Fr

GMa

b

a

ARRA

KWMPtRL

1

aala

hoct

0d

European Journal of Radiology 81 (2012) e269– e276

Contents lists available at ScienceDirect

European Journal of Radiology

jo ur n al hom epage: www.elsev ier .com/ locate /e j rad

DG–PET–CT and whole-body MRI for triage in patients planned foradioembolisation therapy

.P. Schmidta,∗, P. Paprottkaa, T.F. Jakobsa, R.T. Hoffmanna, A. Baur-Melnyka, A. Haugb,. Notohamiprodjoa,b, A. Baur-Melnyka, K. Nikolaoua, M.F. Reisera, C. Rista

Department of Clinical Radiology, University Hospitals Munich – Grosshadern, LMU, Marchioninistr. 15, 81377 München, GermanyDepartment of Nuclear Medicine, University Hospitals Munich – Grosshadern, LMU, Germany

r t i c l e i n f o

rticle history:eceived 7 December 2010eceived in revised form 1 February 2011ccepted 4 February 2011

eywords:hole bodyagnetic resonance imaging

ositron emission tomography computedomographyadioembolisationiver metastases

a b s t r a c t

Purpose: The purpose was to evaluate the potential of FDG–PET–CT and whole-body MRI (WB-MRI) asdiagnostic triage methods for patients planned for radioembolisation of metastatic liver disease.Materials and methods: 135 patients with multifocal liver metastases were evaluated for potential pal-liative therapy with radioembolisation using 90-Yttrium microspheres. All patients were examinedconsecutively with FDG–PET–CT and WB-MRI for exclusion of relevant extra-hepatic tumor mani-festations. All patients underwent 99mTc-albumine angiography followed by scintigraphy to excludesignificant hepato-pulmonary shunting.Results: Out of the 135 patients included into the pre-therapeutic diagnostic algorithm, 56% were eligibleand received radioembolisation, while 44% could not be treated. In 91% the exclusion criteria was diag-nosis of significant extra-hepatic metastatic disease. In 85% exclusion diagnosis was made concordantlyby both FDG–PET–CT and WB-MRI, in 9% diagnosis was provided by PET–CT, in 6% by WB-MRI alone.Patient-based sensitivity for detection of extra-hepatic disease was 94% for PET–CT and 91% for WB-MRI.False-positive diagnosis of extrahepatic disease leading to exclusion for radioembolisation therapy wasmade in 2% of patients, in one patient by PET–CT and in one patient by WB-MRI alone. Overall, speci-ficity for inclusion of radioembolisation therapy by combining both modalities was 99%. In 9% of patientsangiographic diagnosis made radioembolisation impossible, in 7% solely the angiographic findings were

decisive.Conclusion: Both FDG–PET–CT and WB-MRI are efficient diagnostic triage methods for patients plannedfor radioembolisation of liver metastases. Overall, FDG–PET–CT shows a trend to higher diagnostic accu-racy compared to WB-MRI and may be used as imaging method of choice as a standalone examination.In combination, both modalities exhibited high sensitivity for the diagnosis of extra-hepatic tumor man-ifestations and result in high specificity.

. Introduction:

Primary and secondary liver tumors are common malignanciesnd are being treated more aggressively nowadays than decadesgo. Although resection of solitary liver metastases can result inong-term survival, only 10–20% of patients with liver metastasesre amenable surgical candidates [1].

The intrahepatic application of radioactive microspheres via theepatic artery represents a new generation of locoregional therapy

f diffuse or multifocal liver tumors, such as hepatocellular car-inoma or secondary liver metastases, for which to date systemicherapy was the only remaining option [2–4]. The current stan-

∗ Corresponding author. Tel.: +49 89 7095 0; fax: +49 89 7095 8832.E-mail address: gerwin.schmidt@med.uni-muenchen.de (G.P. Schmidt).

720-048X/$ – see front matter © 2011 Elsevier Ireland Ltd. All rights reserved.oi:10.1016/j.ejrad.2011.02.018

© 2011 Elsevier Ireland Ltd. All rights reserved.

dard for radioembolisation (RE) is 90-yttrium glass microspheres.Although it is not considered as a cure, it has been shown to improvequality of life and prolong survival [5,6]. For the patient inclu-sion selection process, an important question to address is theabsence of significant extrahepatic metastatic manifestations. Fur-thermore, a conventional angiography of the hepatic artery withthe use of 99m-macroaggregated albumine (99mTc-MAA) is per-formed for the exclusion of extra-hepatic deposition of particlesand quantitative assessment of hepato-pulmonary shunts, whichcould lead to radiation-induced pneumonitis. Within this setting,also portal venous thrombosis and aberrant visceral vessels shouldbe excluded.

