Fat Containing HCC

Fat Containing HCC: Findings on CT and MRI Including Serial Contrast–Enhanced Imaging1 N. Cem Balci, MD, Alex S. Befeler, MD, B. Kirke Bieneman, MD, Rana Fattahi, MD, Sezer Saglam, MD, Necat Havlioglu, MD

Rationale and Objectives. The purpose of this article is to review the spectrum of computed tomography (CT) and magnetic resonance imaging (MRI) findings of fat containing hepatocellular carcinoma (HCC), including serial contrast–enhanced imaging. Materials and Methods. Imaging findings of 10 fat-containing HCCs on CT (n = 2) or MRI (n = 3) or on both CT and MRI (n = 5) were retrospectively reviewed in 9 patients. Both techniques included serial contrast enhanced imaging in arterial, portal venous, and late venous phases. Results. On non-contrast CT, fat containing HCC was either homogenously hypodense (n = 6) or of mixed density (n = 1). The density values ranged between 11 and 9 HU. On MRI, homogenous (n = 4) or heterogenous (n = 4) signal loss was observed on T1-weighted out-of-phase images as compared to in-phase images. Enhancement patterns on serial contrast–enhanced CT and MRI included: arterial enhancement indistinguishable from the liver with venous wash out (n = 2), arterial capillary blush with venous phase fading (n = 2), and heterogenous arterial enhancement with unenhanced foci and venous phase wash out of enhancements. Larger lesions had late capsular enhancement. Conclusions. Fat containing HCC has spectrum of imaging findings on CT and MRI. MRI with chemical shift technique depicts the fat content. Arterial contrast enhancement with venous washout or fading may help for the diagnosis of HCC in inconclusive cases. Key Words. MRI; CT; fat; HCC; contrast-enhanced study. ª AUR, 2009

Hepatocellular carcinoma (HCC) may rarely contain fat. Histopathologic features and imaging findings of fat containing HCC have been investigated (1,2). Fat content may be the initial finding of HCC during its early development from small hepatocellular regenerative and dysplastic nodules; therefore, its diagnosis and differentiation from other fat containing liver lesions is important (1). Imaging findings of fat containing HCC have been described on computed tomography (CT), ultrasound, and chemical shift magnetic resonance imaging (MRI) (2–4). The latter technique was Acad Radiol 2009; 16:963–968

shown to depict the fat content of HCC not recognized on CT images. Presently, patients are screened for focal liver lesions with the serial contrast–enhanced techniques that comprise arterial, portal venous, and equilibrium phases in both in MRI and CT (5,6). In this report, we retrospectively review spectrum of imaging findings of fat containing HCC on CT and MRI with the use of current imaging techniques that include multiphase contrast-enhanced imaging of the liver.

MATERIALS AND METHODS

1

From the Departments of Radiology (N.C.B., B.K.B., R.F.), Gastroenterology (A.S.B.), and Pathology (N.H.), Saint Louis University School of Medicine, 3635 Vista Ave, St. Louis, MO 63110; Department of Oncology, Istanbul Medical School, Istanbul, Turkey (S.S.). Received January 2, 2009; accepted February 2, 2009. Address correspondence to: N.C.B. e-mail: [email protected]

ª AUR, 2009 doi:10.1016/j.acra.2009.02.010

Patients This study was performed retrospectively and is Health Insurance Portability and Accountability Act–compliant. The institutional review board approved the study and waived the requirement for informed consent for our patient data review.

963

BALCI ET AL

Academic Radiology, Vol 16, No 8, August 2009

Figure 1. A 54-year-old patient with fat containing hepatocellular cancer demonstrated on magnetic resonance imaging. On T1-weighted in-phase spoiled gradient echo (SGE) (repetition time [TR]/echo time [TE]/flip angle [FA]; 145/4.3/70) (a), the lesion is isointense with the liver and cannot be demonstrated. A T1-weighted SGE (TR/TE/FA; 145/2.1/70) out-of-phase image (b) reveals homogenous signal loss of the lesion (arrow); on arterial phase image (SGE, TR/TE/FA; 145/2.1/70) (c), the lesion enhances more than the background liver (arrow) and contrast washes out during portal venous phase (arrow) (d).

