FDG PET/CT in the liver: lesions mimickingmalignancies

Gerald Jit Shen Tan,1,2,3 Salvatore Ugo Berlangieri,3,4 Sze Ting Lee,3,4

Andrew Mark Scott3,4

1Department of Diagnostic Radiology, Tan Tock Seng Hospital, 11 Jalan Tan Tock Seng, Singapore 308433, Singapore2Lee Kong Chian School of Medicine, Nanyang Technological University, 11 Mandalay Road, Singapore 308232, Singapore3Centre for PET, Austin Health, Heidelberg, VIC, Australia4Ludwig Institute for Cancer Research, 145 Studley Road, Heidelberg, VIC 3084, Australia

Abstract

Purpose: 18F-fluorodeoxyglucose (FDG) PET/CT isinvaluable in managing liver lesions, in particular in theevaluation of suspected liver metastases. It is bothsensitive and specific in detecting liver metastases froma wide range of primary cancers, and may change clinicalmanagement, most commonly by detecting additionallesions and decreasing the number of futile surgeries.However, some benign lesions may also show increasedmetabolic activity which can lead to false positive PETfindings. We describe some of these lesions and theirimaging characteristics that may help in differentiatingthem from malignant metastases.Methods: We reviewed all whole body FDG PET/CTstudies performed over a 5-year period in our institution,and identified those with focal liver lesions showingincreased FDG uptake for which histological resultswere available.Results: A majority of lesions showing increased meta-bolic activity were due to malignant disease, such asmetastases or primary liver tumours. However, we alsofound increased FDG uptake in non-neoplastic lesionssuch as Cryptococcosis, abscesses, and secondary inflam-mation from cholecystitis. Increased metabolic activitywas also seen in some benign neoplasms such as hepaticadenomas and hemangioendotheliomas.Conclusion: FDG PET/CT is currently the most sensitivenon-invasive imaging modality for the detection ofhepatic metastases, particularly from the gastrointestinaltract. False positive results are rare, and have beendescribed mainly in abscesses. However, other lesionscan also show increased metabolic activity, and failure to

differentiate these from metastases may result in inap-propriate treatment.

Key words: FDG—PET/CT—Liver—Tumour—Metastases—Abscess

18F-fluorodeoxyglucose (FDG) positron emissiontomography/computed tomography (PET/CT) has pro-ven invaluable in modern oncologic imaging. A commonindication is for the detection of metastases, which typ-ically exhibit increased glucose metabolism and take upFDG as a glucose analogue [1]. The liver is a commonsite for metastases which outnumber, by a ratio of 18 to1, primary liver malignancies [2]. This high incidence,coupled with the high sensitivity of FDG for livermetastases of up to 90% to 95%, has led to the increas-ingly widespread use of FDG PET for the detection ofliver metastases [3–5].

Use of FDG PET/CT in the evaluationof liver lesions

A case in point is colorectal cancer. Colorectal cancerforms the largest proportion of liver metastases, with upto 45% of colorectal cancer patients developing livermetastases, either at the time of diagnosis or in the yearsafter resection of the primary tumour [6]. Despite thepresence of metastases, a select group of these patientswith solitary or unilobar liver metastases are potentialcandidates for curative surgical resection, with signifi-cantly improved 5-year survival rates and long-termoutcome [7–9]. It is however important to ensure thatthese are truly isolated metastases prior to surgery. WhileCT is often the initial scan for the staging of colorectalmetastases, it may fail to detect small liver metastasesCorrespondence to: Gerald Jit Shen Tan; email: radio@geraldtan.com

ª Springer Science+Business Media New York 2013

Published online: 14 November 2013AbdominalImaging

Abdom Imaging (2014) 39:187–195

DOI: 10.1007/s00261-013-0043-3

and has difficulty in characterizing the smaller lesions,potentially understaging the extent of disease [10–12].PET/CT has been shown to be more sensitive thanconventional CT in the detection of both hepatic andextra hepatic metastases, resulting in a change in clinicalmanagement in 31.6% of colorectal cancer patients, mostcommonly by upstaging the disease and decreasing futilesurgery [13].