For the staging and restaging of tumor patients, whole-bodyimaging strategies are increasingly preferred over multi-modalityalgorithms for a precise and rapid assessment of total tumor bur-den, especially in a clinical setting with direct therapeutic impact.

e rnal of

TtalthltdlcfTmafs

Mcma

2

akwbprhc(twtfmawbPaopcwsa

2

siPtwi2eCc

270 G.P. Schmidt et al. / European Jou

he advent of combined positron emission tomography–computedomography (PET–CT) with the use of fluorodesoxyglucose (FDG)s an “allround” tracer has expanded diagnostic options in onco-ogic imaging by adding functional information of a PET exam tohe detailed anatomical data of CT within a single examination. Itas been reported that PET–CT has markedly increased malignant

esion localization and diagnostic sensitivity for various tumor enti-ies [7,8]. Alternatively, advances in scanner technology, sequenceesign and image acquisition algorithms, have made high reso-

ution magnetic resonance imaging over the total body anatomylinically feasible. Whole-body MRI (WB-MRI) has been introducedor staging as well as for surveillance of various neoplastic diseases.he excellent soft tissue contrast in bone marrow or parenchy-al organs and its high spatial resolution make WB-MRI a useful

pplication, especially for imaging of metastases from tumors thatrequently metastasize to the bone, brain and abdominal organs,uch as breast cancer or colorectal cancer [9–11].

The purpose of this study was to compare FDG–PET–CT and WB-RI as applications for total body imaging within pre-therapeutic

linical focus and investigate their diagnostic potential as triageethods for patients planned for selective internal radiation ther-

py of liver metastases.

. Materials and methods

Between March 2004 and December 2008 135 patients (meange 61 years, range 38–78 years, 74 female/61 male) with anown history of therapy-refractory multi-focal liver metastasesere evaluated at our hospital for potential palliative radioem-

olisation therapy with the use of 90-yttrium microspheres. Therimary tumors of our cohort were: colon- and breast (each 30%),ectal- (10%), neuroendocrine-tumors (8%), cholangiocellular- (8%),epatocellular- (2%), ovarian-cancer (2%), melanoma and sar-oman (each 2%). 2% were metastases of unknown primaryCUP) and 3% other tumors. Within the established clinical pre-herapeutic algorithm all patients were examined consecutivelyith a dual-modality PET–CT-scanner (Gemini, Philips Medical Sys-

ems, Cleveland, OH) as well as whole-body-MRI (WB-MRI) at 1.5 Tor exclusion of relevant extrahepatic tumor manifestations. The

ean time period between both exams was 3 ± 6 days. WB-MRInd PET–CT were performed in agreement with the patients asell as the clinicians in charge. Approval of the institutional review

oard and written patient consent were obtained beforehand. BothET–CT and WB-MRI were well tolerated by all patients. Addition-lly, a conventional angiography of the hepatic artery with the usef 99m-Technetium macroaggregated albumine (99mTc-MAA) waserformed to exclude a dystopic, extra-hepatic deposition of parti-les and for a quantitative assessment of hepato-pulmonary shuntsith the use of thoraco-abdominal scintigraphy. Furthermore, pos-

ible portal venous thrombosis and aberrant visceral vessels weressessed and excluded, respectively.

.1. FDG–PET–CT imaging

Examinations were performed on a two-detector row PET–CT-canner (Gemini, Philips Medical Systems, Cleveland, OH) afternjection of an average of 294 MBq [18F]-fluoro-2-deoxy-D-glucose.atients were asked to fast for at least 6 h prior to the examina-ion to assure blood glucose levels below 150 mg/dL. Buscopanas applied intravenously to avoid a first-pass uptake of FDG

nto the smooth muscle of the gastrointestinal tract. Additionally,

0 mg of Furosemide was administered in order to increase renalxcretion of the tracer. 1 h after FDG-administration a low-doseT scan in shallow breathing was performed for PET-attenuationorrection covering neck, thorax, abdomen and pelvis (40 mA s,

Radiology 81 (2012) e269– e276

120 kV, collimation 2 mm × 5 mm, pitch 1.5). Then the emissionscan followed using a 3D-RAMLA (row action maximum likeli-hood) algorithm for reconstruction (10 bed positions, FOV 10 cm,144 × 144 matrix). Finally, a diagnostic contrast-enhanced CT wasconducted (120 mA s, 160 kV, collimation 2 × 5 mm, pitch 1) withapplication of 120 ml of i.v.-contrast agent (Ultravist® 300, BayerSchering Pharma AG, Berlin, Germany) in the venous phase (80 sdelay). The PET- and diagnostic CT-data were fused with the useof special software (Syntegra®, Philips Medical Systems, Cleveland,OH, USA). For the whole examination an average dose of ionizingradiation of 25 mSv has been calculated [12]. Examination time forPET–CT was 103 min (60 min patient preparation, total acquisitiontime 43 min).

2.2. Whole-body MRI

WB-MRI was performed on a 1.5 T (40 mT/min, max. slewrate 20 T/m/s) whole-body scanner (Magnetom Avanto, SiemensMedical Solutions, Erlangen, Germany) using integrated parallelacquisition techniques (iPAT) [13]. The scanner allows the connec-tion of up to 76 receiver coil elements from multiple phased-arraysurface coils, covering the patient from head to toe, with simul-taneous signal reception from 32 independent receiver channels.Using automated table motion, parallel imaging in 3 spatial direc-tions at a total scan range of 205 cm can be performed withoutpatient repositioning. The patients were examined from head tothe proximal calves with STIR-sequences on four body levels incoronal orientation: head/neck, pelvis, thighs/calves (TR 5620/TE92), as well as the thorax/abdomen (TR 3380/TE 101) in breath-hold technique with prospective 2D-navigator correction (PACE,prospective acquisition correction). Subsequently, the body wasscanned with coronal T1-weighted TSE-imaging (TR 79/TE 12;thorax/abdomen TR 400/TE 8.2, breath-hold technique). The lungwas examined in axial orientation with STIR- (TR 3800/TE 100)and HASTE-sequences (TR 1100/TE 27), followed by a respiratory-triggered T2w-fat saturated-TSE scan of the liver (TR 2010/TE 101)and coronal HASTE of the abdomen (TR 900,TE 73). Then, sagittalimaging of the upper and lower spine with T1-weighted-TSE- (TR849/TE11) and STIR-sequences (TR 5700/TE 59) was performed.

