Gastrointestinal Imaging • Original Research Agarwal et al. Inflammatory Hepatocellular Adenomas Can Mimic FNH on MRI

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Gastrointestinal Imaging Original Research

Inflammatory Hepatocellular Adenomas Can Mimic Focal Nodular Hyperplasia on Gadoxetic Acid–Enhanced MRI Sheela Agarwal1 Jorge M. Fuentes-Orrego1 Thomas Arnason2 Joseph Misdraji 3 Kartik S. Jhaveri 4 Mukesh Harisinghani1 Peter F. Hahn1 Agarwal S, Fuentes-Orrego JM, Arnason T, et al.

Keywords: contrast-enhanced MRI, gadoxetate disodium, inflammatory hepatocellular adenoma, liver imaging DOI:10.2214/AJR.13.12251 Received November 15, 2013; accepted after revision March 10, 2014. 1 Division of Abdominal Imaging, Department of Radiology, Massachusetts General Hospital and Harvard Medical School, 55 Fruit St, Boston, MA 02114. Address correspondence to S. Agarwal ([email protected]). 2 Division of Anatomical Pathology, Queen Elizabeth II Health Sciences Centre and Dalhousie University, Halifax, NS, Canada.  3 Department of Pathology, Massachusetts General Hospital and Harvard Medical School, Boston, MA 4 Department of Medical Imaging, University of Toronto, University Health Network and Mount Sinai Hospital, Toronto, ON, Canada. 

WEB This is a web exclusive article. AJR 2014; 203:W408–W414 0361–803X/14/2034–W408 © American Roentgen Ray Society

W408

OBJECTIVE. Inflammatory hepatocellular adenoma (HCA) is a recently categorized entity of hepatocellular neoplasms. We investigated whether gadoxetic acid–enhanced MRI can distinguish inflammatory HCA from focal nodular hyperplasia (FNH). MATERIALS AND METHODS. From January 1, 2009, through January 1, 2013, gadoxetic acid–enhanced MRI examinations from two institutions were reviewed for HCA, with specific histologic features of inflammatory HCA. Biopsy and resection slides were reviewed, and immunohistochemistry for glutamine synthetase was performed in a subset to confirm the initial diagnosis. RESULTS. A total of 10 possible cases of inflammatory HCA were identified in the pathology database. On the basis of glutamine synthetase staining performed for this study, three cases were rediagnosed as FNH and thus were excluded from the study. Therefore, a total of seven patients with inflammatory HCA were identified. On gadoxetic acid–enhanced MRI, four of these patients had classic features of FNH (group A, FNH mimics), and three had imaging features suggestive of HCA (group B, typical inflammatory HCA). Imaging features that were considered diagnostic of FNH included isointense or minimal T2 hyperintensity, arterial enhancement, and diffuse hyperintensity on hepatobiliary phase. Three of the four patients with FNH mimics had slides available for pathologic rereview, and the diagnosis of inflammatory HCA was supported by glutamine synthetase immunohistochemistry findings. The pathology reports of the remaining four cases were rereviewed and were also found to have features consistent with inflammatory HCA. CONCLUSION. Inflammatory HCA can mimic FNH on MRI, including hepatobiliary phase hyperintensity. Moreover, conventional pathology using histopathology alone may lead to misclassification of inflammatory HCA.

H

epatocellular adenoma (HCA) is a relatively uncommon benign hepatic neoplasm, typically found in middle-aged women [1]. These lesions can be difficult to differentiate from focal nodular hyperplasia (FNH), which is frequently seen in women of similar age [2]. Distinguishing FNH from HCA is clinically relevant, because FNHs can be managed conservatively whereas HCAs are usually surgically resected given the risk of rupture, hemorrhage, and malignant transformation [3, 4]. In the last decade, significant advances in the fields of both imaging and pathology have markedly increased our diagnostic capabilities for differentiating FNH from HCA. Within pathology, this can be largely attributed to the understanding of the genetic makeup of the subtypes of HCA and novel

immunohistochemistry stains to differentiate them [5, 6]. Within imaging, the availability and wide adoption of hepatocyte-specific contrast agents, particularly gadoxetic acid, have refined the imaging diagnosis of FNH and HCA [7]. Recently, a number of studies have shown the ability of gadoxetic acid–enhanced liver MRI to differentiate between HCA and FNH with nearly 100% accuracy [7, 8]. FNH appears hyperintense compared with surrounding liver parenchyma on the hepatobiliary phase because of the presence of functioning lesional hepatocytes but abnormal blindending biliary ductules. HCA appears hypointense because of poorly functioning lesional hepatocytes. Although there are rare reports of HCA appearing isointense to surrounding liver on the hepatobiliary phase in the setting of fatty liver, to our knowledge,

