assessed retrospectively. Published online 10.1148/radiol.09090768 Radiology 2010; 254:917–924 Abbreviations: BMI = body mass index CI = confidence interval FDG = fluorine 18 fluorodeoxyglucose NAFLD = non–alcohol-related fatty liver disease ROI = region of interest SD = standard deviation SUV = standardized uptake value SUVm = mean SUV SUVmax = maximal SUV Author contributions: Guarantor of integrity of entire study, J.T.A.; study concepts/study design or data acquisition or data analysis/ interpretation, J.T.A., C.I.F.; manuscript drafting or manuscript revision for important intellectual content, J.T.A., C.I.F.; manuscript final version approval, J.T.A., C.I.F.; literature research, J.T.A., C.I.F.; clinical studies, C.I.F.; statistical analysis, J.T.A., C.I.F.; and manuscript editing, J.T.A., C.I.F. Authors stated no financial relationship to disclose. radiology.rsna.org n Radiology: Volume 254: Number 3—March 2010 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake The medical and imaging data of 31 of the 173 patients were excluded from analysis owing to the following exclusion criteria: patient positioning with both arms down, which results in an artificial elevation of liver attenuation due to beam-hardening effects (22 patients), the presence of confounding focal liver findings at PET or CT (five patients), age younger than 18 years (one patient), diffuse increased attenuation throughout the liver (mean attenuation, 87 HU) for unknown reasons (one patient), and technical error (patient weight accidentally entered as injected FDG dose) that resulted in inaccurate SUV measurements globally (two patients). The remaining 142 patients (mean age, 63.6 years; age range, 19–94 years) were included in the study: 78 (54.9%) men (mean age, 63.8 years; age range, 31–87 years) and 64 (45.1%) women (mean age, 63.4 years; age range, 19– 94 years). The indications for the imaging examinations included lung cancer or nodule assessment for 87 (61.3%) of the 142 patients, esophageal cancer for 10 (7.0%) patients, colon cancer for nine (6.3%) patients, head and neck cancer for nine (6.3%) patients, lymphoma for seven (4.9%) patients, genitourinary cancer for four (2.8%) patients, breast cancer for two (1.4%) patients, melanoma for one (0.7%) patient, and “other” for 13 (9.2%) patients. Three patient groups were derived from the study group of 142 patients: The control group (119 patients) included all patients with a mean liver attenuation value greater than or equal to the mean spleen attenuation value. The fatty liver disease group (23 patients) included all patients in whom the mean liver attenuation value was lower than the mean spleen attenuation value. Because in some of the literature, diffuse fatty infiltration is more strictly defined at nonenhanced CT as the liver having a mean attenuation value that is at least 10 HU lower than the mean spleen attenuation value, the data for a subset of 10 patients from the fatty liver disease group who satisfied this more strict criterion were also analyzed. Radiology: Volume 254: Number 3—March 2010 n Technical Parameters All examinations were performed by using a Gemini TF PET/CT scanner (Philips Medical Systems, Best, the Netherlands). PET was performed 60 minutes after the injection of 5.2 MBq of FDG per kilogram of body weight (maximal dose, 518 MBq). Emission data were acquired at 1 minute per bed position for patients weighing 90 kg or less and at 2 minutes per bed position for patients weighing more than 90 kg. Blood glucose measurements obtained before scanning were lower than 10 mmol/L. PET images were generated by using standard three-dimensional time-offlight iterative reconstruction algorithms with CT-based attenuation correction. Helical CT acquisitions were performed with the following parameters: a tube current of 100 effective mAs, a tube voltage of 120 kVp, a collimation of 16 3 1.5 mm, a pitch of 0.813, and a scanning time of 0.5 second per rotation. For review, the CT images were reconstructed with a section thickness of 3 mm in 1.5mm increments. All patients were given oral contrast material. No intravenous contrast material was administered. ROI Placement For each patient, a single circular ROI at least 3 cm in diameter was drawn over the right lobe of the liver; any obvious vessels were avoided. A single circular ROI measuring a minimum of 3 cm in diameter was also drawn over the spleen. Identical ROIs were used for the PET and CT examinations. The ROIs contained no perceivable focal lesions on either the CT or PET image components. SUVm, SUVmax, and mean attenuation measurements for each ROI were calculated by using commercially available software (PET/CT Application Suite