FDG PET/CT is also highly sensitive and accurate inthe detection of liver metastases from other primarycancers including non-small cell lung cancer, breastcancer and ovarian cancer. D’Souza et al. [14] found thatPET/CT had a sensitivity of 97% and specificity of 75%

for hepatic metastases, compared to 87.9% and 16.7%,respectively, for contrast-enhanced CT. Other studiesevaluating patients with presumed solitary and poten-tially resectable liver metastases found that FDG PET/CT resulted in a change in stage and therapy in 28% to30% of patients [15, 16]. These included positively iden-tifying indeterminate lesions as malignant in 52%, iden-tifying additional metastatic lesions or second primariesin 39% and correctly down staging the remaining 9%

[16]. The use of FDG PET/CT in the evaluation ofpotentially resectable liver metastases is now stronglyrecommended and looks to become routine clinicalpractice. [17, 18].

Primary liver tumours, on the other hand, are not aseasily detected with FDG PET/CT. FDG PET has asensitivity of only 55% to 64% for hepatocellular carci-noma (HCC), which is the commonest malignant pri-mary liver tumour worldwide, as compared to 90% forcontrast-enhanced CT [19, 20]. This lack of FDG avidityis particularly true of the well-differentiated subtypes[21]. FDG PET also fares poorly in the detection ofcholangiocarcinomas, particularly the infiltrative sub-types, where sensitivity is only 18% [22].

In contrast to the poor sensitivity of FDG PET/CTfor primary malignant liver tumours, FDG PET/CT hasgood specificity when characterizing FDG-avid liver le-sions as malignant. Wiering et al. [23] found in 2005 thatFDG PET had a sensitivity of 96.1% for hepatic diseasein patients with colorectal cancer. A meta-analysis of 39studies involving 3391 patients by Niekel et al. [24]concluded that FDG PET/CT had a specificity of 97.2%,compared to 94.9% for CT and 92.5% for MRI, althoughthese differences were not statistically significant.

Delbeke et al. [25] in a study of 120 hepatic lesions,found 88 lesions with increased FDG uptake, of whichall except one were malignant, with the exception being ayeast abscess. The remaining 22 benign lesions allshowed normal or decreased FDG uptake. These con-sisted of 2 other abscesses, as well as regenerative nod-ules, fibronodular hyperplasia, adenomas, cavernoushemangiomas, harmatomas, cysts and post-operativechanges [25]. Other studies involving 8 cases of fibron-odular hyperplasia and 6 cases of hemangiomas found

normal or even decreased accumulation of FDG in allcases [26, 27].

Given the high specificity and positive predictive va-lue of FDG PET/CT, with false positive results limited tothe occasional abscess, some authors propose that PETmay be useful to different malignant from benign lesionsin the liver, particularly in the setting of potentiallyresectable metastatic disease [13, 25]

We performed a review of FDG PET/CT studiesperformed at our institution over a 5-year period andidentified 1427 studies with metabolically active liver le-sions. Definitive histology was available in 214 of thesecases (217 lesions), with the vast majority of lesions (198,or 91%) shown to be malignant. Another 11 lesions(5.1%) remained indeterminate despite biopsy, but ofparticular interest were 8 lesions (3.7%) that were histo-logically proven to be benign. These 8 cases comprised 4abscesses, and 1 case each of cryptococcosis, hepaticadenoma, hemangioendothelioma and focal hepaticinflammation due to adjacent gallbladder empyema.Table 1 lists the possible causes of focal increased met-abolic activity within the liver.

Non-malignant causes of increasedmetabolic activity in the liver

Hepatic abscess

Abscesses are a known cause of increased FDG uptakein the liver [26, 27]. This is not unexpected given theincreased metabolic activity associated with inflamma-tory cells, such as macrophages and neutrophils. SUVvalues of 7.7 ± 2.2 have been reported for abscesses,which overlap those seen in malignant lesions [27].Clinical symptoms and laboratory results such as fever,raised white blood cell count or positive blood culturesmay help in the diagnosis. Imaging features such as asegmental distribution of the lesions and associatedbiliary dilatation (Fig. 1) should raise suspicion of aninfective aetiology. These findings can be seen on low-dose unenhanced CT, although intrahepatic biliarydilatation may be easier to appreciate with intravenouscontrast.