After application of gadolinium-DTPA (Magnevist®, BayerSchering Pharma AG, Berlin, Germany, 3 ml/s; 0.2 mmol/kg, 20 mlsaline flushing) a dynamic (arterial-, portal venous-, late venousphase) axial 3D-VIBE (volume interpolated breath hold exam) liverscan was accomplished (TR 4.38/TE 1.61) including a late venousscan of the lung level (TR 9.1/TE 4.76). Then axial T1w- and T2w-TSE imaging (TR 635/TE 17 and TR 1420/TE 109) of the brain wasperformed. Finally, a fat saturated T1w-GRE-sequence of the wholeabdomen in axial orientation (TR 179/TE 3.33) was performed.Table 1 gives an overview of the applied scan protocol.

For coronal whole-body imaging a PAT-factor of 3 was used. APAT factor of 2 was used for axial imaging of the brain, lung, andabdomen, as well as sagittal imaging of the spine. Examination timefor WB-MRI was 55–60 min (acquisition time 51:24 min).

2.3. Data analysis

With the use of our departmental digital imaging and reportingarchives all imaging findings and examination results were retro-spectively reviewed by one radiologist with 7 years of experienceand one radiologist/nuclear physician with 8 years of experiencein oncologic imaging. The presence, location and number of sus-pected extrahepatic malignant lesions were assessed and recorded

by organ regions: lymph node, lung, bone, peritoneum and other.Brain metastases were described as additional findings in WB-MRIonly and not included in the comparative analysis of diagnosticaccuracy; established region-specific size criteria were applied to

G.P. Schmidt et al. / European Journal of Radiology 81 (2012) e269– e276 e271

Table 1Whole body-MRI protocol with imaging parameters on a 32-receiver channel whole-body scanner at 1.5 T.

Sequence Image plane Matrix/Resolution (mm3) Slice (mm) Acq. time (min)

STIR-WB Coronal 384/1.8 × 1.3 × 5.0 5.0 9:43HASTE-abdo Coronal 384/1.4 × 1.3 × 5.0 5.0 0:38HASTE-lung Axial 320/1.3 × 1.2 × 6.0 6.0 0:44STIR-lung Axial 320/1.8 × 1.2 × 6.0 6.0 1:01T2wfsTSE-liver (free breathing) Axial 320/1.6 × 1.2 × 5.0 5.0 3:41T1wTSE-WB Coronal 384/1.7 × 1.3 × 5.0 5.0 10:30T1wTSE-spine Sagittal 384/1.0 × 1.0 × 3.0 3.0 7:46STIR-spine Sagittal 384/1.0 × 1.0 × 3.0 3.0 7:22Dyn. VIBE liver Axial 384/1.9 × 1.5 × 3.0 3.0 2:20T1wfs GRE pelvis Axial 320/1.5 × 1.2 × 6.0 6.0 1:17

0/0.7 × 0.7 × 5.0 5.0 3:112/0.5 × 0.5 × 5.0 5.0 3:11

51:24

dacHaooobamcisleo

FFotcnoin9fMaa

3

te(thWcppdc

ep

Table 2Overview on exclusion criteria revealed by whole-body imaging and conventionalangiography and patient eligibilty. Patient-based sensitivity for the detection ofsubstantial extra-hepatic disease and specificity for inclusion for radioembolisationtherapy are italicised.

PET–CT WB-MRI

Patients n = 135Eligible for SIRT 56% (76/135)Untreated 44% (59/135)

Exclusion criteriaRelevant extrahepatic metastases 91% (54/59)

Sensitivity 100% (54/54)94% (51/54) 91% (49/54)

Specificity 99% (75/76)97% (74/76) 88% (67/76)

Angiographic diagnosis 9% (5/59)

whole-body imaging.Based on anatomic distribution PET–CT detected malignant

lesions in 56 organ regions compared to 52 regions with WB-MRI(Table 3). Local recurrent tumor was found by PET–CT in n = 2 and

Table 3Extrahepatic malignant tumor manifestations detected with FDG–PET–CT and WB-MRI by anatomic distribution.