V1.2E, Extended Brilliance Workspace; Philips Medical Systems). The ROIs were placed by a 3rd-year medical student (C.I.F.) and were reviewed by a physician with 5 years of experience in diagnostic radiology and nuclear medicine (J.T.A.). Body Weight, Body Mass Index, Diabetes History, and Medication History The data recorded by the clinical nurse at the time of FDG injection, as well as the medication lists provided by the radiology.rsna.org Abele and Fung referring physician on the study requisition, were reviewed. Patient body weight data were available and recorded for all patients. Patient height data were available for most of the patients and enabled calculation of the body mass index (BMI, in kg/m2). Because the imaging protocol at our institution requires that patients with diabetes be imaged first, in the morning, information regarding diabetes status and insulin use was available and recorded for all patients. The available data for all patients were reviewed with respect to the use of tamoxifen, glucocorticoids, methotrexate, zidovudine, lamivudine, amiodarone, diltiazem, or thiazalidinedione therapy at or around the time of the FDG injection. Statistical Analyses Office Excel 2007 software (Microsoft, Redmond, Wash) was used to enter the data onto a spreadsheet and perform statistical analyses. Liver SUVm and SUVmax values for the control group were compared with those for the fatty liver disease and strict fatty liver disease groups by using a two-sample t test for means. The two-sample t test was also used to compare the groups with respect to weight and BMI. A cutoff P value of less than .05 indicated statistical significance. The minimal expected detectable difference between the fatty liver group and control group as well as between the strict fatty liver group and control group was calculated by using a significance criterion of P , .05 and a power of 0.90. To graphically assess the relationship between liver SUVm as a function of mean liver attenuation (in Hounsfield units), we performed least-squares linear regression analysis of these two variables by using liver attenuation as the independent variable. The correlation coefficient for the correlation between these two variables was also calculated. Results With the criterion of a mean liver attenuation value lower than the mean spleen attenuation value (HUL-S , 0) used to indicate diffuse fatty infiltration of the liver, 23 patients had diffuse fatty 919 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake infiltration and 119 patients were control subjects—that is, their mean liver attenuation value was greater than or equal to their mean spleen attenuation value (HUL-S ⱖ 0). These data indicate a prevalence of diffuse fatty infiltration of 16.2% (23 of 142 patients). The average liver SUVm for the control group was 2.18 (standard deviation [SD], 0.36; 95% confidence interval [CI]: 2.12, 2.24), which is not significantly different from previously published data (SUVm, 2.17; SD, 0.33; 95% CI: 2.03, 2.31) (P = .91) (12). The average liver SUVm for the diffuse fatty infiltration group was 2.03 (SD, 0.31; 95% CI: 1.90, 2.16), which was not significantly different from that for the control group (P = .07). The average liver SUVm for the strict fatty liver disease group—those with a mean liver attenuation value minus mean spleen attenuation value difference of less than or equal to 210 HU (HUL-S ⱕ 210)—was 2.07 (SD, 0.24; 95% CI: 1.92, 2.22) and did not differ significantly from the average liver SUVm for the control group (P = .35) (Table 1). The mean liver SUVmax was also measured. There was no significant difference in mean liver SUVmax between the control (mean SUVmax, 2.76; SD, 0.47; 95% CI: 2.68, 2.84) and fatty liver disease (mean SUVmax, 2.63; SD, 0.42; 95% CI: 2.46, 2.80) groups (P = .22) or between the control and strict fatty liver disease (mean SUVmax, 2.54; SD, 0.3; 95% CI: 2.54, 2.92) groups (P = .23). The control group demonstrated a mean spleen attenuation value of 42.3 HU (SD, 3.4; range, 35–54 HU). The mean spleen attenuation value for the fatty liver disease group was 43.9 HU (SD, 2.8; range, 41–50 HU). The mean spleen attenuation value for the strict fatty liver disease group was 43.4 HU (SD, 2.8; range, 41–49 HU). The fatty liver disease group demonstrated a significantly higher mean body weight (89.0 kg; SD, 25.6) compared with the control group (77.1 kg; SD, 18.2) (P = .009). The strict fatty liver disease subgroup also demonstrated a significantly higher mean body weight (99.8 kg; SD, 19.4) compared with the control group (P , .001). There was 920 Abele and Fung Table 1 Liver SUVm Values for Control and Fatty Liver Disease Groups Patient