Table 1. Causes of focal increased metabolic activity in the liver

Malignant causesMetastatic diseaseHepatocellular carcinoma (poorly differentiated subtypes)Cholangiocarcinoma (peripheral nodular subtypes especially)

Benign causesAbscessesCryptococcosisHepatic adenomaHemangioendothelioma

Extrahepatic disease with direct extensionGallbladder carcinomaGallbladder empyema and inflammationAdrenal tumours

188 G. J. S. Tan et al.: FDG PET/CT in the liver

Cryptococcal infection

Cryptococcosis is an uncommon infection caused byencapsulated yeast and usually involves the central ner-vous system and lungs. Although the exact mechanism isunknown, cryptococcus is known to show FDG uptake,and can result in false positives for lung cancer on PET/CT

[28].Disseminated infection is rare andusually seen only inimmunocompromised individuals, with a single case re-port of FDG uptake in the skin, nodal and spleen [29].Liver involvement such as in the case shown in Figure 2has not been previously reported, but presumably wouldinvolve a similar pathophysiology as for the other organs.

Fig. 1. A 61-year-old female presented with painless jaun-dice and a pancreatic head mass. (A) FDG PET/CT showedincreased metabolic activity within the pancreatic head mass(arrow), which was causing biliary obstruction. (B) There wasalso increased FDG uptake in the right lobe of the liver, raisingthe possibility of liver metastases. (C) Contrast-enhanced CTdemonstrates multiple ill-defined lesions in segment 7 of theliver (black arrows), associated with dilated intrahepatic ducts

(white arrows). The possibility of a pancreatic tumour withliver metastases was raised, and biopsy of the liver lesion wasperformed to obtain tissue diagnosis. Histology demonstratedheavily inflamed portal tracts, accompanied by oedema,ductular reaction and acute pattern parenchymal cholestasis.There was no evidence of malignancy. The liver findings werethus attributed to ascending cholangitis with microabscessesrather than metastatic disease.

G. J. S. Tan et al.: FDG PET/CT in the liver 189

Hepatic adenoma

Hepatic adenoma is a rare benign primary liver tumourseen predominantly in young females. It consists of sheetsor cords of hepatocytes with the absence of portal andcentral veins, and bile ducts. Hepatic adenomas are tra-ditionally thought to be non FDG-avid, although therehas been a case report of a single hepatic adenomashowing increased metabolic activity [30]. On multiphasicCT, they are classically described as hypervascular(Fig. 3). This may be a useful distinguishing feature frommetastases which usually enhance in the portal venous

rather than the arterial phase. Some metastases (includingmelanoma) are however hypervascular, as are other pri-mary liver tumours such as focal nodular hyperplasia(FNH) or hepatocellular carcinoma [31]. In such situa-tions, MRI using hepatocyte-specific contrast agents, ortechnetium sulphur colloid or HIDA scans may be useful.

Hemangioendothelioma

Hemangioendotheliomas are rare vascular tumours thatare intermediate in histology and clinical behaviour

Fig. 2. A 39-year-old male with CD4 lymphopenia as well asa history of renal cell carcinoma treated with resection, pre-sented with cryptococcal meningitis and multiple subcentim-eter pulmonary, liver and splenic nodules. FDG PET coronal(A) and PET/CT fusion axial (B) images demonstrate intensemetabolic activity in both the liver and the spleen. The meanSUVbw was 4.7 in the liver and 5.2 in the spleen. Note that ifvisual assessment was performed using the liver alone as thereference, the study may be wrongly interpreted by inexperi-enced readers as showing only increased splenic activity,particularly on the axial images. Only after comparison withother organs and blood pool activity, as well as quantitative

SUV measurements, does it become apparent that the liveralso shows heterogeneously increased generalised FDG up-take. (C) Contrast-enhanced CT demonstrates numeroussmall hypodensities in the spleen, and to a lesser extent, theliver (arrows). These were not visible on the low-dose unen-hanced CT. A liver biopsy demonstrated non-caseating epi-thelioid cell granulomas with mild lymphocytic inflammation, aknown histological appearance of cryptococcal infection.Fungal cultures were negative although this may explained bythe fact that the patient had been on systemic anti-fungaltreatment by the time the biopsy was performed.