PET–CT WB-MRI

Extrahepatic malignant manifestations n = 56 n = 52Local recurrent tumor n = 2 n = 1Lymph nodes n = 21 n = 19Organ metastases n = 33 n = 32Bone n = 16 n = 16

T1wTSE-brain Axial 32T2wTSE-brain Axial 51Total

etermine tumor involvement when assessing lymph nodes, e.g.ccording to the adapted AJCC (American Joint Committee on Can-er) classification for thoracic lymph nodes by Glazer et al. [14,15].ard criteria for malignancy in both modalities were signs ofggressive expansion of a lesion, such as ill-defined borders, erosionr infiltration of neighboring anatomical structures, hemorrhager signs of necrosis. In MRI, malignancy was determined basedn established sequence-specific signal changes, e.g. a hypointenseone marrow signal in T1-weighted imaging in combination with

hyperintense signal in STIR in case of bone lesions [16]. Further-ore, typical findings of abnormal static (CT) or dynamic (MRI)

ontrast uptake characteristics were employed [17]. In addition,n PET–CT a focally increased glucose uptake using a “maximumtandardized uptake value” was used to further classify suspiciousesions using a SUVmax as a cutoff value [18]. According to Beggst al. a SUVmax < 2.5 helped as indicator to rule out malignant causef a suspicious lesion.

Findings classified as suspicious for malignancy in bothDG–PET–CT and WB-MRI were classified as most likely malignant.urthermore, a progressive change in size and number of a lesionr an increase of pathological tracer uptake in non-responderso applied therapy (alternatively a decrease in responders), wereonsidered as hard criteria for malignancy on follow-up exami-ations. Findings classified as suspicious for malignancy in onlyne of both examinations were classified as a questionable find-ng. All discordant findings were cross-checked by radiological oruclear medicine follow-up studies for at least 3 months (average

months, 3–40 months) as a standard of reference. In detail, theollowing imaging studies were performed: PET–CT (n = 64), WB-

RI (n = 34), CT (n = 7) and dedicated MRI (n = 2). Finally, resultsnd findings from pre-therapeutic 99mTc-MAA-angiography andbdomino-thoracic scintigraphy were reviewed.

. Results

From the 135 patients who were included into the pre-herapeutic diagnostic algorithm, 56% (76/135) of patients wereligible and received radioembolisation therapy, while 44%59/135) patients could not be treated. In 91% (54/59) of patientshe exclusion criteria for radioembolisation was substantial extra-epatic metastatic disease diagnosed by FDG–PET–CT and/orB-MRI. In 85% of patients (46/54) exclusion diagnosis was made

oncordantly by both modalities, in 9% (5/54) diagnosis was solelyrovided by the PET–CT exam, in 6% (3/54) by WB-MRI alone. Theatient-based sensitivity for detection of substantial extra-hepaticisease was 94% for PET–CT (51/54) and 91% for WB-MRI (49/54),

ombined patient-based sensitivity was 100% (54/54).

False-positive diagnosis of extrahepatic disease leading toxclusion for radioembolisation therapy was made in 2% (2/135) ofatients, in one patient by PET–CT and in one patient by WB-MRI. In

Significant gastric collaterals n = 3Hepatic artery occlusion n = 1Portal vein thrombosis n = 1

one case multifocal lymph node metastases were falsely diagnosedby PET–CT, in other case multifocal bone metastases were falselydiagnosed in WB-MRI.

False-positive diagnosis not leading to radioembolisation wasmade by PET–CT in 3% (2/76) and by WB-MRI in 12% (9/76) of cases,leading to a specificity of 97% (74/76) and 88% (67/76), respectively.In these patients suspected lymph node- or bone oligometastasiswas diagnosed, yet positive decision for therapy was made due tothe limited, unifocal extent of potential metastatic disease. Fur-thermore suspicion had only been raised in either of one of bothmodalities and not confirmed by both imaging procedures. No signsof suspicious extra-hepatic tumor manifestations were found onfollow-up in these patients. Overall, specificity for inclusion forradioembolisation therapy by combining both modalities was 99%(75/76). Table 2 gives an overview on the diagnostic results of

Lung n = 10 n = 8Peritoneum n = 4 n = 3Other n = 3 n = 3Brain n = 0 n = 2

e272 G.P. Schmidt et al. / European Journal of Radiology 81 (2012) e269– e276

Fig. 1. 46-Year old patient post resection of a NET of the rectum. (A) Sagittal PET–CT reconstruction shows a focally increased tracer uptake at the level of the rectosigmoidanastomosis. (B–D) WB-MRI shows a semicircular perirectal structure with focal contrast-uptake (arrow) which may be due to scarring and local tumor recurrency, respec-t he peP ce. (Ei

bntmvnmpditbC8

rgpTssncso

4

f

ively. CT depicts a solid lesion with a nodular componant in the anterior aspect of tET–CT image exhibits a markedly increased FDG-uptake indicating tumor recurrenliac fossa with enlarged size and increased tracer uptake.

y WB-MRI in n = 1 regions, extensive lymph node metastases in = 21 and n = 19 regions (Figs. 1 and 3). Extrahepatic organ metas-ases were found in n = 33 and n = 32 regions, respectively. Both

odalities performed equally in detecting bone metastases (n = 16s n = 16, Fig. 3), PET–CT found more lung metastases (n = 10 vs

= 8, Fig. 4) and peritoneal lesions (n = 5 vs n = 3, Fig. 3). Furtheretastases were found equally in n = 3 other sites, namely the

ericardium, spleen and adrenals (Fig. 4). WB-MRI and PET–CTiagnosed significant infiltration of hepatic tumor manifestations

nto the abdominal wall and renal fascia in one patient and washerefore excluded from radioembolisation (Fig. 2). Additionalrain metastases were found in 2 patients by WB-MRI only (Fig. 4).oncordance of both methods for regional malignant findings was5% (50/59) (Figs. 1–4).