AJR:203, October 2014

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Inflammatory Hepatocellular Adenomas Can Mimic FNH on MRI there are no other published reports of HCA appearing isointense to hyperintense on hepatobiliary phase in the setting of normal background liver parenchyma. In this study, we focus on a particular subtype of HCA, the inflammatory HCA, also known as telangiectatic adenoma [9, 10]. This entity has pathologic features of both FNH and HCA and has only recently been reclassified as an adenoma by the World Health Organization [11, 12] (Fig. 1). Because of these mixed features, confident diagnosis of this entity is very difficult by conventional histopathologic methods. Differentiating inflammatory HCA from FNH is clinically relevant given the risk of hemorrhage in inflammatory HCA and, rarely, malignant transformation to hepatocellular carcinoma [13–15]. We hypothesized that, given the overlap of histopathologic features of inflammatory HCA and FNH, the MRI characteristics might similarly overlap.

Fig. 1—Schematic representation of spectrum of histologic features commonly seen with benign hepatocellular lesions. Adapted with permission from [12]. A–D, Classic focal nodular hyperplasia is at one of end of spectrum (A), and classic (nontelangiectatic) hepatocellular adenoma (HCA) is at other end both with rare ductules (C) and without ductules (D). Inflammatory HCA, with overlapping histologic features of both, is seen in middle (B). Gray areas represent fibrosis, and red and green dots correspond to arterioles and bile ducts, respectively.

Materials and Methods This retrospective study was approved by the institutional review board of the Massachusetts General Hospital and the University of Toronto and was conducted in accordance with HIPAA. Because it was retrospective, the institutional review boards of both institutions waived the requirement for patient informed consent.

A

B

C

D

E

F

Case Identification The pathology records of two university hospitals were searched for the term “hepatic adenoma” from 2009 to 2013, the period during which gadoxetic acid was used. Twenty-six patients with a pathologic diagnosis of HCA were identified, 14 (54%) of whom were noted in the pathology reports to have histologic features that were diagnostic of or indeterminate for the inflammatory or telangiectatic subtype of adenoma. Four of the cases were received in consultation from outside hospitals and did not have corresponding imaging. The remaining 10 patients all had a gadoxetic acid–enhanced liver MRI available for review.

Pathology Methods Seven of the 10 cases had been diagnosed on core biopsy and three had had a resection. Eight had a pathologic diagnosis of inflammatory HCA and two were reported as having features indeterminate for FNH versus inflammatory HCA. Pathology reports were reviewed for all 10 cases, and archived slides and formalin-fixed paraffinembedded tissue were available and reviewed in six of these cases, including both lesions originally labeled as indeterminate for inflammatory

Fig. 2—44-year-old woman who underwent surgical resection for two lesions. A–F, One lesion (A–C) had MRI features of focal nodular hyperplasia (FNH), and other lesion (D–F) had MRI features of hepatocellular adenoma (HCA). On basis of histologic appearance of surgical specimens (B and E), both lesions were initially diagnosed as inflammatory HCA because of presence of stroma surrounding blood vessels expanded with inflammation, dilated sinusoids, and lack of fibrous septa typical of FNH. Hepatobiliary phase imaging (A and D) shows marked difference in appearance of two lesions, first (asterisk, A) appearing hyperintense on hepatobiliary phase compared with liver parenchyma, and second (arrow, D) appearing diffusely hypointense. However, on staining with glutamine synthetase for purposes of this study, first lesion was reclassified as FNH. Glutamine synthetase staining (×20, C) shows “maplike” pattern of immunopositivity typical of FNH. Glutamine synthetase immunohistochemistry in HCA (×40, F) lacks maplike immunopositivity characteristic of FNH.

