Quantitatively Defining Washout in Hepatocellular Carcinoma

Transcription

1 Gastrointestinal Imaging Original Research Kamaya et al. Washout in Hepatocellular Carcinoma Gastrointestinal Imaging Original Research Yueyi I. Liu 1 Lewis K. Shin 1,2 R. Brooke Jeffrey 1 Aya Kamaya 1 Liu YI, Shin LK, Jeffrey RB, Kamaya A Keywords: delayed phase, equilibrium phase, hepatocellular carcinoma, hypervascular liver lesion, multiphase CT, washout DOI: /AJR Received May 3, 2011; accepted after revision May 21, Department of Radiology, Stanford University Medical Center, 300 Pasteur Dr, H1307, Stanford CA Address correspondence to A. Kamaya (kamaya@stanford.edu). 2 Department of Radiology, VA Palo Alto Health Care System, Palo Alto, CA. CME/SAM This article is available for CME/SAM credit. AJR 2013; 200: X/13/ American Roentgen Ray Society Quantitatively Defining Washout in Hepatocellular Carcinoma OBJECTIVE. Washout on delayed phase (or equilibrium phase) imaging of an arterially hyperenhancing lesion is an excellent predictor of hepatocellular carcinoma (HCC). The purpose of our study was to quantitatively define washout in pathologically proven HCC. A quantitative definition of HCC may minimize interobserver variability and facilitate more accurate diagnosis. MATERIALS AND METHODS. We identified 47 liver lesions that were hyperenhancing in the arterial phase from 24 patients who underwent triphasic MDCT as part of preoperative evaluation for liver transplantation. All HCCs were pathologically proven. Regions of interest were obtained of lesions and areas of adjacent liver on arterial, portal venous, and delayed phase images. Enhancement profiles were assessed by three radiologists. RESULTS. Of the 47 hypervascular lesions, 14 HCCs were identified. There was a statistically significant difference in percentage attenuation ratio (defined as 100 ratio of attenuation of adjacent liver to that of the lesion) between lesions that were HCC (median percentage attenuation ratio, 121) and those that were not (median percentage attenuation ratio, 101) on delayed phase. Percentage attenuation ratio 107 on delayed phase imaging achieved maximal sensitivity (100%) with good specificity (75.8%), positive predictive value (PPV) (63.6%), and negative predictive value (NPV) (100%) in HCC detection. Percentage attenuation ratio also correlated well with radiologists assessments of enhancement profiles of lesions (multinomial logistic regression McFadden R 2, 0.72; chi-square p, < 0.01). CONCLUSION. Our analysis of simple CT attenuation measurements indicates that percentage attenuation ratio offers excellent sensitivity, specificity, PPV, and NPV for HCC detection and very good correlation with radiologists assessments of washout. H epatocellular carcinoma (HCC) is the third leading cause of cancerrelated death in the world, with annual deaths of more than 500,000 worldwide [1]. Diagnosis of HCC depends heavily on imaging characteristics. The appearance of HCC on contrast-enhanced triphasic (i.e., late arterial, portal venous, and delayed or equilibrium phase) MDCT has been well described [2 7]. The classic HCC is hyperdense compared with background liver in the late arterial phase, isodense or hypodense in the portal venous phase, and hypodense on 3-minute delayed phase images. This enhancement difference is thought to be related to the differences in blood supply. HCCs receive blood primarily from the hepatic arteries and therefore tend to enhance more avidly than background liver during late arterial phase imaging, whereas the normal liver receives approximately 25% of its blood sup- ply from the hepatic artery and 75% from the portal vein [8, 9]. During the portal venous phase of contrast enhancement, the liver continues to enhance, and the lack of portal venous blood supply to HCCs results in the characteristic washout in the portal venous phase and especially in the delayed phase. In addition, the liver parenchyma is often fibrotic in patients with HCC. Because fibrotic tissues are known to wash in and wash out slowly, this may also contribute to the hypodense appearance of HCCs in the portal venous and delayed phases [10]. The criteria of hypervascularity in the arterial phase followed by washout in the delayed phase offer good sensitivity (64 89%), specificity (96%), and PPV (93%) in HCC detection [11 13]. Imaging diagnosis of HCC in patients on the transplant waiting list is of utmost importance because it affects the appropriate allocation of the small number of livers that be- 84 AJR:200, January 2013