In 9% (5/59) of patients angiographic diagnosis finally madeadioembolisation impossible and in 7% (4/59) solely the angio-raphic findings were decisive. In one patient WB-MRI indicatedortal vein tumor infiltration, which was excluded in angiography.able 3 gives an overview of the angiographic findings. The averagehunt volume in eligible patients quantified by thoraco-abdominalcintigraphy was 6.3 ± 3.0%. In 9% (7/76) of treated patients a sig-ificant hepato-pulmonary shunt of 10–15% was diagnosed, as aonsequence leading to a 20% dose reduction of applied Yttrium-pheres. In none of the patients a severe hepato-pulmonary shuntf >15% was detected.

. Discussion

Selective internal radiotherapy is a new therapeutic approachor locoregional therapy of diffuse or multifocal secondary liver

rirectal fat tissue and posterior invasion of the mesorectal fascia (arrow). The fused–G) Both MRI and PET–CT additionally reveal a lymph node metastasis in the right

metastases. Recent publications indicate its effectiveness by reduc-ing tumor size, improving the quality of life of patients andprolonging overall survival [19,20]. However, the demands of agentadministration and clinical follow-up require accurate treatmentplanning and patient selection algorithms to assure a cost-efficientand sensible therapy result. For pre-therapeutic eligibility examsand potential therapy, patients are usually admitted for short-termhospitalization where necessary laboratory tests, 99mTc-MAA-angiography and imaging procedures are performed. Within thissetting whole-body imaging procedures such as FDG–PET–CT orWB-MRI represent efficient tools to obtain precise and rapidassessment of total patient tumor burden instead of standard multi-session, dedicated imaging algorithms.

To our knowledge this is the first study comparing the diagnosticperformance of FDG–PET–CT and WB-MRI in a larger patient cohortwith a clearly defined diagnostic focus and therapeutic impact. Inthe presented study, 136 patients with multifocal hepatic metas-tases were triaged for radioembolisation (RE) therapy and 56% wereeligible while 44% had to be excluded. Treatment exclusion deci-sion was based on whole-body imaging results in 91% of all cases bydiagnosing substantial extrahepatic tumor manifestations. Overallspecificity for inclusion to RE therapy by combining both modal-ities was 99%, indicating the high efficacy and reliability of bothprocedures as pre-therapeutic triage modalities. Both FDG–PET–CTand WB-MRI have been proposed for various oncologic applica-tions as modern imaging methods for integrated tumor staging and

follow-up [10,21,22]. High diagnostic sensitivity, especially for thedetection of tumor recurrence in breast-, colon-, head and neckcancer as well as malignant melanoma has been reported, rang-ing between 73 and 91% for FDG–PET–CT and 72–93% for WB-MRI

G.P. Schmidt et al. / European Journal of Radiology 81 (2012) e269– e276 e273

Fig. 2. 75-Year old patient with breast cancer. (A) Fused coronal reconstruction of contrast-enhanced WB-MRI 3D-VIBE imaging of the abdomen and pelvis of a potentialr hows( al 3D-w tion.

[oesdgafmspcdbp

(rpdctPit

adioembolisation candidate. Multifocal metastatic liver disease is depicted. (B) CT sC) Fused PET–CT image confirms an avid FDG-uptrake within the metastasis. (D) Axiall (arrow). Due to imaging findings the patient was not eligible for radioembolisa

11,12,23,24]. These results correlate with our findings showing anverall patient-based sensitivity for detection of extra-hepatic dis-ase of 94% for PET–CT and 91% for WB-MRI. A reliable diagnosticensitivity is an indispensible precondition for a diagnostic proce-ure, especially in a triage situation as it is the case in our patientroup and our experiences indicate that both PET–CT and WB-MRIre suitable imaging procedures for determining inclusion decisionor RE. Although both modalities may also be used as standalone

odalities to improve cost-effectiveness of the diagnostic patientelection algorithm, results indicate a complementary role of bothrocedures as they have individual strong points of diagnostic effi-acy. In 85% of our patients extrahepatic tumor manifestations wereetected in concordance, however in 15% diagnosis was made onlyy one of both modalities meaning that real tumor extent mayotentially be missed by using only either of both procedures.

Based on organ regions PET–CT detected more malignant lesionsn = 56 anatomic regions) than WB-MRI (n = 52), which obviouslyeflects the advantage of the additional metabolic informationrovided by PET–CT. PET–CT has proven particularly efficient inetecting local tumor recurrence, especially after operative pro-edures due to its ability to differentiate scar tissue from viable

umor, for example in colorectal cancer patients [25]. Consequently,ET–CT detected more local recurrences (n = 2) than WB-MRI (n = 1)n our patient cohort, the missed lesion being a tumor recurrence athe site of a rectal anastomosis. Furthermore, it has been reported

a hypointense liver metastasis with continous infiltration beyond the liver capsule.VIBE shows the infiltrating nodular component of the metastasis into the abdominal

that PET–CT is superior to WB-MRI for the assessment of N-stageand detection of nodal tumor recurrence [10,12,23]. Especiallythe metabolic information of the PET exam in borderline sizedlesions ≤1 cm in diameter constitutes an advantage vs WB-MRIor other purely morphologic imaging modalities. Consecutively,previous experiences have described a gain in diagnostic speci-ficity and superiority of the PET–CT exam in the range of 83–95%compared to WB-MRI with 75–77% [10,11,23]. Our findings sup-port these experiences with more lymph node metastases found byPET–CT (n = 21) than by WB-MRI (n = 19). Yet, diffusion-weightedWB-MRI techniques for example promise to facilitate the detectionof nodal lesions and published studies indicate that this methodcan distinctly improve the sensitivity for lymph node metastases.However, with the exception of recent findings regarding locore-gional nodal status of rectal carcinoma, no systematic studies onpotentially improved specificity are yet available [26–28].

