HEAD AND NECK

MRI in the evaluation of facial dermal fillers in normaland complicated cases

Marco Di Girolamo & Mauro Mattei & Alberto Signore &

Francesca Romana Grippaudo

Received: 20 June 2014 /Revised: 1 October 2014 /Accepted: 17 November 2014# European Society of Radiology 2014

AbstractObjectives To ascertain by MRI the presence of filler injectedinto facial soft tissue and characterize complications by con-trast enhancement.Methods Nineteen volunteers without complications were ini-tially investigated to study the MRI features of facial fillers.We then studied another 26 patients with clinically diagnosedfiller-related complications using contrast-enhanced MRI.TSE-T1-weighted, TSE-T2-weighted, fat-saturated TSE-T2-weighted, and TIRM axial and coronal scans were performedin all patients, and contrast-enhanced fat-suppressed TSE-T1-weighted scans were performed in complicated patients, whowere then treated with antibiotics. Patients with soft-tissueenhancement and those without enhancement but who didnot respond to therapy underwent skin biopsy. Fisher’s exacttest was used for statistical analysis.

Results MRI identified and quantified the extent of fillers.Contrast enhancement was detected in 9/26 patients, and skinbiopsy consistently showed inflammatory granulomatous reac-tion, whereas in 5/17 patients without contrast enhancement,biopsy showed no granulomas. Fisher’s exact test showedsignificant correlation (p<0.001) between subcutaneous con-trast enhancement and granulomatous reaction. Cervical lymphnode enlargement (longitudinal axis >10 mm) was found in 16complicated patients (65 %; levels IA/IB/IIA/IIB).Conclusions MRI is a useful non-invasive tool for anatomicallocalization of facial dermal filler; IV gadolinium administra-tion is advised in complicated cases for characterization ofgranulomatous reaction.Key Points• MRI is a non-invasive tool for facial dermal filler detectionand localization.

•MRI-criteria to evaluate complicated/non-complicated casesafter facial dermal filler injections are defined.

• Contrast-enhanced MRI detects subcutaneous inflammatorygranulomatous reaction due to dermal filler.

• 65% patients with filler-related complications showed lymph-node enlargement versus 31.5 % without complications.

• Lymph node enlargement involved cervical levels (IA/IB/IIA/IIB) that drained treated facial areas.

Keywords Dermal filler .MRI . Contrast-enhancedMRI .

Granulomas . Lymph node enlargement

Introduction

Cosmetic tissue augmentation and correction of skin depres-sions using injectable material is not a new concept. In the lastdecade, many new materials have been introduced, claimingto be permanent, inert, nonallergenic, well-tolerated, non-

M. Di Girolamo :M. MatteiRadiology Unit, Faculty of Medicine and Psychology, “Sapienza”University of Rome, Rome, Italy

A. SignoreNuclear Medicine Unit, Faculty of Medicine and Psychology,“Sapienza” University of Rome, Rome, Italy

A. SignoreDepartment of Nuclear Medicine andMolecular Imaging, UniversityMedical Center Groningen, University of Groningen, Groningen,The Netherlands

F. R. GrippaudoPlastic Surgery Unit, Faculty of Medicine and Psychology,“Sapienza” University of Rome, Rome, Italy

M. Di Girolamo (*)Department of Radiology – Sant’Andrea Hospital, Faculty ofMedicine and Psychology, Sapienza University of Rome, Via diGrottarossa 1035, 00189 Rome, Italye-mail: digirolamomarco@hotmail.com

Eur RadiolDOI 10.1007/s00330-014-3513-2

migrating and easily removable in the rare event of complica-tions [1, 2].

Dermal fillers differ in composition and are classified ac-cording to their soft tissue persistence.

Temporary dermal fillers comprise collagen andhyaluronic acid (HA), which are reabsorbed by the body over6–12 months [3].

Long-lasting fillers include poly-L-lactic acid (PLLA)(Sculptra®) [4], synthetic calcium hydroxyapatite micro-spheres (CaHA) suspended in sodium carboxymethyl cellu-lose gel (Radiesse®) [5], and dextran molecules (Sephadex®)and hyaluronic acid (Matridex®), which remain for 2–3 yearsprior to degradation [3].