Retrospective review of patients’ reports in the pathology department between August 2005 and August 2008 revealed 11 patients that had histopathologically confirmed HCC with fat content. In nine patients, the histopathology specimen consisted of collection of multiple true cut core biopsy material from the tumor; in one patient, the specimen was obtained in the autopsy procedure. One patient underwent partial liver resection and a resected liver lobe containing HCC was evaluated. All patients were cross-referenced with the imaging database of the hospital derived from picture archive and communication system (Synapse, Fujifilm USA). Nine patients (six males, three females; ages 29–72; mean, 58.2) were identified that underwent either CT (n = 2) or MRI (n = 3) or combined CT and MRI (n = 4) and had histopathology correlation within a 4-week interval. Patients with combined imaging examination underwent initially CT and MRI thereafter because of the inconclusive CT findings. Imaging MRI was performed at a 1.5 Tesla MR scanner (Intera, Philips Medical Systems, Bothell, WA) with the use of a four-element quadrature phased-array surface coil. A standard upper abdomen MRI protocol was used that con-

964

sisted of following imaging sequences and parameters: T1-weighted spoiled gradient echo in dual phase (repetition time [TR]/echo time [TE]/flip angle [FA]; 140–170 ms/4.4– 2.2 ms/70 ), and T2-weighted echo train fast spin echo (N/ 80 ms) with fat saturation. Gd-DTPA (Magnevist, Bayer, Wayne, NJ) was injected in a dose of 0.1 mmol/kg of body weight as a bolus injection at 2 mL/second using a power injector (Medrad, Indianola, PA). Images were acquired at 18 seconds (arterial dominant phase), 45 seconds (portal venous phase), and 90 seconds (late venous phase) after contrast administration. Lack of contrast enhancement in hepatic veins together with intense contrast enhancement of the aorta was accepted as correct timing for arterial phase images on MRI. For serial contrast–enhanced images T1-weighted, fatsaturated two-dimensional spoiled gradient echo sequence (TR/TE/FA; 160–180 ms/4.3 ms/70 [two patients]) or T1weighted spoiled gradient echo in dual phase (TR/TE/FA; 140–170 ms/4.4–2.2 ms/70 [five patients]) were used. All images were obtained in axial plane with 6-mm section thickness and 160–190  256 matrix with the Sensitivity encoding (SENSE) factor of two. CT examinations were performed with a multidetector CT scanner (Lightspeed plus GE Healthcare, General Electric

Academic Radiology, Vol 16, No 8, August 2009

CT AND MRI OF FAT CONTAINING HCC

Figure 2. Computed tomography (CT) and magnetic resonance imaging (MRI) of fat containing hepatocellular cancer in the posterior segment of the right liver lobe in a 65-year-old male. On arterial phase CT, the lesion reveals partial enhancement (long arrow) with skip areas of enhancement (short arrow) (a); on portal venous phase (b), the enhancement washes out (long arrow) and skip areas of enhancement (short arrow) remain. On MRI, the lesion is mixed hyperintense (arrow) and isointense on T1-weighted in-phase spoiled gradient (SGE) (repetition time [TR]/echo time [TE]/flip angle [FA]; 145/ 4.3/70) (c); in-phase hyperintense signal areas, loose signal on T1-weighted out-ofphase image (TR/TE/FA; 145/2.1/70) (arrow) is consistent with the fat content (d).

Medical Systems, Milwaukee, WI) equipped with four detector rows. Images were acquired through the liver in a craniocaudal direction using 2.5-mm detector collimation with a table speed per rotation of 15 mm/0.8 second, a pitch of 1.5, and an image section thickness of 5 mm for noncontrast, portal venous, and late venous phases. Arterial phase images were acquired with the same detector collimation, but with 7.5 mm/0.8 second table rotation and image section thickness of 2.5 mm. The scanning parameters used for all phases were 120 kVp and 250–350 mAs. After acquisition of unenhanced liver images, a contrast agent with an iodine concentration of 370 mg I/mL (Ultravist 370, Bayer, Wayne, NJ) was administered using a power injector (Medrad). The contrast agent was injected at a rate of 3 mL/second through an 18gauge plastic intravenous catheter placed in an antecubital vein with a total amount of 110–170 mL (2 mL/kg) according to patient’s body weight. Determination of the scanning delay for arterial phase imaging was achieved using an automatic bolus-tracking technique. After the trigger threshold was reached, arterial phase scanning began automatically. Dynamic imaging consisted of three phases (ie, arterial