Regarding organ metastases previous results report a balancedperformance of FDG–PET–CT and WB-MRI. Reported sensitivitiesrange between 73 and 95% for PET–CT and 83–95% for WB-MRI[11,12,24,25]. Due to its intrinsic high bone marrow-, parenchymal-and soft-tissue contrast WB-MRI has inherent strong points in

detecting metastases from tumors which frequently spread to thebone, liver or to the brain [11,12]. Especially the use of dynamiccontrast-enhanced studies of abdominal organs, such as 3D-GRE-sequences (as implemented in our protocol) markedly enhances

e274 G.P. Schmidt et al. / European Journal of Radiology 81 (2012) e269– e276

Fig. 3. 54-Year old patient with cholangiocellular carcinoma. (A) Coronal T1w-TSE WB-MRI reveals a hypointense bone metastasis in the left minor trochanter (arrow).(B + C) PET–CT confirms an osteolysis with increased tracer uptake (arrow). (D + E) WB-MRI and PET–CT concordantly reveal pathologic para-aortic lymph nodes (arrows).Additional pathologic tracer uptake due to metastatic liver disease is indicated. (F + G) Furthermore, peritoneal carcinosis was found in this patient. In the MRI exam a tumorn etwet

lbhtrivtfhtaodns3c

odule in the rectovesical fossa is difficult to identify and is potentially missed in bhe increased tracer uptake.

esion detection and characterization [29]. For the detection ofone metastases the combination T1w-SE- and STIR-sequencesas proved to be most sensitive and allow a reliable discrimina-ion of benign from malignant marrow lesions [30]. Indeed, ouresults confirm reported literature: both PET–CT detected a sim-lar amount of distant extra-hepatic tumor manifestations (n = 33s n = 32). Especially bone metastases were found with equal sensi-ivity (n = 16 each) besides other tumor manifestations (n = 3 each),or example in the spleen. For the detection of lung metastases,owever, WB-MRI has potential limitations in terms of diagnos-ic sensitivity, compared to CT or PET–CT owing to organ motionrtifacts and the lower lung parenchyma contrast. Interestingly,ur observations describe only a moderate disadvantage for theetection of lung metastases for WB-MRI (PET–CT n = 10 vs WB-MRI

= 8). Study results indicate that the implementation of fast single

hot turbo spin echo sequences (HASTE) and contrast-enhancedD-VIBE examinations, as it was also applied in our sequence proto-ol, have significantly improved the performance of WB-MRI in the

en bowel loops (arrow). PET–CT clearly indicates the tumor formation by virtue of

detection of lung pathologies [31]. Concerning peritoneal implants,as observed in three of our patients, nodular lesions may potentiallybe missed in a purely morphologic examination such as WB-MRIas it was the case in one of our patients. Here, the increased traceruptake in tumor lesions may contribute to assess signs of peri-toneal carcinomatosis [32]. Finally, WB-MRI with its anatomicalcoverage from head to toe can potentially reveal additional find-ings of therapeutic and prognostic importance beyond the range ofa standard PET–CT staging protocol [22,23]. Indeed, we found pre-viously unknown brain metastases in two of our patients, detectedby WB-MRI only. Therefore, in the clinical setting of suspicion forbrain lesions whole-body PET–CT in combination with local brainMRI alternatively may be considered for rapid assessment of localstatus. Overall, the high concordance of 85% of detected anatom-ical sites with tumor involvement of both modalities indicates

reproducibility of our findings and diagnostic reliability of bothprocedures. Fortunately, false-positive diagnoses leading to exclu-sion from radioembolisation were described in only 2% of patients,

G.P. Schmidt et al. / European Journal of Radiology 81 (2012) e269– e276 e275

F Contrw ging sf rare p

elc

tamehbiirgthooYta

aoovaa

slF

[

ig. 4. 65-Year old patient with colon cancer. (A) T1w-TSE whole-body MRI. (B + C)ith corresponding CT imaging of the lung. (D) Contrast enhanced abdominal ima

urthermore shows a previously unknown brain metastasis (arrow). (F + G) Finally,

qually misdiagnosed by both procedures. In one case multifocalymph node metastases were falsely diagnosed by PET–CT, in otherase bone metastases were falsely described in WB-MRI.