Permanent dermal fillers include poly(methyl methacry-late) (PMMA), with or without a collagen vector (Artecoll®,Metacrill®) [6], polyacrylamide hydrogel (PAAG)(Aquamid®, Royamid®, Formacryl®) [7], polyalkylimide(PAIG) (BioAlcamid®) [8], acrylic hydrogel (AH)(Dermalive®) [9], and silicone particles suspended in apolyvinilpirrolidone carrier (Bioplastique®) [10]. All of thesesubstances are designed to be permanently encapsulated bythe body’s connective tissue [11]. The cosmetic use of liquidsilicone injections has been banned in several countries, but isstill seen in the elderly and in patients in non-Western coun-tries [12, 13]. It is used off-label in the U.S., primarily for HIV-associated lipoatrophy but also for cosmetic purposes [14,15].

Despite the minimally invasive nature of dermal fillers,there are complications associated with them [16].

Short-term complications usually occur within the first fewweeks after treatment and include bleeding, bruising, infections,redness at the injection site and oedema, usually resolving withina few days. There is always a risk of technical errors fromovercorrection or injection of the product in the wrong site.

Long-term complications (after six months) include theformation of foreign-body abscesses and granulomas, infec-tions, product migration, and cross-reactions among differentproducts injected in nearby sites, producing diffuse oedema,skin discoloration, lumps and ulcers [17–20].

Comprehensive epidemiologic data on the rates of compli-cation associated with these treatments are not available due tothe multitude of fillers in the world market, the small numberof national registries, and—when present—the voluntary na-ture of adverse event reports [21]. Overall complication ratesof dermal fillers are based mainly on case report papers andvary from 3 % after PMMA [22] to 52 % after PLLA, withother fillers within this range [23].

Patients seeking facial cosmetic augmentation are not al-ways aware of the material that was used previously bydifferent physicians. In these patients, high-frequency ultra-sound (HFUS) is a useful tool for detection and characteriza-tion of dermal fillers in soft tissues prior to further injections[24–27].

MRI is a multiplanar and multiparametric diagnostic tech-nique with high intrinsic contrast, allowing evaluation of softtissues without exposure to ionizing radiation, although morecostly in comparison to HFUS. Moreover, MRI obtains goodspatial evaluation of the actual site of any foreign body withrespect to anatomical landmarks [28].

The aim of this study was to investigate the role of MRI fordetection of dermal fillers injected into the facial soft tissues forcosmetic purposes, and to assess any complications related to thisprocedure. In light of its valuable role in inflammatory reactions[29], it seemed logical to investigate the application of intrave-nous administration of paramagnetic contrast media for the eval-uation of inflammatory complications after filler injection.

Materials and methods

Patient selection

To clarify the role of MRI in dermal filler detection, betweenJanuary 2009 and March 2010, a total of 19 asymptomaticpatients referred from the plastic surgery unit after temporary orpermanent facial filler injection for cosmetic purposes (Group 1)were recruited. None had clinical evidence of complications orcontraindications to MRI. Ten patients had received injections oftemporary dermal fillers from four to sevenmonths preceding theclinical evaluation, and the remaining nine patients had receivedinjections of permanent dermal fillers from nine months to10 years preceding the examination.

To assess the role of contrast-enhanced MRI, from January2010 through June 2013, a total of 26 patients with local com-plications (erythema and swelling of different intensity and du-ration, skin ulcers) were referred after a dermal filler treatment ofknown (20 patients) or unknown (6 patients) nature, performedfrom 3 to 12 years previously (Group 2). Twelve of these patientshad received inoculations of different fillers.

Patients in both groups underwent head and neck MRI. Allcomplicated patients (Group 2) received therapy with ceftri-axone (Rocefin) 1 g intramuscularly for seven days, combinedwith ciprofloxacin (Ciproxin) 500 mg twice daily for threeweeks [30].

The protocol of this prospective study was approved by thelocal ethics committee (Prot. CE: 554/2012), and all patientssigned informed consent documentation and gave permissionfor publication of their photos.

