Hepatic Steatosis (Fatty Liver Disease) in Asymptomatic Adults Identified by Unenhanced Low-Dose CT

Transcription

1 Gastrointestinal Imaging Original Research Boyce et al. CT of Hepatic Steatosis Gastrointestinal Imaging Original Research Cody J. Boyce 1 Perry J. Pickhardt 1 David H. Kim 1 Andrew J. Taylor 2 Thomas C. Winter 3 Richard J. Bruce 1 Mary J. Lindstrom 1 J. Louis Hinshaw 1 Boyce CJ, Pickhardt PJ, Kim DH, et al. Keywords: CT, CT colonography, fatty liver disease, hepatic steatosis, screening DOI: /AJR Received February 13, 2009; accepted after revision September 9, Department of Radiology, University of Wisconsin School of Medicine and Public Health, E3/311 Clinical Science Center, 600 Highland Ave., Madison, WI Address correspondence to P. J. Pickhardt (ppickhardt2@uwhealth.org). 2 Virginia Commonwealth University, Richmond, VA. 3 University of Utah, Salt Lake City, UT. AJR 2010; 194: X/10/ American Roentgen Ray Society Hepatic Steatosis (Fatty Liver Disease) in Asymptomatic Adults Identified by Unenhanced Low-Dose CT OBJECTIVE. The purpose of this study was to investigate the prevalence of hepatic steatosis in an asymptomatic U.S. adult population using attenuation values at unenhanced CT as the reference standard. We also assessed the utility of known clinical risk factors for diagnosis. MATERIALS AND METHODS. For 3,357 consecutive asymptomatic adults (1,865 women and 1,492 men; mean age, 57.0 years), hepatic and splenic CT attenuation values (Hounsfield units) were obtained by unenhanced CT using a low-dose colonography technique for colorectal cancer screening. Multiple attenuation criteria for steatosis were applied, including liver thresholds and comparison of liver and spleen attenuation. Relevant clinical risk factors were compared against a CT liver attenuation 40 HU, which has been shown to exclude mild steatosis. RESULTS. Mean liver attenuation was 58.8 ± 10.8 (SD) HU. The prevalence of moderate-to-severe hepatic steatosis (defined by liver attenuation 40 HU) was 6.2% (208/3,357). For CT attenuation criteria that include milder degrees of steatosis, prevalence increased to as high as 45.9% (1,542/3,357) for a liver-to-spleen attenuation ratio of 1.1. Overweight status (body mass index > 25) was a sensitive indicator for moderate-to-severe steatosis (92.8%) but was highly nonspecific (37.5%). Other clinical risk factors, such as diabetes, dyslipidemia, hypertension, alcohol overuse, and hepatitis, were more specific ( %) but highly insensitive ( %). Combining clinical risk factors did not substantially increase the accuracy for screening. CONCLUSION. Assessment of liver attenuation by use of unenhanced CT represents an objective and noninvasive means for detection of asymptomatic hepatic steatosis, whereas clinical risk factor assessment is unreliable. Further longitudinal investigation is needed to determine the most appropriate attenuation threshold and the risk for disease progression to steatohepatitis and cirrhosis. H epatic steatosis (fatty liver disease) describes the discrete accumulation of triglycerides within cytoplasmic vesicles of hepatocytes. Steatosis is a very common condition that is probably of no real clinical significance in the majority of cases [1 3]. However, a poorly defined subset of individuals, including non alcohol drinkers with so-called nonalcoholic fatty liver disease, may progress to steatohepatitis, fibrosis, cryptogenic cirrhosis, and hepatocellular carcinoma [4 7]. Because of a variety of factors, including the use of both widely variable and subjective diagnostic criteria, the reported prevalence of hepatic steatosis has ranged from 3% to 39% in the literature [8 16]. In particular, the prevalence of steatosis in the asymptomatic general adult population has not been established by objective criteria [17]. With the increasing awareness that progression of nonalcoholic fatty liver disease may account for a significant proportion of cryptogenic cirrhosis, there is a growing need for a safe and effective screening tool that noninvasively identifies patients who are potentially at risk. A number of clinical risk factors associated with hepatic steatosis have been identified, most notably alcohol overuse; hepatitis; and obesity, diabetes, insulin resistance, dyslipidemia, and hypertension, which together are known as the metabolic syndrome [16, 18 22]. Cross-sectional imaging studies, such as ultrasound, CT, and MRI, can all show changes of hepatic steatosis in a noninvasive manner. Although ultrasound allows only subjective assessment, both CT and AJR:194, March