Group Control Fatty liver disease Strict fatty liver disease§ HUL-S* No. of Patients Average Liver SUVm† P Value ⱖ0 ,0 ⱕ–10 119 23 10 2.18 (2.12, 2.24) 2.03 (1.90, 2.16) 2.07 (1.92, 2.22) … .07‡ .35‡ * HUL-S = mean liver attenuation value, in Hounsfield units, minus mean spleen attenuation value. † Mean values, with 95% CIs in parentheses. ‡ P values for comparisons with control group values. § The strict fatty liver disease group is a subset of the fatty liver disease group. Table 2 Mean Body Weight and BMI Values for Control and Fatty Liver Disease Groups Patient Group Control Fatty liver disease Strict fatty liver disease§ HUL-S* Body Weight (kg)† BMI† ⱖ0 ,0 ⱕ–10 77.1 (73.9, 80.4) 88.9 (78.5, 99.4)‡ 99.8 (87.8, 111.8)‡ 26.7 (25.7, 27.7) 29.5 (26.3, 32.6)‡ 30.9 (27.5, 34.2)‡ * HUL-S = mean liver attenuation value, in Hounsfield units, minus mean spleen attenuation value. † Mean values, with 95% CIs in parentheses. ‡ P , .05 for comparison with control group values. § The strict fatty liver disease group is a subset of the fatty liver disease group. no statistically significant difference in mean body weight between the fatty liver disease and strict fatty liver disease groups (P = .24). Patient height data were available for most patients and enabled calculations of the BMI. Patient height data were not recorded for nine patients in the control group, one patient in the fatty liver disease group, and one patient in the strict fatty liver disease subgroup. Because the data were analyzed retrospectively, patient height data for these patients could not be obtained. Allowing for this, the fatty liver disease group had a significantly higher mean BMI (29.5; SD, 7.5) compared with the control group (26.7; SD, 5.3) (P = .04). The strict fatty liver disease group also demonstrated a significantly higher BMI (30.8; SD, 5.2) compared with the control group (P = .03). There was no significant difference in BMI between the fatty liver disease and strict fatty liver disease groups (P = .62) (Table 2). The control group of 119 subjects included 19 patients with diabetes mellitus, seven of whom were taking insulin. The fatty liver disease group of 23 subjects included four patients with diabetes mellitus, none of whom was taking insulin. The strict fatty liver disease subgroup of 10 subjects included one patient with diabetes mellitus, who was not taking insulin. Oral antidiabetic medication histories were less reliable owing to the retrospective nature of this study; however, we noted that two patients in the fatty liver disease group, but no patients in the control group or strict fatty liver disease subgroup, were receiving thiazalidinedione therapy. The medication lists provided by the referring physicians at study requisition and the medication history taken by the clinical nurse immediately before the FDG injection revealed that at the time of FDG PET/CT examination, 16 of the 142 study patients (12 from control group, four from fatty liver disease group, one from strict fatty liver disease subgroup) were taking glucocorticoids, two patients (two from fatty liver disease group, one from strict fatty liver disease subgroup) were taking radiology.rsna.org n Radiology: Volume 254: Number 3—March 2010 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake Abele and Fung Figure 1 Figure 1: Graph illustrates FDG uptake in the liver, measured in SUVm values, as a function of mean liver attenuation, in Hounsfield units. Graph data indicate no significant association between these parameters (r = 0.02). The measured slope at linear regression analysis is 0.005. methotrexate, one patient from the control group was receiving lamivudine therapy, and one patient from the control group was taking tamoxifen. No patients were identified as having a history of amiodarone, zidovudine, or diltiazem therapy. Linear regression analysis to assess the relationship between liver SUVm as a function of mean liver attenuation revealed a slope of 0.005, with a correlation coefficient of 0.02. This relationship is plotted graphically in Figure 1. In Figure 2, the same data are plotted on the x (liver attenuation minus spleen attenuation) axis with a more clear depiction of the control and fatty liver disease groups. Representative images obtained in a control patient and a patient with fatty liver disease are shown in Figure 3. With use of a pooled SD for the study group of 0.358, the minimal expected detectable difference between the fatty liver group and the control group was calculated to be 0.26. The Radiology: Volume 254: Number 3—March 2010 n minimal expected detectable difference between the strict fatty liver group and the control group was calculated to