MRI

All of the examinations were performed with a 1.5 Teslasuperconductive MR scanner (MAGNETOM Sonata, SiemensHealthcare, Erlangen, Germany), using a seven-channel headand neck coil. The MR examination was conducted from the

Eur Radiol

frontal bone to the neck base, including all six cervical lymphnode levels, excluding only the supraclavicular region.

Axial and coronal TSE T1-weighted (acquisition parame-ters: TR, 600 ms; TE, 11 ms; ETL, 5) and T2-weighted scans(acquisition parameters: TR, 3600 ms; TE, 108 ms; ETL, 19)were performed with a slice thickness of 3 mm.

TSE T2-weighted axial scans with fat saturation wereperformed (acquisition parameters: TR, 3600 ms; TE,108 ms; ETL, 19) and turbo-inversion-recovery-magnitude(TIRM) sequences in the axial plane with 3 mm slice thick-ness were subsequently obtained using the following acquisi-tion parameters: TR, 91.20 ms; TE, 67 ms; TI, 150 ms.

In Group 2 only, fat-suppressed TSE T1-weighted axialand coronal scans were performed after the IV administrationof 0.1 mmol/kg gadolinium (Gd)-DOTA (Dotarem; GuerbetGroup, Paris, France). MR acquisition started 2-3 minutesafter the IV injection of contrast media.

Image interpretation, skin biopsy, and statistical analysis

MRI was interpreted as follows:

1) Qualitative evaluation of the filler signal intensity on T1-and T2-weighted images in comparison with subcutane-ous fat tissue (hyperintense, isointense, or hypointensesignal)

2) Evaluation of the quantity and site of filler injection usingnormal anatomical landmarks

3) Qualitative analysis concerning the presence or absenceof contrast enhancement at the level of subcutaneous fattissue in complicated patients only (Group 2), reported aspositive or negative

4) Detection of enlarged cervical lymph nodes, consideredpositive when their longitudinal axis was greater than10mm; their level was established using the classificationas established by the American Academy of Otolaryngol-ogy —Head and Neck Surgery (AAO-HNS) and theAmerican Joint Committee on Cancer (AJCC) [31].

The MR examinations were evaluated via consensus bytwo radiologists (MDG and MM) with more than 15 years ofexperience in head and neck pathology, who were unaware ofthe nature of the filler injected but were aware of the compli-cated cases when evaluating the contrast-enhanced images.

Cutaneous biopsy was performed by a plastic surgeon(FRG) in all patients showing positive enhancement after IVadministration of Gd-DOTA and in patients without any con-trast enhancement but with no clinical improvement afterthree weeks of antibiotic therapy, as determined by our localethics committee. Cutaneous biopsy was not performed inpatients with no contrast enhancement on MRI and withclinical improvement after antibiotic therapy.

Statistical analysis of the relationship between subcutaneouscontrast enhancement after Gd-DOTA and histology was per-formed using Fisher’s exact test. Data was analysed using statis-tical software (SigmaPlot version 12.0; Systat Software Inc.).

Results

AllMR examinations were completed within 30minutes, withno patient discomfort or claustrophobia.

MRI detected the presence of dermal filler in all patients,even without clinical evaluation. Three patients presentedferromagnetic artefacts on some images due to the presenceof dental implants, partially impairing the diagnostic results.

MRI consistently and clearly demonstrated the site of fillers,allowing precise anatomical localization, visualizing anatomi-cal landmarks (mandible, lips, nose, zygomatic region, orbitsand glabella) on axial and coronal scans. MRI allowed themeasurement of filler, assessing the extent and depth in facialsoft tissues, with minimum size detected of 2 mm.

In Group 1 patients with temporary fillers, we observedround liquid droplets. HA and C appeared as multiplehypointense spots on T1-weighted images and hyperintenseon T2-weighted images because of their water content(Fig. 1). In normal volunteers with permanent fillers (PAAG,PAIG, S and AH), these appeared as hypointense spots on T1-weighted images, with variable signal intensity on T2-weightedimages. The T2-weighted signal was hyperintense in five pa-tients (three had PAAG and two had PAIG) and hypointense infour patients (two had S and two had AH) (Table 1).