phase, portal venous phase, and equilibrium phase). The mean scanning time delays of the arterial, portal venous, and late venous phase images were 25–30 seconds, 60–70 seconds, and 90–120 seconds, respectively. Image Analysis All images were loaded to the picture archive and communication system (Synapse) and evaluated by two experienced radiologists in consensus. All lesions were identified and measured in their longest diameter. In patients with multiple HCCs, the fat containing HCC was selected according to its biopsy or resected specimen localization. During image analysis, the following features of the HCC were described: lesion density on precontrast CT images, the signal intensity on T2-weighted fat saturated, and T1-weighted in phase and out of phase images. Signal differences between in-phase and out-of-phase T1-weighted images were recorded. On contrast-enhanced CT and MRI, the following enhancement features were assessed: the enhancement of the lesion in the arterial phase, presence of contrast wash out during portal venous and late venous

965

BALCI ET AL

Academic Radiology, Vol 16, No 8, August 2009

Figure 3. A 62-year-old female with small fat containing hepatocellular cancer in the anterior segment of the right liver lobe demonstrated on magnetic resonance imaging. On T1-weighted in-phase spoiled gradient echo (SGE) (repetition time [TR]/echo time [TE]/flip angle [FA]; 145/4.3/70) (a), the lesion is slightly hyperintense (arrow). A T1weighted SGE (TR/TE/FA; 145/2.1/70) out-of-phase image (b) reveals homogenous signal loss of the lesion (arrow); on arterial phase image (fat-saturated SGE, TR/TE/FA; 145/4.3/70) (c), the lesion becomes indistinguishable from the background liver (arrow) and contrast washes out during portal venous phase (arrow) (d).

phases, capsular enhancement in the portal venous and late venous phases, and skip areas of enhancement. RESULTS A total of 10 HCCs were identified on MRI and CT images. One patient had two fat-containing HCCs. Two HCCs were evaluated only on CT; three HCCs only on MRI. Five HCCs were assessed on both CT and MRI. The sizes of the HCCs ranged from 1.2 to 12 cm (mean, 4.75 cm). On noncontrast CT, six of the seven fat-containing HCCs were uniformly hypodense (3–9 Hounsfield units [HU], lesion size: 1.2–1.6 cm), or had mixed density with hypo- and isodense areas (lowest density 11 HU, lesion size: 12 cm) compared to the liver. On MRI, fat-containing HCCs revealed mild hyperintense (n = 5) or isointense (n = 3) signal compared to the liver on T2-weighted fat-saturated images. On T1-weighted in-phase images, the lesions were either uniformly hyper- or isointense compared to the liver or contained heterogenously mixed areas with iso-hyperintense and hypointense signal compared to the liver. On corresponding out-of-phase images, lesions with homogenous in-phase signal intensity revealed homogenous signal loss,

966

lesions with mixed in phase signal intensity showed foci of signal loss that corresponded to the higher signal areas on in phase images. Contrast-enhanced images both on CT and MRI revealed the following imaging findings: arterial phase enhancement, lesion becoming indistinguishable from the background liver, and contrast washout during venous phases (n = 2); arterial capillary blush that faded during venous phases (n = 2); heterogeneous arterial enhancement with skip foci of enhancement and contrast washout of arterial enhancements (n = 6); and late capsular enhancement (n = 4) (Fig. 1–3). The frequency of findings is summarized in Table 1. DISCUSSION The results of this study showed a broad spectrum of imaging findings of fat containing HCC on CT and MRI. The imaging findings of fat containing HCC were previously described. A study investigated the CT findings by including patients that were diagnosed with HCC, which had CT density of less than 10 HU (2). In this study, we did not predefine a selection criteria on imaging and reported the imaging findings on patients with histopathologically

CT AND MRI OF FAT CONTAINING HCC

Academic Radiology, Vol 16, No 8, August 2009

Table 1 Spectrum of Imaging Findings of Fat Containing Hepatocellular Cancer Lesions

CT Findings

MRI Findings

Lesion 1

Homogenously hypodense (3 HU)

Uniformly iso/hyperintense I/P*homogenous signal loss O/P** on T1, isointense on T2 Uniformly iso/hyperintense I/P homogenous signal loss O/P on T1, mildly hyperintense on T2 Uniformly iso/hyperintense I/P homogenous signal loss O/P on T1, isointense on T2

Lesion 2

Homogenously hypodense (4 HU)

Lesion 3

Homogenously hypodense (8 HU)

Lesion 4

Mixed hypo-isodense (lowest-11 HU)

Heterogeneously mixed signal I/P, hetrogenous signal loss O/P, mildly hyperintense on T2