2 Boyce et al. MRI can provide a more objective evaluation that is both reproducible and correlates with the degree of lipid content at liver biopsy [23 30]. In particular, although a variety of CT criteria have been tested, a liver attenuation value of 40 HU, as found by unenhanced CT, has been shown to correlate best with a pathologic fat content of 30%, indicating at least moderate hepatic steatosis [26, 31]. The purpose of this study was to assess the prevalence of hepatic steatosis in an asymptomatic U.S. adult population using low-dose unenhanced CT as the reference standard, as well as to assess the accuracy of known clinical risk factors against the CT diagnosis for moderate-to-severe fatty liver disease. Materials and Methods This HIPAA-compliant study was performed under an institutional review board approved protocol at a single institution. The requirement for informed consent for this investigation was waived. Low-dose unenhanced abdominal CT scans were obtained for 3,357 consecutive asymptomatic adults undergoing routine CT colonography screening at our institution between April 16, 2004, and November 6, The demographic data for this screening cohort are provided in Table 1. Because unenhanced CT images of the liver and spleen are obtained as part of a standard CT colonography examination, no additional scanning or radiation dose was necessary for the purposes of this study. The low-dose unenhanced MDCT protocol used in this study has been described elsewhere [32] and consisted of 120 kvp and mas. The estimated radiation dose for the supine CT series used for steatosis analysis is approximately 2.5 msv, less than annual background radiation. All studies were performed on 8- or 16-MDCT scanners (LightSpeed series, GE Healthcare), which are calibrated for attenuation on a daily basis. The extracolonic supine images were reviewed on a standard PACS workstation as 5-mm-thick sections reconstructed at 3-mm intervals [33]. Mean CT attenuation values (in Hounsfield units) were obtained and recorded for both the liver and spleen by one of six radiologists using a validated single-slice standard region-of-interest method [34] (Fig. 1). Care was taken to sample homogeneous areas representative of the parenchyma, with avoidance of vessels, bile ducts, focal lesions, focal changes of fatty liver or fatty sparing, and the surface margins. To assess the prevalence of steatosis, five unenhanced CT criteria for diagnosis were applied: liver attenuation 40 HU, liver attenuation less than or equal to spleen attenuation minus 10 HU, liver TABLE 1: Demographic and clinical characteristics of the screening cohort (n = 3,357) Characteristic a attenuation less than or equal to spleen attenuation, liver attenuation less than or equal to spleen attenuation plus 5 HU, and liver-to-spleen attenuation ratio 1.1. These criteria vary in their sensitivity for detecting mild fatty liver, which also determines the specificity for moderate-to-severe disease. A liver attenuation value 40 HU represents the most accurate criterion for moderate-tosevere disease [26, 31]. Because this more specific criterion may allow identification of the subpopulation at greatest risk for disease progression, the 40-HU threshold was chosen as the CT reference standard against which the clinical and demographic factors were compared. Demographic data and clinical risk factors for steatosis were recorded from a combination of the standard clinical intake form (e.g., age, sex, race, Value Age, mean ± SD (y) 57.0 ± 7.2 Male:female ratio 1,492:1,865 Overweight b 2,162 (64) Obese c 878 (26) Overweight but not obese d 1,284 (38) Diabetes mellitus 295 (9) Dyslipidemia 738 (22) Hypertension 784 (23) Alcohol overuse 260 (8) Hepatitis 12 (< 1) Note Except where noted, data are no. (%) of patients. a See Materials and Methods section for more details regarding clinical risk factors. b Body mass index > 25. c Body mass index > 30. d Body mass index = A Fig. 1 Unenhanced low-dose CT images from two asymptomatic adults undergoing screening CT colonography. A, 57-year-old woman with severe hepatic steatosis. Mean liver attenuation is 5 HU, and mean splenic attenuation