MRI detected cervical lymph node enlargement in 6 of 19(31.5 %) asymptomatic volunteers, and the enlarged lymphnodes had a longitudinal axis greater than 10 mm [32]. Thelymph node level was IA in one volunteer (bilateral), IB intwo (bilateral in one and unilateral in one), IIA and IIB in two(bilateral), III in one (bilateral), and V in a single volunteer(monolateral and retroauricular) (Table 1).

On T2-weighted images in Group 2 (Table 2), fillers ap-peared with a hyperintense signal in 11 patients, a hypointensesignal in 6 patients, and with both hyperintense andhypointense lesions in 9 patients due to different dermal fillersin the same patient.

After the injection of Gd-DOTA, 9 of 26 patients (34.6 %)presented areas of enhancement of the facial subcutaneous fattissues. Subsequent skin biopsy ascertained an inflammatorygranuloma at the level of the enhanced area (Figs. 2 and 3). In5 of 17 patients with filler-related complications but with noGd-enhancement and with no clinical improvement after an-tibiotic therapy, biopsy did not demonstrate a granulomatousinflammatory reaction (Fig. 4); in three of these patients,biopsy showed fibrotic tissue.

Eur Radiol

Fischer’s exact test revealed a significant correlation (p<0.001) in complicated cases between facial subcutaneous con-trast enhancement and granulomatous reaction detected withskin biopsy. Fischer’s exact test considered only 14 patients: 9with subcutaneous contrast enhancement and detection ofinflammatory granulomatous reaction and 5 without contrastenhancement and no histological finding of granuloma.

In 17 of the 26 Group 2 patients (65.3 %), MRI detectedcervical lymph node enlargement with a longitudinal axisgreater than 10 mm [32].

The lymph node level was IA in one patient (bilateral), IBin seven (bilateral in six and unilateral in one), IIA in 13 (allbilateral), and IIB in 10 patients (all bilateral) (Table 2). Thelargest lymph node had a longitudinal axis of 26 mm. Theenlarged lymph nodes all had a longitudinal and axial axisratio greater than 2 and none showed the same signal intensity

as the injected subcutaneous filler. In two patients who hadreceived silicone injections, no silicone was detected withMRI in the enlarged lymph nodes. None of the lymph nodesshowed gadolinium enhancement and thus were interpreted assites of previous inflammatory reactions.

MRI allowed the detection of incidental pathologies in fivepatients (one frontal meningioma, one temporal arachnoidcyst, one chronic ischemic cerebral gliosis, one frontal areaof malacia secondary to a previous head trauma, and oneapical dental granuloma).

Discussion

Dermal filler-related complications are a frequent pathology[16], often presenting several years after the original cosmetic

Fig. 1 (A, B, C, D): Group 1patient with temporary dermalfiller (HA) in the zygomaticregion and nasolabial folds. Onaxial FSE T1-weighted scan (TR:600–800 ms, TE: 11 ms, ETL: 5)(A) HA appears with ahypointense signal (white arrow),while on axial FSE T2-weightedscans with or without fatsaturation (TR: 2750 ms, TE:108 ms, ETL: 19) (B and C) andTIRM acquisitions (TR:9000 ms, TE: 67 ms, TI: 150 ms,ETL: 11) (D), HA appears with ahomogeneous hyperintense signal(white arrow)

Eur Radiol

treatment. Because of the time span between treatment andcomplications, it is common to deal with patients who do notremember the filler they received, do not have the productleaflet and have lost contact with the performing physician.Knowledge of the site, size and nature of the filler improvesthe chances of successful treatment.

CT imaging has been proposed as a valuable tool to eval-uate dermal fillers [33], but in our opinion, has the disadvan-tage of ionizing radiation exposure to critical organs (crystal-line lens).