AJR:203, October 2014 W409

Agarwal et al. TABLE 1: Clinical Characteristics of Seven Women With Inflammatory Hepatocellular Adenoma

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Characteristic

Value

Age (y), median (range)

43 (25–50)

Body mass index (kg/m2), median (range)

34.7 (22–38.1)

Asymptomatic presentation

7 (100)

Lesion maximum diameter (cm), median (range)

2.96 (1.45–6.91)

Oral contraceptive use

1 (14)

Alcohol intakea

4 (57)

Hepatic steatosis

3 (43)

Abnormal liver function tests

0 (0)

Note—Except where noted otherwise, data are number (%) of patients. aDefined as at least four drinks per month.

HCA versus FNH. H and E slides were reviewed by two pathologists for histologic features of inflammatory HCA, including thin-walled small arteries without accompanying veins, subtle ductular reaction, lymphocytic inflammation, and dilated sinusoids [16]. The slides were also assessed to exclude histologic features of FNH, including central scarring, nodular architecture, and fibrous septa–containing abnormal vessels [17]. In the six cases with available tissue, a glutamine synthetase immunohistochemical stain was applied to 5-μm sections of formalin-fixed paraffinembedded tissue using an automated slide stainer (Bond III, Leica) with Bond Polymer DAB Detection kits (Leica). After antigen retrieval in Epitope Retrieval Solution 1 (Leica) for 20 minutes, the glutamine synthetase antibody (clone GS-6, Millipore Billerica) was applied at a dilution of 1:8000. Glutamine synthetase–stained slides were assessed to identify patterns consistent with HCA, which include the absence of staining in hepatocytes (apart from surrounding intratumoral veins) or, alternatively, diffuse immunopositivity, which is typical of β-catenin-mutated tumors. This pattern of staining contrasts to the broad anastomosing or “maplike” pattern of glutamine synthetase staining associated with FNH [5, 18] (Fig. 2). On pathology review, the two tumors that were originally reported as indeterminate for inflammatory HCA versus FNH were determined to be most consistent with FNH on the basis of H and E features and a maplike pattern of glutamine synthetase staining considered characteristic of FNH. Additionally, one of the cases originally reported as inflammatory HCA had subtle histologic features suggestive of FNH, and glutamine synthetase stain revealed a maplike pattern, resulting in a change of diagnosis to FNH (Fig. 2). Thus, these three cases were excluded from the imaging review. The other four cases originally diagnosed as HCA were all histologically consistent with the di-

W410

agnosis of inflammatory HCA and had glutamine synthetase expression supporting that diagnosis. The histologic descriptions of the three cases without available slides or tissue were reevaluated and were also compatible with inflammatory HCA. These seven cases of inflammatory HCA formed the study population for detailed imaging review; for each patient, one lesion was pathologically diagnosed as already described and served as the index lesion. In addition, the seven patients had another 17 lesions among them that were presumed to represent inflammatory HCAs because they had imaging features identical to those of the biopsied lesion. The seven patients were all women (age range, 25–50 years). Six patients were clinically obese, with body mass index greater than 30.0 kg/m2. Clinical information was also collected including oral contraceptive use, liver function tests, and serum inflammatory markers. Details of the clinical data are outlined in Table 1.

MRI Acquisition All patients underwent MRI of the liver with a 1.5-T system (Magnetom Avanto, Siemens Health-

care). All patients were positioned supine with a phased-array receiver coil covering the upper abdomen. The sequences were obtained by applying parallel imaging sensitivity encoding with a reduction factor of 2. The image acquisition protocol is summarized in Table 2. T1-weighted gradient-echo volumetric interpolated breath-hold examination sequences were performed before, during, and after IV administration of 10 mL of gadoxetic acid (automatic injection rate of 1 mL/s followed by 20-mL saline flush) by using an automated power injector. Pulse sequence parameters remained unchanged from the unenhanced to contrast-enhanced acquisitions. The arterial, portal venous, and late dynamic phases were acquired after the administration of contrast agent at 25–35, 70, and 180 seconds, respectively. The hepatobiliary phase was obtained 20 minutes after the administration of contrast agent. In addition, the acquisition protocol encompassed turbo spin-echo T2-weighted imaging with fat saturation and 3D gradient-recalled echo in-phase and out-of-phase technique.