is 53 HU. All five diagnostic CT attenuation criteria for hepatic steatosis are easily met. Note how hepatic vessels stand out against steatotic liver. B, 58-year-old man with normal liver attenuation (62 HU). height, weight, and body mass index) and electronic medical record review (e.g., diabetes or insulin resistance, dyslipidemia, hypertension, alcohol overuse, and hepatitis). Chart review included comprehensive text searches, which included pertinent International Classification of Diseases, Ninth Revision [35], codes. Overweight status was defined as a body mass index > 25, obesity was defined as body mass index > 30, and a third category, overweight but not obese, was defined as body mass index of Alcohol overuse was defined as was defined as > 2 drinks/d (> 20 g/d). Diabetes or insulin resistance, dyslipidemia, hypertension, and hepatitis were primarily identified with International Classification of Diseases, Ninth Revision, codes. The prevalence data for these clinical risk factors are listed in Table 1. B 624 AJR:194, March 2010

3 Frequency (%) CT of Hepatic Steatosis Statistical Analysis The diagnostic performance (sensitivity, specificity, accuracy, positive predictive value, and negative predictive value) of the various clinical risk factors in predicting steatosis was assessed according to the 40-HU threshold CT reference standard. A univariate analysis was also performed to assess the clinical risk factors for steatosis relative to the 40-HU criterion. In addition, a multivariate analysis was used to evaluate risk factors in combination. The Fisher s exact test was used for the univariate analysis, and a stepwise logistical regression model was used for multivariate analysis. A logistic regression model was used to identify the combination of clinical risk factors that best predict steatosis. This analysis was displayed using a receiver operating characteristic (ROC) curve. Results In this asymptomatic cohort of 3,357 adults, the mean liver attenuation at unenhanced CT was 58.8 ± 10.8 HU (SD), with a range of 14 to 91 HU (Fig. 2). The mean splenic attenuation was 54.9 HU. The prevalence of hepatic steatosis, according to the various diagnostic criteria, is summarized in Table 2. Prevalence varied widely depending on the specific diagnostic CT attenuation criterion, ranging from 6.2% (208 individuals) when we used the 40-HU threshold indicative of moderateto-severe disease to nearly half the population (45.9%; 1,542 individuals) when we used a liver-to-spleen ratio of 1.1. The sensitivity, specificity, accuracy, positive predictive value, and negative predictive value for the clinical risk factors in predicting moderate-to-severe steatosis are shown in Table 3; 95.2% (198/208) of patients with steatosis had at least one clinical risk factor, but this combined panel approach yielded a specificity of 0.4% (9/2,288) and an accuracy of 31.8% (1,069/3,357). An overweight condition (body mass index > 25) or obesity (body mass index > 30) showed the highest sensitivity among the clinical factor factors (92.8% and 63.5%, respectively) but also had very low specificity (37.5% and 76.3%, respectively). In contrast, diabetes or insulin resistance and hepatitis showed much higher specificities (92.2% and 99.7%, respectively) but were insensitive indicators (24.0% and 1.9%, respectively) for steatosis. Similarly, a history of alcohol overuse had a specificity of 92.4% but a sensitivity of only 10.1%, which also indicates that most steatosis cases in this cohort are nonalcoholic in nature (nonalcoholic fatty liver disease). In the univariate analysis of the individual risk factors Fig. 2 Histogram showing frequency of liver attenuation values. Note extended tail at lower end of Hounsfield unit (HU) spectrum, indicating varying degrees of hepatic steatosis (Table 3), body mass index, hypertension, and diabetes and insulin resistance were associated with steatosis. Other risk factors, including alcohol abuse, dyslipidemia, and hepatitis, were less associated with steatosis. In addition, demographic characteristics, such as age, sex, and ethnicity, were not associated with steatosis. The best-fit risk factor model chosen by a stepwise logistic regression is graphically represented as an ROC curve in Figure 3. The model used body mass index, diabetes and insulin resistance, dyslipidemia, alcohol overuse, and hepatitis