190 G. J. S. Tan et al.: FDG PET/CT in the liver

between hemangiomas and angiosarcomas [32]. Amajority of cases have an indolent course, although asmall proportion may undergo malignant degenerationinto angiosarcomas [33]. Hemangioendotheliomas tendto show heterogeneously increased signal on T2-weightedMRI, with some lesions having a target appearance dueto the presence of a central sclerotic zone and a periph-eral region of cellular proliferation [34, 35]. This mayexplain our findings of peripheral increased FDG uptakeand central photopenia on PET (Fig. 4). Other possibleexplanations for the central areas of reduced metabolicactivity include hemorrhage, coagulation necrosis andcalcification. [36]. Distinguishing between these causeson PET/CT alone may not be possible, although MRI(particularly heme-sensitive sequences) may be useful inthis regard.

Extrahepatic lesions

Increased hepatic FDG uptake due to extra-hepatic le-sions may occur in two ways. The first is due to super-imposition of extra-hepatic FDG activity on the PETscan onto inappropriate anatomic structures on the CT,termed misregistration. This may occur if the patientmoves in between or during scans, or more subtly as aresult of scan technique, for example, if the CT is acquiredin maximal inspiration, while the PET is obtained duringnormal ‘‘free’’ breathing [21]. This can result in FDGuptake in the lung base being wrongly projected into theliver, and vice versa [37]. With the faster PET acquisitiontimes and respiratory gating techniques available in thenewer generation of PET/CT scanners, the incidence ofsuch artefacts may potentially be reduced.

Fig. 3. A 51-year-old female with a history of melanoma ofthe left ankle treated with surgical resection 15 years ago,presented with regional dermal lymphatic metastases. FDGPET (A) and PET/CT fusion (B) images demonstrate multipleareas of increased metabolic activity in the liver (arrows).Contrast-enhanced CT in the arterial (C) and portal venous(D) phases demonstrated corresponding anatomic lesions,some showing early enhancement with subsequent washout

(arrows). Given the findings of multiple hypervascular, hyper-metabolic liver lesions in the setting of malignant melanoma,these were felt to represent metastases. However histology ofthe largest lesion demonstrated a hepatic adenoma with noevidence of metastatic melanoma. Immunohistochemistry formelanocytic antigens was also negative. These lesions re-mained stable in size on follow up CT 1 year later, and a finaldiagnosis of hepatic adenomatosis was made.

G. J. S. Tan et al.: FDG PET/CT in the liver 191

The second process by which increased intrahepaticuptake from extrahepatic sources may occur is as a resultof direct invasion or secondary inflammation of the liverfrom pathology in adjacent organs.

These organs would include the gallbladder, kidneyor adrenals on the right, and the stomach on the left,

although other sources such as the hepatic flexure of thecolon, portal hilar structures and even peritoneal nodulesmay also present with similar findings. Figure 5 shows acase of an empyema of the gallbladder resulting in sec-ondary inflammation of the adjacent liver, due to in-creased FDG uptake by activated macrophages [38].

Fig. 4. A 43-year-old man with multiple liver hemangioen-dotheliomata, presented with increase in size of one of the le-sions on serial CT, raising possibility of malignant degenerationinto an angiosarcoma. (A) FDG PET coronal image demon-strates subcapsular areas of mild to moderately increasedmetabolic activity most prominent in the right lobe of the liver(arrows). (B) PET/CT fusion axial image through the enlarginglesion in segment 7, demonstrates mild increase in FDG uptakeperipherally (black arrows), with a central photopenic area

(white arrow). Considerations at this point included peripheralmalignant change and central hemorrhage. (C) The non-con-trast CT component of the PET/CT showed a uniformly hypo-dense lesion, and could not explain the differential FDG uptake.(D) T2-weighted MRI showed an area of central cystic change(arrow) corresponding to the area of photopenia. The finaldiagnosis was that of a peripheral rim of benign hemangioen-dothelioma, with an area of central haemorrhage.