MRI Analysis All MRI examinations were evaluated by three abdominal radiologists (with 7, 15, and 28 years of experience). All hepatic lesions were qualitatively evaluated for signal intensity (SI) on unenhanced T1- and T2-weighted images compared with the adjacent parenchyma (hypointense, isointense, or hyperintense); the enhancement pattern in arterial, portal venous, and hepatobiliary phases (none, mildly intense, moderately intense, and intense); and the homogeneity of the enhancement. Other morphologic features were also noted, including the presence of fat (SI decrease on out-of-phase imaging), T2 hyperintense central scar, capsule (peripheral T1 hypointensity on contrast-enhanced imaging), or hemorrhage (regions of high SI on T1-weighted images) [19]. Lesion number, size, and location were also documented. Whenever dis-

TABLE 2: MRI Protocol

Sequence

TR/TE

Flip Angle (°)

Section Thickness (mm)

T1-weighted 3D (volumetric interpolated breath-hold examination)

3.27/1.19

25

3.5

300 × 400

256 × 154

T1-weighted in phase (3D gradient-echo)

8.232/4.168

12

5.0

400

320 × 192

T1-weighted out of phase (3D gradient-echo)

8.232/2.084

12

5.0

400

320 × 192

90

5.0

420

512 × 256

Fast spin-echo T2-weighted 1.125 × 10 4 /85.89 with fat saturation

FOV (mm)

Matrix

Note—Volumetric interpolated breath-hold examination sequence was performed before and after the administration of contrast agent.

AJR:203, October 2014

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Inflammatory Hepatocellular Adenomas Can Mimic FNH on MRI agreement was recorded, the readers reviewed the images and reached consensus. The quantitative analysis was performed by one radiologist (with 7 years of experience) on a diagnostic PACS workstation (Impax RS 3000 review station, AGFA Technical Imaging Systems). The SI was quantified in the liver lesions, background liver parenchyma, and paraspinal muscles both before and after the administration of contrast agent on T1-weighted volumetric interpolated breath-hold examination images. A circular ROI (350–450 mm2) was placed in the liver, avoiding vessels, liver lesions, and artifacts. The lesions were evaluated at the same axial level drawing the largest possible ROI. All ROIs were carefully placed to be similar in size and in location on each phase of the dynamic acquisition. The lesions’ ROIs ranged in size from 61 to 469 mm2 (mean, 354 mm2). We restricted our quantitative analysis to one lesion per patient with histopathologic correlation. The contrast enhancement ratio was calculated by applying the following formula: (SILe − SINc) / SILe × 100, where SILe represents the SI of the lesion on enhanced images and SINc represents the SI of the unenhanced liver lesion. In addition, the lesion-to-liver contrast ratio was also analyzed for each tumor during the arterial, portal venous, and hepatobiliary phases, respectively, by using the following formula: (SILe − SILiv) / SIMs, where SILiv and SIMs represent the SI in the enhanced liver parenchyma and paraspinal muscles, respectively. The SI of the paraspinal muscles was used for normalization of lesion-to-liver contrast ratios; background noise could not be used because of the application of parallel imaging.

Statistical Analysis For the purpose of this study, all lesions were classified into two groups—FNH mimics (isointense to hyperintense on hepatobiliary phase, group A) and typical inflammatory HCAs (hypointense on hepatobiliary phase, group B)—to evaluate for overlapping features between FNH and HCA, respectively. The qualitative features of each of the tumors were recorded and compared using the Fisher exact test. The median contrast enhancement ratio and lesion-to-liver contrast ratio were recorded and qualitatively compared for each of the dynamic phases between groups A and B. In addition, the contrast enhancement ratio and lesion-to-liver contrast ratio of the two groups of lesions were compared with established contrast enhancement ratio and lesion-to-liver contrast ratio for FNH and HCA from published data [7, 20]. All the data analysis was performed using statistical software (JMP Pro version 10.0, SAS Institute).

Results Qualitative Analysis Seven patients (with a total of 24 lesions) with inflammatory HCA were identified who also underwent gadoxetic acid–enhanced MRI. Four patients (11 lesions) were found by all three radiologists to have classic diagnostic features of FNH, including minimal T2 hyperintensity, central scar, homogeneous enhancement during the late arterial phase, and diffuse hyperintensity on the hepatobiliary phase. From the remaining three patients, one patient (eight lesions) had features compatible with HCA, and two patients (five lesions) had equivocal features, including heterogeneous arterial enhancement, a subtle pseudocapsule during the portal venous phase, and hypointensity during the hepatobiliary phase. On the basis of the readers’ qualitative assessment, four patients (11 lesions) formed group A, and the remaining three patients (13 lesions) formed group B (Figs. 3 and 4). Statistically significant differences were found between both groups A and B for MRI findings, such as homogeneous enhancement during the arterial phase, hyperintensity on T2-weighted images, presence of central scar, and pseudocapsule, respectively, as described in Table 3 (p < 0.05). Quantitative Analysis Our quantitative analysis supported readers’ categorization of lesions. Characteristic trends were observed between both lesion groups for the magnitude of contrast enhancement and uptake pattern (lesion-to-liver contrast ratio) during the portal venous and hepatobiliary phases, as shown in Figures 5 and 6. The difference in lesion-to-liver contrast ratio between the two groups was greatest during the hepatobiliary phase (lesionto-liver contrast, 0.01 vs −2.41, for groups