be 0.38. Discussion The data in our study demonstrated no clinically important correlation between liver attenuation at CT and SUVm. In addition, the average liver SUVm and mean liver SUVmax measurements in the patients with evidence of diffuse fatty infiltration of the liver were not significantly different from those in the healthy control subjects. From a clinical perspective, these data are important, given the high prevalence of diffuse fatty infiltration in the Western population (20%–35%) and the continuously increasing problem of obesity, which is strongly associated with NAFLD (1,2). Also, the increased use of hybrid PET/CT rather radiology.rsna.org than PET alone has led to increased recognition of background hepatic steatosis in the population of patients who undergo clinical PET owing to the added information gained from the CT component of the hybrid examination. Given that the average FDG uptake in the liver is often used as a background comparator to determine the clinical importance of focal FDG uptake elsewhere in the abdomen (ie, adrenal glands, pancreas, lymph nodes), it is important to validate this approach in patients with diffuse liver disease. On the basis of the data in our study, it is acceptable to use average liver FDG uptake as a comparator in patients with obvious hepatic steatosis. The results of this study contradict trends recently reported in the limited literature. Qazi et al (13) described a small study with 33 patients in which liver SUVmax–to–spleen SUVmax ratios and liver SUVmax–to–mediastinum SUVmax ratios were compared between a healthy volunteer group and a fatty liver neoplasm 921 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake Abele and Fung Figure 2 Figure 2: Graph illustrates liver FDG uptake, measured in SUVm values, as a function of mean liver attenuation value minus mean spleen attenuation value difference. Circles indicate control group, and triangles indicate fatty liver disease group. The measured slope at linear regression analysis is 0.005. patient group. Although the liver-spleen SUVmax ratio for the fatty liver disease group was noted to be significantly lower than that for the healthy group (1.1 versus 1.4, P = .002), there was no significant difference in liver-mediastinum SUVmax ratio between these groups (P = .1). Concerns with this report include uncertainty in the abstract regarding how the diagnosis of fatty liver disease was made, uncertainty regarding how the ROIs were applied, and the use of SUVmax as the measured parameter. SUVmax is commonly used as a marker of FDG uptake in tumors because it is less susceptible to volumeaveraging effects. In contrast, for large anatomic structures such as the liver, where ROIs can be reliably applied, the SUVm is a more robust measure of FDG uptake statistically, because few or no unwanted voxels are included. From the clinical point of view of using the liver as a comparator for foci of FDG uptake in the abdomen and pelvis, the SUVm can be applied more directly because these foci generally are compared with 922 liver activity as a whole rather than as specific voxels. A case report by Purandare et al (14) describes a case of focal increased FDG uptake in a region of focal fatty sparing in the liver. The measured SUV in this focus was 4.0. This value is greater than that expected for a normal liver and indicates increased FDG uptake in the focus as opposed to decreased FDG uptake in the background regional fatty change. Assuming the correct site was sampled at biopsy, this finding may have been a reflection of increased focal FDG delivery related to vascular flow differences rather than of a true difference in metabolic activity. In a recent report, Borra and colleagues (15) describe an inverse association between liver fat content and hepatic glucose uptake in patients with type 2 diabetes mellitus. In that study, FDG was injected in the presence of a euglycemic hyperinsulinemic clamp. It is likely that the diminished glucose uptake in the examined population was related to increased insulin resistance in the liver rather than a direct effect of increased liver fat content. In any event, the results of the Borra et al study do not apply to the general population of patients with neoplasms examined with PET in our study. In particular, oncology patients generally are prohibited from oral caloric intake and insulin injection for a period before the FDG injection. As a result, insulin levels are generally low and the effects of insulin resistance are likely to be reduced. It should be noted that a number of medications used in oncologic and cardiology applications are known to be secondary causes of NAFLD. These medications