2 Washout in Hepatocellular Carcinoma come available. Retrospective studies of the database of the United Network for Organ Sharing (UNOS) and Organ Procurement and Transplantation Network (OPTN) showed unacceptably inaccurate performance of radiologic staging criteria at that time, which included only hypervascularity on the arterial phase ( vascular blush corresponding with the area of suspicion ), compared with final pathologic staging [14]. As a result, a national conference was held to address specific HCC issues related to liver allocation and a new draft policy was proposed [15]. Acting on this new policy, clinical trials, such as the American College of Radiology Imaging Network (ACRIN) trial to evaluate CT and MRI for detecting HCC [16], are currently underway to improve our ability to detect HCC and reduce false-positive findings. Meanwhile, both the American Association for the Study of Liver Diseases (AASLD) and UNOS have updated their policy to include washout as part of their criteria for imaging diagnosis of HCC. The most recent recommendations by AASLD state that a diagnosis of HCC can be made on the basis of imaging alone if a nodule > 1 cm shows typical features of HCC (hypervascularity in the arterial phase and washout in the early or delayed venous phase) on contrast-enhanced CT or MRI [17]. Similarly, UNOS has also recently started using hypervascularity on the arterial phase and washout on the portal venous or delayed phase as part of its criteria for imaging diagnosis of HCC [18]. Interestingly, the washout phenomenon has never been quantitatively defined for HCC and relies solely on the radiologist s subjective visual assessment. However, similar temporal enhancement changes are quantitatively measured to differentiate adrenal adenomas from carcinomas using either absolute percentage washout or relative percentage washout. Absolute percentage washout is defined as 100 ([EA DA] / [EA PA]), where EA is attenuation on contrast-enhanced scans, DA is attenuation on delayed contrast-enhanced scans, and PA is unenhanced attenuation. Relative percentage washout is defined as 100 (EA DA) / EA). Adrenal adenomas typically exhibit rapid washout, i.e., absolute percentage washout > 60% or relative percentage washout > 40%. Adrenocortical carcinomas, on the other hand, typically have relative percentage washout < 40% [19]. The purposes of this study were to quantitatively define washout, assess the accuracy of quantitative formulas for washout, and correlate these findings with radiologists impressions of washout in small hypervascular liver lesions, with pathologic correlation. Materials and Methods Patients Institutional review board approval was obtained for this retrospective study and the requirement of informed consent was waived. We reviewed 229 consecutive liver transplantations from 220 patients who underwent orthotopic liver transplantation at our institution from 2004 to Seventy-five patients who underwent a dedicated triphasic MDCT liver protocol within 6 months of transplantation were considered. Of the remaining patients in our study, 106 had no CT within 6 months of the transplantation. Because ours is a large referral center, many patients had their workup done elsewhere before coming to our institution, and 39 patients did not have triphasic CT performed. Some patients underwent biphasic CT instead of triphasic CT; others underwent only unenhanced CT because of renal impairment. Lesion Selection Among the 75 patients who underwent a triphasic MDCT liver protocol within 6 months of transplantation, 44 had hypervascular lesions identified on arterial phase images. We excluded 12 patients who had only lesions that had been previously chemoembolized and three patients who had only small lesions (< 1 cm in the longest axis). In addition, five patients were excluded because their hypervascular lesions were too difficult to correlate with the final pathology report. Forty-seven