192 G. J. S. Tan et al.: FDG PET/CT in the liver

Approach to an FDG-avid liver lesion

It is therefore insufficient to rely merely on the fusedPET/CT axial images when evaluating FDG-avid lesionsin the liver. The reconstructed 3D volume dataset forboth the PET and CT scans should be visualized intransaxial, coronal and sagittal planes, and the maximumintensity projections should also be reviewed [39].

This is important for the detection of misregistra-tion artefacts, and to differentiate extra-hepatic sourcesof activity from true intra-hepatic lesions. Featuressuch as preservation of a fat plane between the massand the liver, formation of an acute angle with theliver or the absence of a ‘‘claw sign’’ (a rim of livertissue extending partially around the mass), shouldraise the possibility of an extra-hepatic mass as theprimary lesion [40].

Additional imaging modalities, such as contrast-en-hanced CT, multiphasic CT or MRI, are particularlyuseful in the liver, given the poor soft tissue contrast oflow dose unenhanced CT. These do not have to be ac-quired simultaneously with the PET study, and can bereviewed either using visual fusion (traditional side-by-side review) or software image fusion [41]. Such a reviewshould be standard practice and may explain focal dif-ferences in FDG uptake within a lesion, thus helping tobetter characterise lesions.

The distribution of FDG uptake may also be useful incharacterising the nature of disease. If the increasedmetabolic activity is generalised or segmental in nature, adiffuse process such as inflammation or infection shouldbe considered, particularly if there are supporting clinicalsigns or biochemical markers. There have however been

Fig. 5. A 45-year-old man presented with a gallbladdermass on ultrasound. He had no signs of infection and thepossibility of a gallbladder carcinoma was considered. (A)Coronal PET demonstrates a bilobed area of increased met-abolic activity, with the inferior component centred on thegallbladder (black arrow), and the superior aspect extending

into the liver (white arrowhead). FDG PET/CT (B) and CT (C)images show that axial images alone are insufficient to dis-tinguish this from an intrahepatic lesion. Explorative surgerydemonstrated a gallbladder empyema with only inflammatorychange in the adjacent liver.

G. J. S. Tan et al.: FDG PET/CT in the liver 193

reports of primary liver malignancies such as intrahe-patic cholangiocarcinoma or lymphoma demonstratingdiffuse FDG uptake [42, 43].

Finally, it is important to avoid relying on visualassessment of lesions using only the liver as a reference.Diffusely increased liver uptake may be overlooked insuch a situation. Qualitative visual assessment can still beperformed, but using other organs or blood pool activityas a reference. Alternatively, a SUVbw cut-off of greaterthan 3.5, as suggested by Delbeke et al. [25] may providea more objective method of evaluating the liver. This isparticularly important in cases of infection and inflam-mation, which can result in generalised liver disease asopposed to the more heterogeneous or focal pattern ofinvolvement by neoplasms.

Conclusion

FDG PET/CT is highly sensitive for the detection ofmalignant liver lesions, particularly metastases. Howeverit is important to recognise that lesions other thanmalignancy can result in increased FDG uptake in theliver. Careful evaluation of the pattern of distribution,together with multiplanar reconstruction and correlationwith other imaging modalities, may help to arrive at thecorrect diagnosis.

References

1. Khandani AH, Wahl RL (2005) Applications of PET in liverimaging. Radiol Clin N Am 43:849–860

2. Brant WE (1998) Liver. In: Webb WR, Brant WE, Helms CA (eds)Fundamentals of Body CT, 2nd edn. Philadelphia: WB Saunders, pp195–212

3. Lai DT, Fulham M, Stephen MS, et al. (1996) The role of whole-body positron emission tomography with [18F] fluorodeoxyglucosein identifying operable colorectal cancer metastases to the liver.Arch Surg 131:703–707

4. Kinkel K, Lu Y, Both M, Warren RS, Thoeni RF (2002) Detectionof hepatic metastases from cancers of the gastrointestinal tract byusing noninvasive imaging methods (US, CT, MR imaging, PET): ameta-analysis. Radiology 224:748–756

5. Bipat S, van Leeuwen MS, Comans EF, et al. (2005) Colorectalliver metastases: CT, MR imaging, and PET for diagnosis-meta-analysis. Radiology 237:123–131