as risk factor variables. This model and ROC curve could then be used with a specific patient s clinical risk factors and a desired sensitivity and specificity to judge the test positive or negative depending on which side of the graphed line the output specificity fell, for a given sensitivity. However, even using this best-fit model, specificity decreases quickly as sensitivity is increased (area under the curve = 0.744), again indicating that clinical risk factors are relatively poor predictors of steatosis. Discussion Nonalcoholic fatty liver disease is increasingly being recognized as an important clinical Liver Attenuation (HU) TABLE 2: Prevalence of hepatic steatosis in 3,357 consecutive asymptomatic adults According to various unenhanced CT criteria Diagnostic CT Criterion Prevalence, % (No. of Patients) Liver attenuation 40 HU a 6.2 (208) Liver attenuation (spleen attenuation 10 HU) 10.2 (341) Liver attenuation spleen attenuation 25.6 (860) Liver attenuation (spleen attenuation + 5 HU) 45.5 (1,527) Liver attenuation / spleen attenuation (1,542) a The 40-HU threshold, which indicates moderate-to-severe steatosis, was used as the main criterion for assessing the clinical risk factors. entity with a disease spectrum potentially ranging from isolated steatosis to more ominous conditions, including nonalcoholic steatohepatitis, cirrhosis, and hepatocellular carcinoma [4 7]. In fact, it is now believed that most cases of cirrhosis labeled as cryptogenic in nature are probably related to this disease spectrum. There is also increasing evidence that nonalcoholic fatty liver disease may represent the hepatic component of a systemic metabolic syndrome that includes obesity, diabetes and insulin resistance, dyslipidemia, and hypertension [16, 22]. Although these clinical factors are clearly associated with nonalcoholic fatty liver disease, our findings suggest that they represent poor diagnostic screening tools that are either unacceptably insensitive or nonspecific. Likewise, liver biopsy would be neither a suitable nor feasible front-line screening test given its invasive nature, relatively high cost, potential for complications, and inherent difficulty in selecting appropriate candidates [36 38]. Therefore, the need clearly exists for a noninvasive yet objective screening test that can accurately identify hepatic steatosis in a safe and reliable manner. For a number of reasons, low-dose unenhanced CT is a very attractive screening option for detecting hepatic steatosis. Low-dose CT AJR:194, March

4 True-Positive Rate (Sensitivity) Boyce et al. TABLE 3: Diagnostic performance of individual clinical risk Factors for detecting moderate-to-severe hepatic steatosis Risk Factor Sensitivity Specificity Accuracy Overweight not obese a 61/208 (29.3 [ ]) Overweight b 193/208 (92.8 [ ]) Obese c 132/208 (63.5 [ ]) Diabetes and insulin resistance 50/208 (24.0 [ ]) Dyslipidemia 55/208 (26.4 [ ]) Hypertension 78/208 (37.5 [ ]) Alcohol overuse 21/208 (10.1 [ ]) Hepatitis 4/208 (1.9 [ ]) 1,926/3,149 (61.2 [ ]) 1,180/3,149 (37.5 [ ]) 2,403/3,149 (76.3 [ ]) 2,904/3,149 (92.2 [ ]) 2,466/3,149 (78.3 [ ]) 2,443/3,149 (77.6 [ ]) 2,910/3,149 (92.4 [ ]) 3,141/3,149 (99.7 [ ]) 1,987/3,357 (59.2 [ ]) 1,373/3,357 (40.9 [ ]) 2,535/3,357 (75.5 [ ]) 2,954/3,357 (88.0 [ ]) 2,521/3,357 (75.1 [ ]) 2,521/3,357 (75.1 [ ]) 2,931/3,357 (87.3 [ ]) 3,145/3,357 (93.7 [ ]) technique with either static milliamperes or automated tube current modulation allows objective liver assessment with minimal radiation dose and without affecting mean liver attenuation values. Not only does CT provide fast, reproducible, objective, and noninvasive assessment, it also correlates well with pathologic fat content, obviating biopsy in most cases. A number of methods have been proposed to identify steatosis using CT, including the five criteria used here. However, a liver attenuation value of 40 HU has been shown to be the most indicative of moderate-to-severe macrovesicular steatosis [26, 31]. Because this attenuation threshold excludes cases with milder degrees of steatosis that are of doubtful clinical relevance (decreasing the overall prevalence by as much as 40%, compared with other CT criteria), it may represent a potentially useful diagnostic cutoff for screening. Unenhanced CT evaluation of the liver represents a free by-product of unenhanced CT examinations performed