A and B, respectively). The contrast enhancement ratio and lesion-to-liver contrast curves of lesions in group A mirrored those reported for FNH; group B lesions displayed imaging characteristics for HCA as described in prior studies [7, 20, 21] (Figs. 4 and 5). Discussion As the use of hepatocyte-specific contrast agents becomes more widespread, there is an increased reliance on imaging for noninvasive differentiation of FNH from HCA [7, 8]. Experience at two institutions shows that, although noninvasive differentiation is reliable for most patients, this may not be the case for inflammatory HCA. In our series, we identified four patients who had imaging features overlapping with those of FNH, as shown by the contrast enhancement ratio and lesionto-liver contrast ratio curves (Figs. 5 and 6), including during the hepatobiliary phase, because these lesions became isointense to hyperintense relative to the background liver parenchyma. Indeed, our results provide corroboration with authors who have reported isolated cases of HCAs that showed hyperintensity during the hepatobiliary phase after injection of hepatocyte-specific contrast agents [7, 22, 23]. These imaging features of inflammatory HCA are closely related to the histologic composition, as stated by Laumonier et al. [24], who postulated that sinusoidal dilatation, in addition to the presence of numerous arteries and cavities, could play a role in explaining the MRI features of inflammatory HCA. These features of inflammatory HCAs are thought to reduce the blood flow, accounting for the hyperintensity seen on T2-weighted images. Furthermore, we theorize that the strong enhancement seen during the arterial phase could be inherently related to the presence of large arteries, whereas the hyperintensity observed during the hepatobiliary

TABLE 3: Imaging Features of Inflammatory Hepatocellular Adenoma (HCA) on MRI for Seven Women With 24 Lesions Group A, FNH Mimics (n = 11 Lesions)

Group B, Typical Inflammatory HCA (n = 13 Lesions)

p

11 (100)

4 (31)

0.0006

T2 hyperintensity

0 (0)

9 (69)

0.0006

MRI Finding Homogeneous enhancement Central scar

7 (64)

0 (0)

0.001

Pseudocapsule

0 (0)

13 (100)

0.0001

Hyperintensity on hepatobiliary phase

8 (73)

0 (0)

0.0002

Note—Data are number (%) of lesions. FNH = focal nodular hyperplasia.

AJR:203, October 2014 W411

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Agarwal et al.

A

B

C

D

E

F

Fig. 3—Examples of inflammatory hepatocellular adenoma (HCA) on T1-weighted hepatobiliary phase images in six different patients. A–C, Three lesions show hyperintensity (arrows) on hepatobiliary phase and represent group A, focal nodular hyperplasia mimics. D–F, Three lesions show hypointensity (asterisks) on hepatobiliary phase and represent group B, typical inflammatory HCA.

phase could be the result of reduced blood flow due to large cavities and sinusoidal dilatation, as described by van Aalten et al. [25]. Another factor that plays a major role in hepatobiliary uptake is the expression of the organic anionic transport protein (OATP1B1 and OATP1B3). It is known that OATP1 mediates the uptake of gadoxetic acid into the hepatocyte. As a result, hepatobiliary phase SI is inherently related to the level of expression of transporter proteins within each lesion. Vander Borght et

W412

al. [26] studied diverse types of focal liver lesions in the setting of normal background liver parenchyma and found that the distribution and expression of transporter proteins among FNH and inflammatory HCA varies. In our study, despite the similarities seen between hyperintense inflammatory HCA and FNH regarding contrast dynamics, we found that hyperintense inflammatory HCAs often showed slightly increased SI on T2-weighted imaging. This might provide an imaging feature for discriminating

between FNH and inflammatory HCA; similar observations have been described recently by Grazioli et al. [27]. Given the overlapping features found on MRI between FNH and inflammatory HCA, there is a need to rely on additional information to distinguish these entities. We think that incorporation of clinical information can help toward this end. For instance, in the past, the development of HCAs has been linked to use of oral contraceptive pills and alcohol [28]. However, recently various au-