6. Wiering B, Vogel WV, Ruers TJ, Oyen WJ (2008) Controversies inthe management of colorectal liver metastases: role of PET andPET/CT. Dig Surg 25:413–420

7. Yamaguchi A, Kurosaka Y, Kanno M, et al. (1993) Analysis ofhepatic recurrence of colorectal cancer after resection of hepaticmetastases. Int Surg 78:16–19

8. Nordlinger B, Guiguet M, Vaillant JC, et al. (1996) Surgicalresection of colorectal carcinoma metastases to the liver: a prog-nostic scoring system to improve case selection, based on 1568patients. Association Francaise de Chirurgie. Cancer 77:1254–1262

9. Gayowski TJ, Iwatsuki S, Madariaga JR, et al. (1994) Experiencein hepatic resection for metastatic colorectal cancer: analysis ofclinical and pathologic risk factors. Surgery 116:703–710

10. Delbeke D, Vitola JV, Sandler MP, et al. (1997) Staging recurrentmetastatic colorectal carcinoma with PET. J Nucl Med 38:1196–1201

11. Abdel-Nabi H, Doerr RJ, Lamonica DM, et al. (1998) Staging ofprimary colorectal carcinomas with fluorine-18 fluorodeoxyglucosewhole-body PET: correlation with histopathologic and CT find-ings. Radiology 206:755–760

12. Topal B, Flamen P, Aerts R, et al. (2001) Clinical value of whole-body emission tomography in potentially curable colorectal livermetastases. Eur J Surg Oncol 27:175–179

13. Sacks A, Peller PJ, Surasi DS, et al. (2011) Value of PET/CT in themanagement of liver metastases, part 1. AJR Am J Roentgenol.197:W256–W259

14. D’Souza MM, Sharma R, Mondal A, et al. (2009) Prospectiveevaluation of CECT and 18F-FDG-PET/CT in detection of hepaticmetastases. Nucl Med Commun 30:117–125

15. Grassetto G, Fornasiero A, Bonciarelli G, et al. (2010) Additionalvalue of FDG-PET/CT in management of ‘‘solitary’’ liver metas-tases: preliminary results of a prospective multicenter study. MolImaging Biol 12:139–144

16. Briggs RH, Chowdhury FU, Lodge JP, Scarsbrook AF (2011)Clinical impact of FDG PET-CT in patients with potentiallyoperable metastatic colorectal cancer. Clin Radiol 66:1167–1174

17. Engledow AH, Skipworth JR, Pakzad F, et al. (2012) The role of18FDG PET/CT in the management of colorectal liver metastases.HPB (Oxford) 14:20–25

18. O’Connor OJ, McDermott S, Slattery J, Sahani D, Blake MA(2011) The use of PET-CT in the assessment of patients withcolorectal carcinoma. Int J Surg Oncol 2011:846512.

19. Khan MA, Combs CS, Brunt EM, et al. (2000) Positron emissiontomography scanning in the evaluation of hepatocellular carci-noma. J Hepatol 32:792–797

20. Wudel LJ Jr, Delbeke D, Morris D, et al. (2003) The role of [18F]fluorodeoxyglucose positron emission tomography imaging in theevaluation of hepatocellular carcinoma. Am Surg 69:117–124

21. Blake MA, Singh A, Setty BN, et al. (2006) Pearls and pitfalls ininterpretation of abdominal and pelvic PET-CT. Radiographics26:1335–1353

22. Anderson CD, Rice MH, Pinson CW, et al. (2004) Fluorodeoxy-glucose PET imaging in the evaluation of gallbladder carcinomaand cholangiocarcinoma. J Gastrointest Surg 8:90–97

23. Wiering B, Krabbe PF, Jager GJ, Oyen WJ, Ruers TJ (2005) Theimpact of fluor-18-deoxyglucose-positron emission tomography in themanagement of colorectal liver metastases. Cancer 104:2658–2670

24. Niekel MC, Bipat S, Stoker J (2010) Diagnostic imaging of colo-rectal liver metastases with CT, MR imaging, FDG PET, and/orFDG PET/CT: a meta-analysis of prospective studies includingpatients who have not previously undergone treatment. Radiology257:674–684