hypervascular lesions from 24 patients (20 men, four women; mean age, 53 years; age range, years) fit our criteria and formed the study group. Indications for liver transplantation included hepatitis C in 11 patients, hepatitis B in six patients, cryptogenic cirrhosis in three patients, alcoholic cirrhosis in two patients, primary sclerosing cholangitis in one patient, and cryptogenic fibrosis (without cirrhosis) in one patient. Pathology Explanted livers were serially sectioned at 5- to 10-mm slices and examined for focal lesions after being fixed in formalin. Only lesions with clear histopathologic correlation by location and distance to anatomic landmarks were included. Five HCCs were excluded because of lack of clear indication of location in the histopathologic report. CT Triphasic Liver Technique All studies were performed with 8-, 16-, or 64- MDCT scanners (LightSpeed Ultra and Light- Speed 16, GE Healthcare) and Somatom Sensation 64 (Siemens Healthcare). Typically, 120 ml (or lower for patients with low weight) of IV contrast material (iohexol 350 mg/ml) was power injected at a rate of 4 5 ml/s followed by a saline chaser. No oral contrast material was administered. Bolus-tracking software (SmartPrep, GE Healthcare) was used, and arterial phase, portal venous phase, and delayed phase scans were obtained at seconds, seconds, and 3 minutes after initiation of contrast administration, respectively. The scanning parameters were 2.5-mm section collimation, 3.0-mm effective section thickness, and 3.0-mm reconstruction interval. The kilovoltage was set at 120. Median volume CT dose index (CTDI vol ) was 53.7 mgy and median dose-length product (DLP) was 1592 mgy cm for the 19 patients for whom these data were available. Image Analysis Objective attenuation measurements Circular or oval regions of interest (ROIs) were manually drawn by one radiologist over the lesion and two areas of adjacent liver on arterial phase, portal venous phase, and delayed phase images, and Hounsfield units were measured. Care was taken to avoid any blood vessels, ducts, cysts, or artifacts. The CT numbers of the two ROIs drawn over the adjacent liver were averaged. Percentage attenuation ratio (PAR), attenuation change (AC), and relative washout ratio (RWR) on the portal venous phase or delayed phase were calculated as follows: PAR PV / D = 100 (AA PV / D / LA PV / D ), AC PV / D = (AA PV / D LA PV / D ), and RWR PV/D = 100 (LA A LA PV / D ) / LA A, where AA PV / D is the average of the two areas adjacent to the lesion on portal venous (PV) or delayed phase (D) and LA A / PV / D is the attenuation of the lesion on arterial (A), portal venous (PV), or delayed phase (D). All attenuation measurements are in Hounsfield units. Subjective radiologist evaluation All images were independently reviewed by three board-certified radiologists specializing in abdominal imaging. Results of the subjective radiologist evaluation are published in a recent article [20]. In the current study, we correlated PAR values with radiologists subjective evaluation. Statistical Analysis The two-tailed Student t test was used to compare the mean of groups of attenuation values. Fleiss kappa was used to evaluate interobserver agreement. Standard binomial ROC curves were generated for percentage attenuation ratio, attenuation change, and relative washout ratio using maximum likelihood estimation with the Rockit 1.1B software (Kurt Rossmann Laboratories for Radiologic Image Research). We used McFadden AJR:200, January

3 Kamaya et al. R 2 test, which is a type of multinomial logistic regression, to correlate PAR (a continuous variable) to lesion enhancement profile (a categoric variable). All calculations were performed using the statistical programming language R (R Project for Statistical Computing). A Results Forty-seven hypervascular lesions in 24 patients were identified on arterial phase images. Average lesion size was 1.8 cm (range, cm; median, 1.5 cm). Twelve patients had one lesion, seven patients had two lesions, three patients had three lesions, one patient had five lesions, and one patient had seven lesions. Fourteen of the 47 lesions were HCC on the basis of pathologic examination of the explanted liver. Figure 1 shows a