AJR:203, October 2014

Fig. 4—47-year-old woman. Core needle biopsy of group A focal nodular hyperplasia (FNH) mimic lesion is shown (see Fig. 3C for hepatobiliary phase image). This adenoma shows same histologic features as group B, typical inflammatory hepatocellular adenoma lesions. A, H and E stain (×40) shows dilated sinusoids and absence of fibrous septa. B, Glutamine synthetase immunohistochemistry (×40) shows absence of maplike immunopositivity characteristic of FNH.

A

B 1.5

140 120

CER Group A (FNH mimics)

100

CER of FNH*

80 60

×

40

× ×

20 0

CER Group B (Typical inflammatory HCA)

Arterial Phase

×

CER of HCA*

Portal Hepatobiliary Venous Phase Phase

1 0.5

LLC Group A (FNH Mimics)

×

0

×

−0.5

LLC of FNH*

×

LLC Group B (Typical inflammatory HCA)

−1

×

−1.5

LLC of HCA*

−2 −2.5

Fig. 5—Median contrast enhancement ratios (CERs) of focal nodular hyperplasia (FNH) mimics (group A) and typical inflammatory hepatocellular adenoma (HCA) (group B), as compared with previously published CERs of FNH and HCA during arterial, portal venous, and hepatobiliary phases. Asterisks denote previously published values [7, 20].

thors have reported an increased occurrence of HCA and particularly inflammatory HCA in female patients in the setting of obesity and metabolic syndrome [10, 29]. Indeed, in our cohort, all except one patient were either overweight (n = 1) or obese (n = 5). This finding was also correlated with the fact that 43% had hepatic steatosis. These data are important because female overweight or obese patients with inflammatory HCA are being acknowledged as a new clinical entity. Obesity under these circumstances is worrisome, because it has been postulated that activation of interleukin-6 pathway might produce malignant transformation in these patients [10, 30]. Although the number of lesions presented in this series is small, the percentage of inflammatory HCAs (57%) with an imaging appearance mimicking FNH should raise concern that these lesions are being misdiagnosed. One reason why this may not have been apparent until recently is because these lesions, presumed to be FNH, were not sampled because of a low index of suspicion.

Liver-to-Lesion Contrast Ratio

160 Contrast Enhancement Ratio (%)

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Inflammatory Hepatocellular Adenomas Can Mimic FNH on MRI

Arterial Phase

Portal Hepatobiliary Venous Phase Phase

Fig. 6—Median temporal lesion-to-liver contrast (LLC) ratios for focal nodular hyperplasia (FNH) mimics (group A), typical inflammatory hepatocellular adenoma (HCA) (group B) lesions, and published LLC ratios of FNH and HCA after administration of gadoxetic acid during arterial, portal venous, and hepatobiliary phases. Asterisks denote previously published values [7].

Moreover, our pathology review reveals that misclassification of these lesions by pathologists is another potential source of radiologic and pathologic disagreement in classifying these lesions. Of seven cases identified with pathologic features of inflammatory HCA for which slides and tissue were available for pathology review, only four were confirmed as inflammatory HCA after independent slide review and application of glutamine synthetase immunohistochemistry. The other three cases were determined to be most consistent with FNH. Two of these FNH cases were originally reported as having pathologic features indeterminate for HCA or FNH, but the third case was originally reported as inflammatory HCA on the basis of evaluation of the complete surgical resection specimen (Fig. 2). This finding highlights the difficulty of accurately separating FNH from inflammatory HCA by pathologic criteria, particularly on small biopsy specimens and without supporting immunohistochemistry. The two entities can have overlapping histologic features, partic-

ularly when bleeding and necrosis occur in inflammatory HCA, causing secondary fibrosis that mimics the fibrotic septa of FNH [30]. The addition of glutamine synthetase immunohistochemistry, introduced only recently as a clinical tool, has been shown to increase pathologist accuracy for separating these two entities [5, 18]. The results of our study highlight the need for wider use and adoption of glutamine synthetase staining. The major limitation of our study is its small sample size. The second limitation is the small volume of tissue that was available for reexamination with new immunohistochemistry stains. Ideally, all of our cases would have confirmation with glutamine synthetase instead of only cases with questionable imaging features. We speculate that more widespread awareness of this potential clinical conundrum will lead to moreaggressive management. This, in turn, will produce larger series that can help to define the scope of the problem. Until additional studies are conducted to confirm these results, the use of gadoxe-