include tamoxifen, glucocorticoids, methotrexate, zidovudine, amiodarone, and diltiazem (1,2). Because FDG PET is commonly used for oncologic and cardiology applications at both initial presentation and follow-up, secondary NAFLD may develop over a period of clinical imaging examinations. For example, the limited available medication data on the radiology.rsna.org n Radiology: Volume 254: Number 3—March 2010 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake Figure 3 Figure 3: Transverse (a, c) CT and (b, d) PET images acquired from hybrid FDG PET/CT examinations performed in two patients: one from fatty liver disease group (a, b) and the other from control group (c and d). CT scanning was performed without intravenous contrast material administration. While a clearly depicts diffuse fatty liver disease, there is no significant qualitative difference in average FDG uptake between the two patients. patients in our study indicated that 19 patients were taking at least one of these medications. The results of this study help to validate the use of background liver activity as a comparator when there has been an interval development of hepatic steatosis in these patients. Furthermore, our study results confirm previously published values for liver SUVm. The control group in our study was much larger (119 patients) than control groups in previously published studies (11,12), and the control group results validated the use of the liver as a stable comparator, with an SUVm of 2.18 (95% CI: 2.12, 2.24). The liver not only demonstrates stable FDG uptake over time (11) and between patients (12) but also is stable regardless of diffusely decreased CT attenuation. These factors are key to the use of the liver as a baseline organ for comparison in the abdomen and pelvis. It is interesting that the fatty liver disease groups demonstrated a signifiRadiology: Volume 254: Number 3—March 2010 n cantly greater mean body weight and mean BMI compared with the control group. This finding correlates with the known association between obesity and hepatic steatosis (1). It is also interesting that in this study, even the control group demonstrated a mean BMI higher than 25 (26.7). A BMI higher than 25 is categorized as overweight according to the World Health Organization classification (16), and an increased BMI has been linked to numerous cancers, including lung cancer (17), the most prevalent malignancy in our study population. One potential problem with our study was type II error—that is, the possibility that the power of the study was not high enough for the detection of a true difference between the groups. This issue in particular is raised owing to the relatively low but non–significance-indicating P value of .07 for the difference in average SUVm between the fatty liver disease and control groups. The minimal expected detectable difference between radiology.rsna.org Abele and Fung these groups was calculated to be 0.26. The minimal expected detectable difference between the strict fatty liver disease and the control groups was calculated to be 0.38. Because most clinically important changes in SUVm are greater than 0.50 (18,19), these values suggest that our study had an adequate number of subjects for the detection of a clinically importance difference in average liver SUVm. One limitation of this study was the potential for confounding lesions in the ROI that were not identifiable with use of the nonaugmented CT or PET component of the hybrid examination. While those patients with identifiable focal abnormalities in the liver were excluded from analysis, it is possible that lesions such as hemangiomata, small cysts, vascular anomalies, or even isoattenuating isometabolic metastases were included, particularly given the absence of intravenous contrast material. The importance of this potential confounder was mitigated by the relative rarity of this scenario combined with the relatively large ROIs (.3 cm) applied. Because fatty liver disease was defined as a mean liver attenuation value lower than the mean spleen attenuation value in this study, abnormally increased spleen attenuation could have had a confounding effect. The spleen attenuation measurements were reviewed and noted to have relatively narrow ranges, with no obvious outliers. Given these results, there is no evidence to suggest that the group stratification was confounded by spleen abnormalities. We also noted that no patient had a reported history of amiodarone therapy or thorium dioxide exposure. The prevalence of hepatic steatosis in this study was 16.2%. This is similar to but slightly lower than published prevalences of 20%–35%. While a component of this result may be related