Lesion 5

Homogenously hypodense (9 HU)

Uniformly iso/hyperintense I/P homogenous signal loss O/P on T1, isointense on T2

Lesion 6

Homogenously hypodense (6 HU)

Lesion 7

Heterogeneously mixed signal I/P, heterogenous signal loss O/P, mildly hyperintense on T2

Lesion 8

Heterogeneously mixed signal I/P, hetrogenous signal loss O/P, mildly hyperintense on T2

Lesion 9

Heterogeneously mixed signal I/P, hetrogenous signal loss O/P, mildly hyperintense on T2

Lesion 10

Homogenously hypodense (6 HU)

Enhancement

Size

Arterial enhancement indistinguishable from liver parenchyma with venous phase washout Arterial capillary blush that fades during venous phase

1,4 cm

Heterogeneous arterial enhancement with unnenhanced foci and venous phase contrast washout of arterial enhancements Heterogeneous arterial enhancement with unnenhanced foci and venous phase contrast washout of arterial enhancements and late capsular enhancement Arterial capillary blush that fades during venous phase Arterial enhancement indistinguishable from liver parenchyma with venous phase washout Heterogeneous arterial enhancement with unnenhanced foci and venous phase contrast washout of arterial enhancements and late capsular enhancement Heterogeneous arterial enhancement with unnenhanced foci and venous phase contrast washout of arterial enhancements Heterogeneous arterial enhancement with unnenhanced foci and venous phase contrast washout of arterial enhantments Heterogeneous arterial enhancement with unnenhanced foci and venous phase contrast washout of arterial enhancements and late capsular enhancement

1,6 cm

1,5 cm

12 cm

1,6 cm

1,2 cm

2,4 cm

2,6 cm

2,1 cm

2,4 cm

I/P*, in phase; O/P**, out of phase.

diagnosed HCC. This also explains why lesions with higher density values than fat existed in our database. Fat metamorphosis in HCC occurs in the early stages of HCC being homogenously dispersed in the lesion or in advanced stages of growth in which this process may become a lipomatous component of the tumor (1,2). Fatty metamorphosis and fibrous capsule are reported to occur more frequently in the Asian population as compared to non-Asians in imaging and pathology literature. In our patient population, we also observed limited number of lesions with fatty metamorphosis within a large number of HCCs in our institutions database,

which correlates with the reported rarity of this entity in nonAsians (7). The reason why the density values of smaller sized HCCs resembled cysts on CT in our cases can be explained by uniform composition of fat and tumor and partial volume effect during density measurement (8). MRI revealed homogenous out-of-phase signal loss in all small-sized HCCs and determined the fat content. In a comparative study for the diagnosis of fat containing HCC on chemical shift MRI, the superior diagnostic ability of this technique was shown over CT in cases with inconclusive density values (3). In four

967

BALCI ET AL

cases with five fat-containing HCCs in our database, MRI was superior to demonstrate the fat content of the lesion, especially in lesions with inconclusive density values on CT. MRI was also able to assess the enhancement pattern on serial contrast–enhanced images when fat suppression techniques were used. Demonstration of fat content on out-of-phase images as a single finding cannot be called HCC because a variety of liver lesions may have fat content (9,10). Therefore, the enhancement pattern of the fat containing HCC needs to be evaluated. The enhancement of fat containing HCC has been described on CT (2). According to that report, areas of fat metamorphosis do not enhance despite of overall arterial enhancement of the tumor. On the other hand, intratumoral arterioles are present in fat containing HCC that may increase in number and density when the tumor reaches the size of 1.6–3.0 cm (1). We also observed arterial enhancement of the HCCs in our patient population with skip areas of enhancement in both small- and large-sized lesions. HCC enhances more than the background liver in the arterial phase images because of its rich blood supply (11). Two of our cases revealed arterial enhancement of the lesion that became indistinguishable from the background liver. This diminished arterial enhancement pattern may be explained with the lack of enhancement of the homogenously dispersed fat content causing perception of overall lesser degree of arterial enhancement. In such cases, venous phase washout of contrast favored the diagnosis of HCC. The contrast washout has been considered characteristic of the arterial phase enhancement for the diagnosis of HCC. Rich arterial blood supply of HCC is the cause of its rapid contrast blush and clearance. Incorrect timing may miss the capillary blush, but the imaging window of contrast washout is longer and less effected by contrast administration timing (12). In some of our cases, arterial enhancement of the fat containing HCC faded during the venous phases. This may occur in HCC and dysplastic nodules and with the presence of fat content, hepatocellular adenoma, and—less frequently— focal nodular hyperplasia (FNH) can be taken in consideration for the differential diagnosis (10). A focal fat may be a differential diagnosis than fat containing HCC, but negligible enhancement of the focal fat can rule out HCC. Follow-up examination is recommended in arterial enhancing small (<2 cm) nodules that do not reveal washout to differentiate between HCC and benign lesions (13). In greater sized HCCs, we observed heterogeneous enhancement pattern with skip areas of enhancement and contrast washout of arterial enhancements. Greater sized HCCs may mimic other fat containing lesions such as angiomyolipoma and hepatocellular adenoma. Angiomyolipoma reveals its peak enhancement at a later time than HCC and can be distinguished. Adenomas enhance arterially and fade during the venous phase (ie, becoming equally enhanced compared