representative arterially hypervascular lesion with clear washout, best visualized on the delayed phase, that was proven to be HCC. Figure 2 shows the distribution of attenuation values of HCCs and other hypervascular lesions (non-hcc) and their adjacent areas in the arterial, portal venous, and delayed phases. There was a significant difference between the attenuation values of HCC and non-hcc lesions in the portal venous phase (Student t test, p < 0.01; difference, 21.8; 95% CI, 33.7 to 9.9) and delayed phase (Student t test, p < 0.01; difference, 19.9; 95% CI, 28.6 to 11.2) but not in the arterial phase (Student t test, p < 0.19; difference, 8.5; 95% CI, 22.2 to 5.2). To better delineate HCC washout, we calculated the percentage attenuation ratio for each of the lesions on both portal venous and delayed phase (Fig. 3). There was no significant difference between these values for HCC and other hypervascular lesions on the portal venous phase (p < 0.14; difference, 0.069; 95% CI, to 0.165), but a significant difference was found on delayed phase (p < 0.01; difference, 0.22; 95% CI, ). Similarly, there was also a significant difference between the attenuation change of adjacent liver and lesion for HCC and other hypervascular lesions on the delayed phases (p < 0.01; difference, 18.6; 95% CI, ). Percentage attenuation ratio and attenuation change offer essentially the same sensitivity and specificity in differentiating HCC from other hypervascular lesions, except at the sensitivity level of 85.7% (when there were 12 true-positive HCCs and two false-negative ones), where percentage attenuation ratio has slightly higher true-negative findings (28 vs 27), resulting in a slightly higher specificity value (84.8% vs 81.9%). Figure 4 shows the ROC curve for the detection of HCC using either percentage attenuation ratio or attenuation change. Relative washout ratio was also calculated for each lesion. The relative washout ratio is analogous to lesion deenhancement used in characterizing adrenal masses. However, it did not perform as well as percentage attenuation ratio or attenuation change in the identification of HCC lesions (Fig. 4). To maximize lesion detection, we chose a percentage attenuation ratio threshold of 107 on the delayed phase, which offered sensitivity B Fig. 1 Hepatocellular carcinoma in 49-year-old woman. A C, CT images show tumor (arrow, A) is hyperdense compared with background liver on arterial phase (A), isodense to slightly hypodense on portal venous phase (B), and hypodense on delayed phase (C). C 86 AJR:200, January 2013

4 Washout in Hepatocellular Carcinoma of 100%, specificity of 75.8%, PPV of 63.6%, and NPV of 100% for HCC detection. Percentage attenuation ratio also correlated well with radiologists assessments of the enhancement profile of liver lesions. No statistically significant difference was seen among the ratios of hyperdense lesions on all phases (Student t test, p = 0.08), between isodense lesions on portal venous and delayed phases (Student t test, p = 0.07); or between hypodense lesions on portal venous or delayed phases (Student t test, p = 0.11). As a result, we combined the three phases and evaluated the correlation between percentage attenuation ratio and radiologists assessments of enhancement profiles regardless of phase. They were strongly correlated (McFadden R 2, 0.72; chi-square test, p < 0.01) (Fig. 5). Regardless of imaging phase, percentage attenuation ratio values of 113 or higher have a 50% or more likelihood to be deemed hypodense by radiologists, whereas percentage attenuation ratio values of 86 or lower have a 50% or more likelihood of being deemed hyperdense by radiologists. With a sensitivity of 100%, a percentage attenuation ratio threshold of 107 is more sensitive at HCC detection than radiologists subjective evaluation. In the same set of 14 HCCs, radiologists deemed 13 lesions to be hypodense (i.e., washout) on the delayed phase, resulting in sensitivity of 92.9%. One HCC was deemed isodense on both portal venous and delayed phases [20]. On the other hand, radiologists had fewer false-positive findings and better PPV for the presence of HCC (92.9% vs 63.6% for percentage attenuation ratio). Arterial Portal Venous Delayed HCC Liver HCC