AJR:203, October 2014 W413

Downloaded from www.ajronline.org by East Carolina University on 10/27/14 from IP address 150.216.68.200. Copyright ARRS. For personal use only; all rights reserved

Agarwal et al. tic acid–enhanced MRI to diagnose FNH should be considered as a complementary tool to clinical information, including risk factors and associations with inflammatory HCA (obesity, high alcohol intake, anicteric cholestasis, systemic inflammatory syndrome, fatty liver, use of oral contraceptive pills, and other HCAs in the liver). Although we cannot make firm recommendations with regard to which lesions to biopsy given our small sample size, our findings serve to alert radiologists to this problem. Radiologists and their referring surgeons and gastroenterologists will have to decide in individual cases whether to pursue biopsy. We caution, however, that pathologists should be prepared to use glutamine synthetase staining on their histologic material; otherwise, even conventional histologic analysis can be misleading. In conclusion, inflammatory HCAs can mimic FNH both histologically and on MRI, including uptake of gadoxetic acid during the hepatobiliary phase. However, factoring of clinical information, such as obesity, alcohol intake, and use of oral contraceptive pills, may help to correctly distinguish them. Furthermore, in light of the overlapping histologic features between FNH and inflammatory HCA, the addition of glutamine synthetase staining should become standard to characterize lesions as inflammatory HCA. Larger multiinstitutional studies are needed to confirm the results presented here and to further evaluate the clinical significance of the lesions that are found to mimic FNH on hepatobiliary phase imaging. References 1. Rooks JB, Ory HW, Ishak KG, et al. Epidemiology of hepatocellular adenoma: the role of oral contraceptive use. JAMA 1979; 242:644–648 2. Miller FH, Hammond N, Siddiqi AJ, et al. Utility of diffusion-weighted MRI on distinguishing benign and malignant hepatic lesions. J Magn Reson Imaging 2010; 32:138–147 3. Cherqui D, Rahmouni A, Charlotte F, et al. Management of focal nodular hyperplasia and hepatocellular adenoma in young women: a series of 41 patients with clinical, radiological and pathological correlations. Hepatology 1995; 22:1674–1681 4. Farges O, Ferreira N, Dokmak S, Belghiti J, Bedossa P, Paradis V. Changing trends in malignant transformation of hepatocellular adenoma. Gut 2011; 60:85–89 5. Bioulac-Sage P, Cubel G, Taouli S, et al. Immuno-

W414

histochemical markers on needle biopsies are helpful for the diagnosis of focal nodular hyperplasia and hepatocellular adenoma subtypes. Am J Surg Pathol 2012; 36:1691–1699 6. Rebouissou S, Bioulac-Sage P, Zucman-Rossi J. Molecular pathogenesis of focal nodular hyperplasia and hepatocellular adenoma. J Hepatol 2008; 48:163–170 7. Grazioli L, Bondioni MP, Haradome H, et al. Hepatocellular adenoma and focal nodular hyperplasia: value of gadoxetic acid-enhanced MR imaging in differential diagnosis. Radiology 2012; 262:520–529 8. Bieze M, van den Esschert JW, Nio CY, et al. Diagnostic accuracy of MRI in differentiating hepatocellular adenoma from focal nodular hyperplasia: prospective study of the value of gadoxetate disodium. AJR 2012; 199:26–34 9. Bioulac-Sage P, Cubel G, Zucman-Rossi J. Subtype classification of hepatocellular adenoma. Dig Surg 2010; 27:39–45 10. Paradis V, Champault A, Ronot M, et al. Telangiectatic adenoma: an entity associated with increased body mass index and inflammation. Hepatology 2007; 46:140–146 11. Attal P, Vilgrain V, Brancatelli G, et al. Telangiectatic focal nodular hyperplasia: US, CT, MR imaging findings with histopathologic correlation in 13 cases. Radiology 2003; 228:465–472 12. Walther Z, Jain D. Molecular pathology of hepatic neoplasms: classification and clinical significance. Pathol Res Int 2011; 2011:403929 13. Hechtman JF, Raoufi M, Fiel MI, et al. Hepatocellular carcinoma arising in a pigmented telangiectatic adenoma with nuclear β-catenin and glutamine synthetase positivity: case report and review of the literature. Am J Surg Pathol 2011; 35:927–932 14. Stoot JH, Coelen RJ, De Jong MC, Dejong CH. Malignant transformation of hepatocellular adenomas into hepatocellular carcinomas: a systemic review including more than 1600 adenoma cases. HBP (Oxford) 2010; 12:509–522 15. Dokmak S, Paradis V, Vilgrain V, et al. A singlecenter surgical experience of 122 patients with single and multiple hepatocellular adenomas. Gastroenterology 2009; 137:1698–1705 16. Burt AD, Portmann BC, Ferrell LD. MacSween’s pathology of the liver, 6th ed. Philadelphia, PA: Churchill Livingston Elsevier, 2012 17. Shafizadeh N, Kakar S. Diagnosis of well differentiated hepatocellular lesions: role of immunohistochemistry and any other ancillary techniques. Adv Anat Pathol 2011; 18:438–445 18. Bioulac-Sage P, Taouji S, Le Bail B, Possenti L, Balabaud C. Value and limits of routine