968

Academic Radiology, Vol 16, No 8, August 2009

to the liver parenchyma) (10). Capsular enhancement is a helpful imaging feature of HCC. The fibrous capsule of HCC becomes enhanced on venous phases, reaching its peak at the equilibrium enhancement phase (10). The fibrous capsule is more frequently seen in lesions grater than 2 cm, as was observed in our cases. The limitations of our study were the limited number of cases to better categorize the spectrum of findings and compare between imaging modalities and the absence of sectional histopathologic analysis of the entire lesion to show corresponding imaging findings of the tumor composition. In conclusion, spectrum of imaging findings of fat containing HCC was demonstrated on CT and MRI. The fat content was consistently discernible on chemical shift MRI, although inconclusive density values on CT may exist. Partial or uniform arterial enhancement was observed in all cases, with venous phase contrast washout of arterial enhancements in the majority and fading in fewer lesions. Greater lesions revealed characteristic late capsular enhancement.

REFERENCES 1. Kutami R, Nakashima Y, Nakashima O, et al. Pathomorphologic study on the mechanism of fatty change in small hepatocellular carcinoma of humans. J Hepatol 2000; 33:282–289. 2. Yoshikawa J, Matsui O, Takashima T, et al. Fatty metamorphosis in hepatocellular carcinoma: radiologic features in 10 cases. Am J Roentgenol AJR 1988; 151:717–720. 3. Martı´n J, Sentı´s M, Zidan A, et al. Fatty metamorphosis of hepatocellular carcinoma: detection with chemical shift gradient-echo MR imaging. Radiology 1995; 195:125–130. 4. Kanno T, Kurioka N, Kim S, et al. Implications of hyperechoic lesions in small hepatocellular carcinoma. Gastroenterol Jpn 1989; 24:528–534. 5. Lauenstein TC, Salman K, Morreira R, et al. Gadolinium-enhanced MRI for tumor surveillance before liver transplantation: center-based experience. Am J Roentgenol AJR 2007; 189:663–670. 6. Valls C, Cos M, Figueras J, et al. Pretransplantation diagnosis and staging of hepatocellular carcinoma in patients with cirrhosis: value of dual-phase helical CT. Am J Roentgenol AJR 2004; 182:1011–1017. 7. Freeny PC, Baron RL, Teefey SA. Hepatocellular carcinoma: reduced frequency of typical findings with dynamic contrast-enhanced CT in a non-Asian population. Radiology 1992; 182:143–148. 8. Burgener FA, Hamlin DJ. Contrast enhancement of focal hepatic lesions in CT: effect of size and histology. Am J Roentgenol AJR 1983; 140: 297–301. 9. Prasad SR, Wang H, Rosas H, et al. Fat-containing lesions of the liver: radiologic-pathologic correlation. Radiographics 2005; 25:321–331. 10. Hussain SM, Zondervan PE, Izermans JN, et al. Benign versus malignant hepatic nodules: MR imaging findings with pathologic correlation. Radiographics 2002; 22:1023–1036. 11. Tajima T, Honda H, Taguchi K, et al. Sequential hemodynamic change in hepatocellular carcinoma and dysplastic nodules: CT angiography and pathologic correlation. Am J Roentgenol AJR 2002; 178:885–897. 12. 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; 234:460–467. 13. Jeong YY, Mitchell DG, Kamishima T. Small (<20 mm) enhancing hepatic nodules seen on arterial phase MR imaging of the cirrhotic liver: clinical implications. Am J Roentgenol AJR 2002; 178:1327–1334.