Liver HCC Liver HCC Liver HCC Liver HCC Discussion Differentiating HCC from other common hypervascular liver lesions, such as arterioportal shunts or regenerative nodules is critical because accurate early identification allows early intervention and improved long-term patient survival. Moreover, accurate diagnosis of HCC in patients on transplant waiting lists is of utmost importance to allow appropriate allocation of the small number of livers that become available. Although the term washout has been adopted by major societies, such as UNOS and AASLD [17], to identify HCC without biopsy confirmation, the term has never been defined quantitatively. As a result, washout detection depends on many factors, including the interpreting radiologist s subjective visual assessment and level of experience as well as window and level settings that often vary between radiologists. A quantitative definition of washout would be helpful in improving the confidence in HCC diagnosis as well as facilitating more accurate diagnosis. Interobserver variability may be minimized by a quantitative definition. It may also help less-experienced radiologists or referring physicians understand the concept of washout. Lastly, a quantitative definition can help lay the foundation for future computer-aided programs that may use automated liver lesion detection or characterization algorithms. Our analysis of simple CT attenuation measurements indicates that a threshold percentage attenuation ratio of 107, which corresponds to parenchymal attenuation at least 7% higher than the lesion, offers 100% sensitivity with excellent PPV for HCC detection and correlates well with radiologists assessments of washout. Although percentage attenuation ratio calculations showed excellent results in our study, the radiologists subjective assessment for the presence of HCC is still critical in the overall diagnosis of HCC, as evidenced by a higher PPV. For instance, diffuse background liver alteration, such as hepatic steatosis, can skew percentage attenuation ratio calculations but radiologists may intuitively reconcile these changes when evaluating focal liver lesions. Other radiographic features of HCC, such as a delayed enhancing pseudocapsule, are not necessarily reflected in percentage attenuation ratio calculations. Percentage attenuation ratio values therefore can be used as a valuable quantitative tool to assist radiologists in overall assessment of lesions with questionable visible washout on delayed images in combination with other imaging features that are characteristic of HCC. Washout has previously been applied to differentiate adrenal adenomas from carcinomas. For adrenal masses, washout is defined as deenhancement of the lesion from contrast-enhanced to delayed phase by a certain percentage; this concept is analogous to the relative washout rate in our study. However, we found that deenhancement is not helpful in liver lesions and should not be applied in evaluation for HCC. Rather, direct comparison of the lesion with adjacent liver parenchyma, or percentage attenuation ratio, is better at predicting HCC. Furthermore, percentage attenuation ratio correlated with radiologists assessments of the enhancement profiles of the lesions. Region-of-interest-based calculations have previously been applied to detect liver metastasis [21]. Kim et al. [22] used quantitative CT color mapping of the arterial enhancement Liver Attenuation (HU) 200 Fig. 2 Boxplot shows distribution of attenuation values of hepatocellular carcinoma (HCC) and other hypervascular lesions (non-hcc) and their adjacent areas in arterial, portal venous, and delayed phases. AJR:200, January