for other indications, such as urolithiasis and colorectal cancer screening. This information is available without the need for additional radiation dose. In the case of CT colonography screening, assessment of liver attenuation may provide additional diagnostic value along with screening for aortic aneurysms and abdominal malignancy, which has been shown to be cost effective [39]. Maximizing the diagnostic yield from extracolonic evaluation at CT colonography in a responsible manner is critical to ensure an overall net benefit to the individual. Because further workup of otherwise asymptomatic patients with steatosis would incur additional costs and potential risks, the choice of an appropriate attenuation threshold is even more critical. To effectively weigh the cost benefit considerations, one would need to take into account the costs of subsequent diagnostic workup and treatment against any potential savings related to the prevention of nonalcoholic steatohepatitis, cirrhosis, and hepatocellular carcinoma. Unfortunately, there are too many data gaps at this time for reliable cost-effectiveness analysis. Positive Predictive Value 61/1,284 (4.8 [ ]) 193/2,162 (8.9 [ ]) 132/878 (15.0 [ ]) 50/295 (16.9 [ ]) 55/738 (7.5 [ ]) 78/784 (9.9 [8 12.3]) 21/260 (8.1 [ ]) 4/12 (33.3 [ ]) Negative Predictive Value 1,926/2,073 (92.9 [ ]) 1,180/1,195 (98.7 [ ]) 2,403/2,479 (96.9 [ ]) 2,904/3,062 (94.8 [ ]) 2,466/2,619 (94.2 [ ]) 2,443/2,573 (94.9 [ ]) 2,910/3,097 (94.0 [ ]) 3,141/3,345 (93.9 [ ]) Note Data are no. of patients/total (% [95% CI]). Moderate-to-severe steatosis is defined as a liver attenuation of 40 HU or less at unenhanced CT (see text). a Body mass index = b Body mass index > 25. c Body mass index > 30. Fig. 3 Receiver operating characteristic curve of best-fit clinical risk factor model from stepwise logistic regression in screening for hepatic steatosis (area under receiver operating characteristic curve = 0.744, standard error = 0.020, and 95% CI = ) Hepatic steatosis can also be identified with other cross-sectional imaging tests, including ultrasound and MRI. However, both tests have limitations as potential primary screening tools relative to unenhanced CT assessment. Ultrasound evaluation for steatosis is subjective and can be associated with substantial observer variation [23]. Although MRI is proving to be valuable for the quantification of hepatic fat content in patients with known steatosis [30, 40, 41], cost and availability issues may preclude its use as a primary screening tool. Because of a number of confounding issues, such as widely variable and subjective False-Positive Rate (1 Specificity) p AJR:194, March 2010

5 CT of Hepatic Steatosis diagnostic criteria and inclusion of symptomatic patients, the reported prevalence of fatty liver disease in the literature has ranged from 3% to 39% [8 16]. Shortcomings with various studies have included subjective diagnostic criteria, small sample sizes, patient selection biases, and poorly defined reference standards. Depending on the specific CT criterion used, our findings also show a wide range in prevalence, from 6.2% to 45.9%. However, when a positive diagnosis is restricted to cases of at least moderate steatosis, the prevalence drops to the low end of this range, which is a more manageable rate that most likely identifies patients at greatest risk. There are a number of limitations to our study. This was a cross-sectional investigation and not a longitudinal evaluation. Therefore, we cannot yet ascertain whether the 6% of individuals with at least moderate steatosis truly represent the subset at greater risk for disease progression to steatohepatitis or cirrhosis. We plan to follow this entire screening cohort over time with the hopes of elucidating the natural history of this disease process. Another limitation is the lack of histologic correlation such that we cannot absolutely exclude the coexistence of nonalcoholic steatohepatitis or early cirrhosis. Furthermore, it is conceivable that iron deposition could mask some cases of steatosis. However, it is unlikely that these factors would play a significant role in an asymptomatic screening population. In addition, histopathologic evaluation for steatosis is typically assessed and reported in a semiquantitative manner that does not actually measure lipid content [42, 43]. Therefore, given the