25. Delbeke D, Martin WH, Sandler MP, et al. (1998) Evaluation ofbenign vs malignant hepatic lesions with positron emissiontomography. Arch Surg 133:510–515

26. Kurtaran A, Becherer A, Pfeffel F, et al. (2000) 18F-fluorodeoxy-glucose (FDG)-PET features of focal nodular hyperplasia (FNH)of the liver. Liver 20:487–490

27. Son HB, Han CJ, Kim BI, et al. (2002) Evaluation of various he-patic lesions with positron emission tomography (in Korean).Taehan Kan Hakhoe Chi 8:472–480

28. Igai H, Gotoh M, Yokomise H (2006) Computed tomography (CT)and positron emission tomography with [18F] fluoro-2-deoxy-D-glucose (FDG-PET) images of pulmonary cryptococcosis mimick-ing lung cancer. Eur J Cardiothorac Surg 30:837–839

29. Seshadri N, Kaur B, Balan K (2007) Disseminated cryptococcosis:detection by F-18 FDG PET. Clin Nucl Med 32:476–478

30. Patel PM, Alibazoglu H, Ali A, Fordham E, LaMonica G (1997)‘False-positive’ uptake of FDG in a hepatic adenoma. Clin NuclMed 22:490–491

31. Grazioli L, Federle MP, Brancatelli G, et al. (2001) Hepatic ade-nomas: imaging and pathologic findings. Radiographics 21:877–892

32. Weiss SW, Enzinger FM (1982) Epithelioid hemangioendotheli-oma: a vascular tumor often mistaken for a carcinoma. Cancer50:970–981

33. Woo RK, Albanese CT (2008) Pediatric Surgery. In: Norton JA,Barie PS, Bollinger RR, et al. (eds) Surgery: Basic Science andClinical Evidence, 2nd edn. New York: Springer, p 684

34. Lyburn ID, Torreggiani WC, Harris AC, et al. (2003) Hepaticepithelioid hemangioendothelioma: sonographic, CT, and MRimaging appearances. AJR Am J Roentgenol 180:1359–1364

35. Ros LH, Fernandez L, Villacampa VM, Ros PR (1999) Epithelioidhemangioendothelioma of the liver: characteristics on magneticresonance imaging: case report. Can Assoc Radiol J 50:387–389

194 G. J. S. Tan et al.: FDG PET/CT in the liver

36. Kehagias DT, Moulopoulos LA, Antoniou A, et al. (2000) Hepaticepithelioid hemangioendothelioma: MR imaging findings. Hepa-togastroenterology 47:1711–1713

37. Osman MM, Cohade C, Nakamoto Y, et al. (2003) Clinically sig-nificant inaccurate localization of lesions with PET-CT: frequencyin 300 patients. J Nucl Med 44:240–243

38. Bohm B, Voth M, Geoghegan J, et al. (2004) Impact of positronemission tomography on strategy in liver resection for primary andsecondary liver tumors. J Cancer Res Clin Oncol 130:266–272

39. Boellaard R, O’Doherty MJ, Weber WA, et al. (2010) FDG PETand PET/CT: EANM procedure guidelines for tumour PETimaging: version 1.0. Eur J Nucl Med Mol Imaging 37:181–200

40. Routhier JR, Woodfield CA, Mayo-Smith WW (2009) AJRteaching file: fat-containing retroperitoneal mass presenting withacute flank pain. AJR Am J Roentgenol 192:S122–S124

41. Vogel WV, Oyen WJ, Barentsz JO, Kaanders JH, Corstens FH(2004) PET/CT: panacea, redundancy, or something in between? JNucl Med 45(Suppl 1):15S–24S

42. Kim YJ, Yun M, Lee WJ, Kim KS, Lee JD (2003) Usefulness of18F-FDG PET in intrahepatic cholangiocarcinoma. Eur J NuclMed Mol Imaging 30:1467–1472

43. Lam WW, Osmany S (2009) Biliary non-Hodgkin lymphoma de-tected by F-18 FDG PET/CT. Clin Nucl Med 34:791–792

G. J. S. Tan et al.: FDG PET/CT in the liver 195