to the study population, this prevalence is also probably a reflection of the limited sensitivity of CT for detecting mild fatty liver disease (3). However, the fact that this study revealed no relationship between SUV and either moderate or severe fatty liver disease perceivable at CT makes it reasonable to assume that 923 NUCLEAR MEDICINE: Effect of Hepatic Steatosis on Liver FDG Uptake this relationship absence also applies to mild nonperceivable steatosis. Finally, the results from this study apply to a broad cross section of patients in a typical oncologic PET practice at a single point in time. The effect of transition from no steatosis to steatosis on liver SUV in individual patients was not directly assessed. Given that the use of some chemotherapeutic agents can result in secondary hepatic steatosis, it would be useful to evaluate specifically this group over time in future studies. In conclusion, we observed no significant association between liver attenuation and SUVm. Specifically, hepatic steatosis does not appear to have any significant effect on FDG uptake by the liver as determined by using SUVm values. The data presented in this study further validate the use of background liver activity as a stable comparator when determining the clinical importance of equivocal focal FDG uptake elsewhere in the abdomen. Specifically, the liver is acceptable to use as a comparator in patients with fatty liver disese. Acknowledgment: The authors thank Michael Grace, PhD, for his help in preparing this manuscript. References 1. Adams LA, Angulo P, Linder KD. Nonalcoholic fatty liver disease. CMAJ 2005; 172:899–905. 924 Abele and Fung 2. Ramesh S, Sanyal AJ. Evaluation and management of non-alcoholic steatohepatitis. J Hepatol 2005;42:S2–S12. 3. Saadeh S, Younossi ZM, Remer EM, et al. The utility of radiological imaging in nonalcoholic fatty liver disease. Gastroenterology 2002;123:745–750. 4. Yokoo T, Bydder M, Hamilton G, et al. Nonalcoholic fatty liver disease: diagnostic and fat-grading accuracy of low-flipangle multiecho gradient-recalled-echo MR imaging at 1.5 T. Radiology 2009;251:67– 76. 5. Mehta SR, Thomas EL, Bell JD, Johnston DG, Taylor-Robinson SD. Non-invasive means of measuring hepatic fat content. World J Gastroenterol 2008;14:3476–3483. 6. Mihmanli I, Kantarli F, Yilmaz MH, et al. Effect of diffuse fatty infiltration of the liver on hepatic artery resistance index. J Clin Ultrasound 2005;33:95–99. 7. Jacobs JE, Birnbaum BA, Shapiro MA, et al. Diagnostic criteria for fatty infiltration of the liver on contrast-enhanced helical CT. AJR Am J Roentgenol 1998;171:659– 664. 8. Khandani AH, Wahl RL. Applications of PET in liver imaging. Radiol Clin North Am 2005;43:849–860. 9. Kumar R, Xiu Y, Yu JQ, et al. 18F-FDG PET in evaluation of adrenal lesions in patients with lung cancer. J Nucl Med 2004;45:2058–2062. 10. van Kouwen MC, Jansen JB, van Goor H, de Castro S, Oyen WJ, Drenth JP. FDG-PET is able to detect pancreatic carcinoma in chronic pancreatitis. Eur J Nucl Med Mol Imaging 2005;32:399–404. 11. Paquet N, Albert A, Foidert J, Hustinx R. Within-patient variability of 18F-FDG: standardized uptake values in normal tissues. J Nucl Med 2004;45:784–788. 12. Ramos CD, Erdi YE, Gonen M, et al. FDG-PET standardized uptake values in normal anatomic structures using iterative reconstructed segmented attenuation correction and filtered back-projection. Eur J Nucl Med 2001;28:155–164. 13. Qazi F, Oliver D, Nguyen N, Osman M. Fatty liver: impact on metabolic activity as detected with 18F FDG-PET/CT [abstr]. J Nucl Med 2008;49(suppl 1):263P. 14. Purandare NC, Rangarajan VR, Rajnish A, Shah S, Arora A, Pathak S. Focal fat spared area in the liver masquerading as hepatic metastasis on F-18 FDG PET imaging. Clin Nucl Med 2008;33:802–805. 15. Borra R, Lautamäki R, Parkkola R, et al. Inverse association between liver fat content and hepatic glucose uptake in patients with type 2 diabetes mellitus. Metabolism 2008;57:1445–1451. 16. Health Canada. Canadian guidelines for body weight classification in adults. Ottawa, Ontario, Canada: Minister of Public Works and Government Services Canada, 2003. 17. Rauscher GH, Mayne ST, Janerich DT. Relation between body mass index and lung cancer risk in men and women never and former smokers. Am J Epidemiol 2000;152:506–513. 18. Nahmias C, Wahl LM. Reproducibility of standardized uptake value measurements determined by 18F-FDG PET in malignant tumors. J Nucl Med 2008;49:1804–1808. 19. Wahl RL, Jacene H, Kasamon Y, Lodge MA. From RECIST to PERCIST: evolving considerations for PET response criteria in solid tumors. J Nucl Med 2009;50:122S– 150S. radiology.rsna.org n Radiology: Volume 254: Number 3—March 2010