Besides the importance of detecting substantial extrahepaticumor manifestations as exclusion criteria for radioembolisation,n important precondition of a staging method is its specificity,eaning that cases of potentially eligible patients being falsely

xcluded from therapy are kept to a minimum level. Overall aigh specificity was observed by both methods with 97% achievedy PET–CT compared to 88% by WB-MRI. Overall, specificity for

nclusion to radioembolisation therapy by combining both modal-ties was 99%, meaning that almost all eligible patients receivedadioembolisation. In almost 1 out of 10 patients (9%) angio-raphic findings made radioembolisation impossible underlininghe importance of pre-therapeutic 99mTc-MAA angiography on oneand to assess the degree of hepato-pulmonary shunting and on thether hand to reveal vascular problems, potentially complicating REr making it impossible (as it was the case in four of our patients).et, none of our patients was excluded from radioembolisation dueo a significant shunt situation, in 9% of patients RE was successfullypplied with reduced radiation dose according to guidelines [33].

A limitation of this study was the lack of a histological proof as true gold standard for the detected lesions and verification basedn radiological follow-up alone. On the other hand, as with numer-us studies of similar design, obtaining multiple biopsies for tissueerification would have been impractical and ethically unaccept-ble, especially within a clinical setting with the need of a rapidssessment of total body tumor burden [21,23,24,34].

In summary, both FDG–PET–CT and WB-MRI are efficienttaging methods for patients planned for radioembolisation ofiver metastases with a defined therapeutic impact. Overall,DG–PET–CT shows a trend to higher diagnostic accuracy com-

[

ast-enhanced 3D-GRE-sequences showing a lung metastasis in the left lower lobehows a centrally necrotic metastasis of the left adrenal gland (arrow). (E) WB-MRIresentation of an epidural metastasis is found at the level of L1 (arrow).

pared to WB-MRI and may be used as imaging method of choiceas a standalone examination as a rapid, integrative triage strategyin the pre-therapeutic setting of RE as an alternative to establishedmulti-modality algorithms.

References

[1] Schlitt HJ, Arnold D, Knoefel WT, et al. Surgical and perioperative therapy ofliver metastases. Onkologie 2008;Suppl. 5:9–13.

[2] Clark HP, Carson WF, Kavanagh PV, Ho CP, Shen P, Zagoria RJ. Staging andcurrent treatment of hepatocellular carcinoma. Radiographics 2005;25(Suppl.1):S3–23.

[3] Kennedy AS, Coldwell D, Nutting C, Murthy R, Wertman Jr DE, Loehr SP,et al. Resin 90Y-microsphere brachytherapy for unresectable colorectal livermetastases: modern USA experience. Int J Radiat Oncol Biol Phys 2006;65:412–25.

[4] Jakobs TF, Hoffmann RT, Fischer T, et al. Radioembolization in patients withhepatic metastases from breast cancer. J Vasc Interv Radiol 2007;19(5):683–90.

[5] Khodjibekova M, Szysko T, Khan S, Nijran K, Tait P, Al-Nahhas A. Selectiveinternal radiation therapy with Yttrium-90 for unresectable liver tumours. RevRecent Clin Trials 2007;2:212–6.

[6] Jakobs TF, Hoffman RT, Poepperl G, et al. Mid-term results in otherwisetreatment refractory primary or secondary liver confined tumours treatedwith selective internal radiation therapy (SIRT) using (90)Yttrium resin-microspheres. Eur Radiol 2007;17:1320–30.

[7] Pelosi E, Messa C, Sironi S, Picchio M, Landoni C, Bettinardi V, et al. Value of inte-grated PET/CT for lesion localisation in cancer patients: a comparative study.Eur J Nucl Med Mol Imaging 2004;31:932–9.

[8] Cohade C, Osman M, Leal J, Wahl RL. Direct comparison of 18F-FDG–PET andPET–CT in patients with colorectal carcinoma. J Nucl Med 2003;44:1797–803.

[9] Antoch G, Vogt FM, Freudenberg LS, Nazaradeh F, Goehde SC, Barkhausen J, et al.Whole-body dual-modality PET/CT and whole-body MRI for tumor staging inoncology. JAMA 2003;290:3199–206.

10] Schmidt GP, Baur-Melnyk A, Haug A, Heinemann V, Bauerfeind I, Reiser MF,

Schoenberg SO. Comprehensive imaging of tumor recurrence in breast cancerpatients using whole-body MRI at 1.5 and 3 T compared to FDG–PET–CT. Eur JRadiol 2008;65:47–58.

11] Schmidt GP, Baur-Melnyk A, Haug A, Becker CR, Tiling R, Utzschneider S,et al. Whole-body MRI at 1.5 and 3 Tesla compared to FDG–PET–CT for the

e rnal of

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

276 G.P. Schmidt et al. / European Jou

detection of tumor recurrence in patients with colorectal cancer. Eur Radiol2009;19:1366–78.

12] Brix G, Lechel U, Glatting G, Ziegler SI, Muenzing W, Mueller SP, et al. Radiationexposure of patients undergoing whole-body dual-modality 18F-FDG–PET/CTexaminations. J Nucl Med 2005;46:608–13.

13] Griswold MA, Jakob PM, Heidemann RM, Nittka M, Jellus V, Wang J, et al. Gen-eralized autocalibrating partially parallel acquisitions (GRAPPA). Magn ResonMed 2002;47:1202–10.