5 Kamaya et al. Percentage Attenuation Ratio HCC HCC Portal Venous Delayed Fig. 3 Boxplot shows distribution of percentage attenuation ratio of hepatocellular carcinoma (HCC) and other hypervascular lesions (non-hcc) on portal venous and delayed phases. Horizontal line marks percentage attenuation ratio of 107. fraction to detect HCC and reported sensitivities of 88.8%, comparable to our results. In their study, however, they used unenhanced (U), arterial (A), and portal venous (P) phase images and did not obtain or analyze delayed phase images. They defined arterial enhancement fraction as 100 (HU A HU U ) / (HU P HU U ). Our study differs from Kim et al. in that we were specifically interested in delayed phase images in addition to the arterial and portal venous phase images in lesion characterization. In fact, our study found that delayed phase percentage attenuation ratio values were significantly more useful than percentage attenuation ratio values in the portal venous phase, highlighting the improved diagnostic yield of routinely obtaining delayed phase images in HCC evaluation. We did not use unenhanced images in our analysis because prior studies have shown them to be of limited value in lesion detection and characterization [23, 24]. Therefore, we believe that our study more accurately reflects current standard-of-care CT protocols for HCC diagnosis while minimizing unnecessary radiation. There are several limitations to our study. First, our study has a relatively small number of lesions. Although we started with 229 consecutive liver transplantations, only 47 lesions fit our selection criteria. In the future, we hope to validate our results with a larger dataset. Second, our results depend on where the ROIs are drawn. Two lesions in our study were heterogeneous in enhancement: One was pathologically confirmed to represent a mixed HCC and cholangiocarcinoma, and the other was found to Sensitivity Specificity Fig. 4 Graph shows receiver operating characteristic curve for hepatocellular carcinoma detection using either percentage attenuation ratio (red), attenuation change (green), or relative washout ratio (blue). represent a focally hemorrhagic HCC. In these cases, ROIs were drawn over the most hyperdense area in the lesion on the arterial phase and care was taken to place the ROIs in the same location on the corresponding portal venous and delayed phase images. Another difficulty in ROI selection is in placing the ROI on the background liver, which potentially can be heterogeneous due to fatty infiltration or cirrhosis. We recommend selecting two ROIs adjacent to the lesion in the same lobe of the liver and avoiding large blood vessels or areas of ductal dilatation that can skew the liver attenuation measurements. Another limitation of our study is that the threshold we identified is only applicable to hypervascular lesions seen on CT. Some HCCs, especially small or well-differentiated lesions, may not be hyperenhancing on arterial phase images, in which case our criteria may not be applicable. In a study by Monzawa et al. [24], only 58% of HCCs < 2 cm were hypervascular on the arterial phase [24]. Conversely, a minority of hypervascular HCCs may not exhibit washout on portal venous or delayed phase images. In the same study by Monzawa et al., 88% of well-differentiated HCCs and 93% of moderately to poorly differentiated HCCs showed washout on delayed phase images [24]. Finally, we relied on retrospective review of pathology reports to determine the presence of HCCs. As a result, HCCs smaller than the histologic sectioning size (1 cm) could potentially have been missed by the pathologists. The retrospective nature of the pathology review also limited identification of most of the hypervascular liver lesions that were not HCC. Predicted Probability for Each Category Hypodense Isodense Hyperdense Percentage Attenuation Ratio Fig. 5 Multinomial logistic regression chart shows correlation between percentage attenuation ratio and radiologists assessments of enhancement profiles. Vertical lines mark percentage attenuation ratio of 86 and 113. As a result, we do not have an alternative truth diagnosis for these lesions. Future studies with thinner slice sectioning of explanted livers may be needed for this purpose. In conclusion, the presence of washout in arterially hyperenhancing hepatic lesions in patients with risk factors for HCC is a strong predictor of the presence of HCC. We have quantitatively defined a threshold value for washout (percentage attenuation ratio, 107) that correlates well with pathologically proven diagnoses of HCC in explanted livers. Acknowledgments We thank Erik Mittra for assistance with the statistical analysis and Jeslyn A. Rumbold for editorial assistance. References 1. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, CA Cancer J Clin 2005; 55: Baron RL, Brancatelli G. Computed tomographic imaging of hepatocellular carcinoma. Gastroenterology 2004; 127(5 suppl 1):S133 S Hollett MD, Jeffrey RB Jr, Nino-Murcia M, Jorgensen MJ, Harris DP. Dual-phase helical CT of the liver: value of arterial phase scans in the detection of small 1.5 cm) malignant hepatic neoplasms. AJR 1995; 164: Hwang GJ, Kim MJ, Yoo HS, Lee JT. Nodular hepatocellular carcinomas: detection with arterial-, portal-, and delayed-phase images at spiral CT. Radiology 1997; 202: Kim T, Murakami T, Takahashi S, et al. Optimal phases of dynamic CT for detecting hepatocellular carcinoma: evaluation of unenhanced and triple- 88 AJR:200, January 2013

6 Washout in Hepatocellular Carcinoma phase images. Abdom Imaging 1999; 24: Laghi A, Iannaccone R, Rossi P, et al. Hepatocellular carcinoma: detection with triple-phase multi-detector row helical CT in patients with chronic hepatitis. Radiology 2003; 226: Mitsuzaki K, Yamashita Y, Ogata I, Nishiharu T, Urata J, Takahashi M. Multiple-phase helical CT of the liver for detecting small hepatomas in patients with liver cirrhosis: contrast-injection protocol and optimal timing. AJR 1996; 167: Quiroga S, Sebastia C, Pallisa E, Castella E, Perez- Lafuente M, Alvarez-Castells A. Improved diagnosis of hepatic perfusion disorders: value of hepatic arterial phase imaging during helical CT. Radio- Graphics 2001; 21:65 81; questionnaire, Fournier LS, Cuenod CA, de Bazelaire C, et al. Early modifications of hepatic perfusion measured by functional CT in a rat model of hepatocellular carcinoma using a blood pool contrast agent. Eur Radiol 2004; 14: Kamaya A, Maturen KE, Tye GA, Liu YI, Parti NN, Desser TS. Hypervascular liver lesions. Semin Ultrasound CT MR 2009; 30: Ronzoni A, Artioli D, Scardina R, et al. Role of MDCT in the diagnosis of hepatocellular carcinoma in patients with cirrhosis undergoing orthotopic liver transplantation. AJR 2007; 189: Valls C, Andia E, Sanchez A, Moreno V. Selective use of low-osmolality contrast media in computed tomography. Eur Radiol 2003; 13: Marrero JA, Hussain HK, Nghiem HV, Umar R, Fontana RJ, Lok AS. Improving the prediction of hepatocellular carcinoma in cirrhotic patients with an arterially-enhancing liver mass. Liver Transpl 2005; 11: Freeman RB Jr. Transplantation for hepatocellular carcinoma: the Milan criteria and beyond. Liver Transpl 2006; 12(11 suppl 2):S8 S Pomfret EA, Washburn K, Wald C, et al. Report of a national conference on liver allocation in patients with hepatocellular carcinoma in the United States. Liver Transpl 2010; 16: ACRIN Website. Protocol ACRIN www. acrin.org/tabid/679/default.aspx. Accessed August 15, Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. Hepatology 2011; 53: Organ Procurement and Transplantation Network. Allocation of livers. optn.transplant.hrsa. gov/policiesandbylaws2/policies/pdfs/policy_8. pdf. Accessed August 15, Johnson PT, Horton KM, Fishman EK. Adrenal mass imaging with multidetector CT: pathologic conditions, pearls, and pitfalls. RadioGraphics 2009; 29: Liu Y, Kamaya A, Jeffrey RB, Shin LK. Multidetector computed tomography triphasic evaluation of the liver before transplantation: importance of equilibrium phase washout and morphology for characterizing hypervascular lesions. J Comput Assist Tomogr 2012; 36: Platt JF, Francis IR, Ellis JH, Reige KA. Liver metastases: early detection based on abnormal contrast material enhancement at dual-phase helical CT. Radiology 1997; 205: Kim KW, Lee JM, Klotz E, et al. Quantitative CT color mapping of the arterial enhancement fraction of the liver to detect hepatocellular carcinoma. Radiology 2009; 250: Iannaccone R, Laghi A, Catalano C, et al. Hepatocellular carcinoma: role of unenhanced and delayed phase multi-detector row helical CT in patients with cirrhosis. Radiology 2005; 234: Monzawa S, Ichikawa T, Nakajima H, Kitanaka Y, Omata K, Araki T. Dynamic CT for detecting small hepatocellular carcinoma: usefulness of delayed phase imaging. AJR 2007; 188: FOR YOUR INFORMATION This article is available for CME/SAM credit. Log onto click on AJR (in the blue Publications box); click on the article name; add the article to the cart; proceed through the checkout process. AJR:200, January