objective nature of CT assessment, it is conceivable that this noninvasive technique is equally or perhaps even more reliable than typical pathologic assessment. Finally, the presence or absence of many of the clinical risk factors was based on an extensive medical records review, including text searches for key words and International Classification of Diseases, Ninth Revision, codes. However, the search was not infallible, and it is likely that some risk factors were not completely accounted for in this large cohort. We do not think that this significantly affected our results. In conclusion, hepatic steatosis, or fatty liver disease, is being increasingly recognized as a potentially significant clinical condition that can progress to chronic or even fatal liver disease. However, because mild forms of steatosis are extremely common even in an asymptomatic adult population, a more specific screening tool is needed that identifies cases of greater severity. Our findings show that low-dose unenhanced CT represents a promising screening test for detecting cases of moderate-to-severe hepatic steatosis by applying a 40-HU threshold. Because liver attenuation assessment is readily available from routine unenhanced low-dose CT colonography examination, this screening tool provides a reliable, objective, and safe noninvasive means for identifying the subset of individuals with potentially relevant fatty liver disease. Furthermore, we have shown that the established clinical risk factors provide an inadequate means for effectively identifying this group of patients. It remains unproven whether this subset of individuals with moderate-to-severe fatty liver disease are truly at greater risk for progression to steatohepatitis or cirrhosis, but longitudinal follow-up is planned to answer this question. References 1. Teli MR, James OF, Burt AD, Bennett MK, Day CP. The natural history of nonalcoholic fatty liver: a follow-up study. Hepatology 1995; 22: Powell EE, Cooksley WG, Hanson R, Searle J, Halliday JW, Powell LW. The natural history of nonalcoholic steatohepatitis: a follow-up study of forty-two patients for up to 21 years. Hepatology 1990; 11: Lee RG. Nonalcoholic steatohepatitis: a study of 49 patients. Hum Pathol 1989; 20: Ekstedt M, Franzen LE, Mathiesen UL, et al. Long-term follow-up of patients with NAFLD and elevated liver enzymes. Hepatology 2006; 44: Adams LA, Lymp JF, St Sauver J, et al. The natural history of nonalcoholic fatty liver disease: a population-based cohort study. Gastroenterology 2005; 129: Angulo P. Nonalcoholic fatty liver disease. N Engl J Med 2002; 346: Bugianesi E, Leone N, Vanni E, et al. Expanding the natural history of nonalcoholic steatohepatitis: from cryptogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 2002; 123: Hultcrantz R, Glaumann H, Lindberg G, Nilsson LH. Liver investigation in 149 asymptomatic patients with moderately elevated activities of serum aminotransferases. Scand J Gastroenterol 1986; 21: Clark JM, Brancati FL, Diehl AM. The prevalence and etiology of elevated aminotransferase levels in the United States. Am J Gastroenterol 2003; 98: Hilden M, Christoffersen P, Juhl E, Dalgaard JB. Liver histology in a normal population: examinations of 503 consecutive fatal traffic casualties. Scand J Gastroenterol 1977; 12: Kammen BF, Pacharn P, Thoeni RF, et al. Focal fatty infiltration of the liver: analysis of prevalence and CT findings in children and young adults. AJR 2001; 177: Browning JD, Szczepaniak LS, Dobbins R, et al. Prevalence of hepatic steatosis in an urban population in the United States: impact of ethnicity. Hepatology 2004; 40: Park SH, Jeon WK, Kim SH, et al. Prevalence and risk factors of non-alcoholic fatty liver disease among Korean adults. J Gastroenterol Hepatol 2006; 21: Byron D, Minuk GY. Clinical hepatology: profile of an urban, hospital-based practice. Hepatology 1996; 24: Tominaga K, Kurata JH, Chen YK, et al. Prevalence of fatty liver in Japanese children and relationship to obesity: an epidemiological ultrasonographic survey. Dig Dis Sci 1995; 40: Hamaguchi M, Kojima T, Takeda N, et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med 2005; 143: Angulo P. GI epidemiology: nonalcoholic fatty liver disease. Aliment Pharmacol Ther 2007; 25: Day CP, Yeaman SJ. The biochemistry of alcoholinduced fatty liver. Biochim Biophys Acta 1994; 1215: Wanless IR, Lentz