14] Glazer GM, Gross BH, Quint LE, Francis IR, Bookstein FL, Orringer MB. Normalmediastinal lymph nodes: number and size according to American thoracicsociety mapping. AJR 1985;144:261–5.

15] Dorfman RE, Alpern MB, Gross BH, Sandler MA. Upper abdominal lymph nodes:criteria of normal size determined with CT. Radiology 1991;180:319–22.

16] Vanel D, Bittoun J, Tardivon A. MRI of bone metastases. Eur Radiol1998;8:1345–51.

17] Danet IM, Semelka RC, Leonardou P, Braga L, Vaidean G, Woosley JT, et al. Spec-trum of MRI appearances of untreated metastases of the liver. Am J Roentgenol2003;181:809–17.

18] Beggs AD, Hain SF, Curran KM, O’Doherty MJ. FDG–PET as a “metabolic biopsy”tool in non-lung lesions with indeterminate biopsy. Eur J Nucl Med Mol Imaging2002;29:542–6.

19] Sato KT, Lewandowski RJ, Mulcahy MF, Atassi B, Ryu RK, Gates VL, et al.Unresectable chemorefractory liver metastases: radioembolization with 90Ymicrospheres—safety, efficacy, and survival. Radiology 2008;247:507–15.

20] Pwint TP, Midgley R, Kerr DJ. Regional hepatic chemotherapies in the treatmentof colorectal cancer metastases to the liver. Semin Oncol 2010;37:149–59.

21] Schlemmer HP, Schäfer J, Pfannenberg C, Radny P, Korchidi S, Müller-HorvatC. Fast whole-body assessment of metastatic disease using a novel mag-netic resonance imaging system: initial experiences. Invest Radiol 2005;40:64–71.

22] Schmidt GP, Baur-Melnyk A, Herzog P, Schmid R, Tiling R, Reiser MF, et al. High-resolution whole-body MRI tumor staging with the use of parallel imaging

versus dual modality PET–CT: experience on a 32-channel system. Invest Radiol2005;40:743–53.

23] Ng SH, Chan SC, Yen TC, Liao CT, Chang JTC, Ko SF, et al. Comprehensive imagingof residual/recurrent nasopharyngeal carcinoma using whole-body MRI at 3 Tcompared with FDG–PET–CT. Eur Radiol 2010;April 15 [Epub ahead of print].

[

[

Radiology 81 (2012) e269– e276

24] Laurent V, Trausch G, Bruot O, Olivier P, Felblinger J, Régent D. Com-parative study of two whole-body imaging techniques in the case ofmelanoma metastases: advantages of multi-contrast MRI examination includ-ing a diffusion-weighted sequence in comparison with PET–CT. Eur J Radiol2009;June 2 [Epub ahead of print].

25] Even-Sapir E, Parag Y, Lerman H, et al. Detection of recurrence in patients withrectal cancer: PET/CT after abdominoperineal or anterior resection. Radiology2004;232:815–22.

26] Takahara T, Imay Y, Yamashita T, Yasuda S, Nasu S, Van Cauteren M. Diffu-sion weighted whole body imaging with background body signal suppression(DWIBS): technical improvement using free breathing, STIR and high resolution3D display. Radiat Med 2004;22:275–82.

27] Lichy MP, Aschoff P, Plathow C, Stemmer A, Horger W, Mueller-Horvat C, et al.Tumor detection by diffusion-weighted MRI and ADC-mapping—initial clinicalexperiences in comparison to PET–CT. Invest Radiol 2007;42:605–13.

28] Rao SX, Zeng MS, Chen CZ, et al. The value of diffusion-weighted imaging incombination with T2-weighted imaging for rectal cancer detection. Eur J Radiol2008;65:299–303.

29] Semelka RC, Worawattanakul S, Kelekis NL, et al. Liver lesion detection, char-acterization, and effect on patient management: comparison of single phasespiral-CT and current MR techniques. J Magn Reson Imaging 1997;7:1040–7.

30] Mehta RC, Marks MP, Hinks RS, Glover GH, Enzmann DR. MR evaluation ofvertebral metastases: T1-weighted, short-inversion-time inversion recovery,fast spin-echo, and inversion-recovery fast spin-echo sequences. AJNR Am JNeuroradiol 1995;16:281–8.

31] Frericks BB, Meyer BC, Martus P, Wendt M, Wolf KJ, Wacker F. MRI of the thoraxduring whole-body MRI: evaluation of different MR sequences and compari-son to thoracic multidetector computed tomography (MDCT). J Magn ResonImaging 2008;27:538–45.

32] Satoh Y, Ichikawa T, Motosugi U, Kimura K, Sou H, Sano K, Araki T. Diagnosis ofperitoneal dissemination: comparison of 18F-FDG PET/CT, diffusion-weightedMRI, and contrast-enhanced MDCT. AJR Am J Roentgenol 2011;196:447–53.

33] Ho S, Lau WY, Leung TW, Chan M, Johnson PJ, Li AK. Clinical evaluation of thepartition model for estimating radiation doses from yttrium-90 microspheresin the treatment of hepatic cancer. Eur J Nucl Med 1997;24:293–8.

34] Brennan DD, Gleeson T, Coate LE, Cronin C, Carney D, Eustace SJ. A comparisonof whole-body MRI and CT for the staging of lymphoma. AJR 2005;185:711–6.