histology alone or combined with glutamine synthetase immunostaining in the diagnosis of hepatocellular adenoma subtypes on surgical specimens. Int J Hepatol 2013; 2013:417323 19. Hussain SM, van den Bos IC, Dwarkasing RS, Kuiper JW, den Hollander J. Hepatocellular adenoma: findings at state-of-the-art magnetic resonance imaging, ultrasound, computed tomography and pathologic analysis. Eur Radiol 2006; 16:1873–1886 20. Gupta RT, Iseman CM, Leyendecker JM, Shyknevsky I, Merkle EM, Taouli B. Diagnosis of focal nodular hyperplasia with MRI: multicenter retrospective study comparing gadobenate dimeglumine to gadoxetate disodium. AJR 2012; 199:35–43 21. Denecke T, Steffen IG, Agarwal S, et al. Appearance of hepatocellular adenomas on gadoxetic acid-enhanced MRI. Eur Radiol 2012; 22:1769–1775 22. Ringe KI, Husarik DB, Sirlin CB, Merkle EM. Gadoxetate disodium–enhanced MRI of the liver. Part 1. Protocol optimization and lesion appearance in the noncirrhotic liver. AJR 2010; 195:13–28 23. Giovanoli O, Heim M, Terraciano L, Bongartz G, Ledermann HP. MRI of hepatic adenomatosis: initial observations with gadoxetic contrast agent in three patients. AJR 2008; 190:[web]W290–W293 24. Laumonier H, Bioulac-Sage P, Laurent C, Zucman-Rossi J, Balabaud C, Trillaud H. Hepatocellular adenomas: magnetic resonance imaging features as a function of molecular pathological classification. Hepatology 2008; 48:808–818 25. van Aalten SM, Thomeer MGJ, Terkivatan T, et al. Hepatocellular adenomas: correlation of MR imaging findings with pathologic subtype classification. Radiology 2011; 261:172–181 26. Vander Borght S, Libbrecht L, Blokzijl H, et al. Diagnostic and pathogenetic implications of the expression of hepatic transporters in focal lesions occurring in normal liver. J Pathol 2005; 207:471–482 27. Grazioli L, Olivetti L, Mazza G, Bondioni MP. MR imaging of hepatocellular adenomas and diagnosis dilemma. Int J Hepatol 2013; 2013:374170 28. Bioulac-Sage P, Laumonier H, Laurent C, Zucman-Rossi J, Balabaud C. Hepatocellular adenoma: what is new in 2008. Hepatol Int 2008; 2:316–321 29. Bioulac-Sage P, Taouji S, Possenti L, Balabaud C. Hepatocellular adenoma subtypes: the impact of overweight and obesity. Liver Int 2012; 32:1217–1221 30. Bioulac-Sage P, Sempoux C, Possenti L, et al. Pathological diagnosis of hepatocellular cellular adenoma according to the clinical context. Int J Hepatol 2013; 2013:253261

AJR:203, October 2014

Inflammatory hepatocellular adenomas can mimic focal nodular hyperplasia on gadoxetic acid-enhanced MRI.

Inflammatory hepatocellular adenoma (HCA) is a recently categorized entity of hepatocellular neoplasms. We investigated whether gadoxetic acid-enhance...
804KB Sizes 0 Downloads 3 Views