JS. Fatty liver hepatitis (steatohepatitis) and obesity: an autopsy study with analysis of risk factors. Hepatology 1990; 12: Knobler H, Schattner A, Zhornicki T, et al. Fatty liver: an additional and treatable feature of the insulin resistance syndrome. QJM 1999; 92: Pamilo M, Sotaniemi EA, Suramo I, Lahde S, Arranto AJ. Evaluation of liver steatotic and fibrous content by computerized tomography and ultrasound. Scand J Gastroenterol 1983; 18: Fan JG, Zhu J, Li XJ, et al. Fatty liver and the metabolic syndrome among Shanghai adults. J Gastroenterol Hepatol 2005; 20: Strauss S, Gavish E, Gottlieb P, Katsnelson L. Interobserver and intraobserver variability in the sonographic assessment of fatty liver. AJR 2007; 189:1449 [web]w320 W Longo R, Ricci C, Masutti F, et al. Fatty infiltration of the liver: quantification by 1 H localized magnetic resonance spectroscopy and comparison with computed tomography. Invest Radiol 1993; 28: Limanond P, Raman SS, Lassman C, et al. Macrovesicular hepatic steatosis in living related liver donors: correlation between CT and histologic AJR:194, March

6 Boyce et al. findings. Radiology 2004; 230: Park SH, Kim PN, Kim KW, et al. Macrovesicular hepatic steatosis in living liver donors: use of CT for quantitative and qualitative assessment. Radiology 2006; 239: Levenson H, Greensite F, Hoefs J, et al. Fatty infiltration of the liver: quantification with phasecontrast MR imaging at 1.5 T vs biopsy. AJR 1991; 156: Thomsen C, Becker U, Winkler K, Christoffersen P, Jensen M, Henriksen O. Quantification of liver fat using magnetic resonance spectroscopy. Magn Reson Imaging 1994; 12: Taylor KJ, Carpenter DA, Hill CR, McCready VR. Gray scale ultrasound imaging: the anatomy and pathology of the liver. Radiology 1976; 119: Yokoo T, Bydder M, Hamilton G, et al. Nonalcoholic fatty liver disease: diagnostic and fat-grading accuracy of low-flip-angle multiecho gradientrecalled-echo MR imaging at 1.5 T. Radiology 2009; 251: Kodama Y, Ng CS, Wu TT, et al. Comparison of CT methods for determining the fat content of the FOR YOUR INFORMATION liver. AJR 2007; 188: Kim DH, Pickhardt PJ, Taylor AJ, et al. CT colonography versus colonoscopy for the detection of advanced neoplasia. N Engl J Med 2007; 357: Pickhardt PJ, Hanson ME, Vanness DJ, et al. Unsuspected extracolonic findings at screening CT colonography: clinical and economic impact. Radiology 2008; 249: Speliotes EK, Massaro JM, Hoffmann U, et al. Liver fat is reproducibly measured using computed tomography in the Framingham Heart Study. J Gastroenterol Hepatol 2008; 23: Centers for Disease Control and Prevention Website. International Classification of Diseases, Ninth Revision (ICD-9). icd9.htm. Updated September 1, Accessed November 29, Garcia-Tsao G, Boyer JL. Outpatient liver biopsy: how safe is it? Ann Intern Med 1993; 118: Friedman LS. Controversies in liver biopsy: who, where, when, how, why? Curr Gastroenterol Rep 2004; 6: Dam-Larsen S, Franzmann M, Andersen IB, et al. Long term prognosis of fatty liver: risk of chronic liver disease and death. Gut 2004; 53: Hassan C, Pickhardt PJ, Laghi A, et al. Computed tomographic colonography to screen for colorectal cancer, extracolonic cancer, and aortic aneurysm: model simulation with cost-effectiveness analysis. Arch Intern Med 2008; 168: Longo R, Pollesello P, Ricci C, et al. Proton MR spectroscopy in quantitative in vivo determination of fat content in human liver steatosis. J Magn Reson Imaging 1995; 5: Guiu B, Loffroy R, Cercueil JP, Krause D. Multiecho MR imaging and proton MR spectroscopy for liver fat quantification. Radiology 2008; 249: Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. Nonalcoholic steatohepatitis: a proposal for grading and staging the histological lesions. Am J Gastroenterol 1999; 94: Franzen LE, Ekstedt M, Kechagias S, Bodin L. Semiquantitative evaluation overestimates the degree of steatosis in liver biopsies: a comparison to stereological point counting. Mod Pathol 2005; 18: PQI Connect is the latest addition to the ARRS Website and serves as a source for information on meeting the growing demand for quality review programs in today s radiology practices and facilities. The interactive and easy-to-navigate site focuses on five critical topics that guide you through news items, relevant articles, and links to important information on each topic. 628 AJR:194, March 2010