HFUS is a reliable, diffuse and economic diagnostic toolthat is useful in evaluating the site, extension, and amount ofinjected fillers. With this procedure, it is possible to ascertainthe temporary or permanent nature of the product and tosuggest a diagnosis of granuloma [24, 25]. Disadvantagesinclude the lack of certain anatomical landmarks in evaluationby the plastic surgeon and the absence of consolidated criteriato diagnose inflammatory reactions. Moreover, HFUS is anoperator-dependent investigative method, not allowing a sec-ond opinion in the evaluation of diagnostic images. In contro-versial cases, therefore, HFUS evaluation may require thecollaborative presence of both radiologist and plastic surgeonduring the examination. Nevertheless, because of its highavailability, HFUS is often the first-line diagnostic examina-tion performed in these patients. Considering the limits ofHFUS, MRI could be proposed as second-line diagnostic

examination, despite its higher costs and reduced availability.The advantages of MRI are its ability to provide excellent softtissue evaluation with multi-parametric diagnostic images[28].

In our study, TSE T1- and T2-weighted axial and coronalscans were used, with a total acquisition time of 20 minutes.The only drawbackwas the reduction in contrast-to-noise ratiobetween the hyperintense signal of subcutaneous fat tissue andfiller on TSE T2-weighted images. This problem can beminimized using particular fat saturation acquisitions thatnullify the subcutaneous fat tissue’s signal such that the signalfrom other tissue is more conspicuous.

There are various techniques for achieving fat saturationwithMRI [34]. All techniques for fat suppression are based onthe fact that, due to the different chemical environment, hy-drogen nuclei in water and in fat tissue have different valuesfor some MRI-relevant parameters, primarily the relaxationtime and the resonance frequency (chemical shift). The mostcommon method for achieving fat saturation is by applying anarrow-band frequency-selective RF pulse at the beginning ofany sequence and following it immediately with a spoiler orcrusher gradient that shifts the net magnetization vector of fatso that it has no longitudinal magnetization at the beginning ofthe MR acquisition. This method was used in our fat-saturatedT2-weighted and contrast-enhanced fat-saturated T1-weighted acquisitions. The second technique for obtaining

Table 1 Types of filler, sites of injection and MRI findings in normal volunteers with facial dermal fillers

PTS Type of filler T2-w. signal intensity Sites of injection Lymph node enlargement and site

Female HA Hyper Lips No

Female HA Hyper Nasolabial folds Right IB

Female HA Hyper Nasolabial folds No

Female HA Hyper Nasolabial folds, glabella No

Female HA Hyper Nasolabial folds No

Female HA Hyper Lips, nasolabial folds Bilateral IIA, IIB, III

Female HA Hyper Lips, nasolabial folds, zygomas No

Female HA Hyper Lips No

Female C Hyper Lips, glabella, nasolabial folds Bilateral IIA, IIB

Female C Hyper Lips No

Female PAAG Hyper Nasolabial folds Right V

Female PAAG Hyper Lips No

Female PAAG Hyper Lips No

Female PAIG Hyper Lips No

Female PAIG Hyper Nasolabial folds, glabella Bilateral IA, IB

Female S Hypo Lips, nasolabial folds Bilateral IIA, IIB, III

Male AH Hypo Nasolabial folds, lips No

Female AH Hypo Zygomas, lip No

Legend

Temporary fillers: HA Hyaluronic acid; C Collagen

Permanent Fillers: S Silicone; PAAG Polyacrylamide gel; PAIG Polyalkylimide gel; AH Acrylic hydrogel particles, copolymer of 40 % hydroxyl-ethyl-methacrylate (HEMA) and ethyl-methacrylate (EMA) and HA 60 %

Eur Radiol

Tab

le2

Typesof

filler,sitesof

injection,MRIfindings,biopsyandhistologicalspecim

ensin

patientswith

localcom

plications

afterfacialfillerinjections

PTS

Type

offiller

T2-w.signal

intensity

Sitesof

injection

Lymph

node

enlargem

ent

andsite

Contrastenhancement

afterGadolinium

BiopsyandHistology

Occasionalfindings

♀HA

Hyper

Zygom

asNasolabialfolds

No

No

No

Chronicischem

iccerebral

gliosis

♀HA

Hyper

Zygom

asNasolabialfolds

BilateralIB,IIA

,IIB

No

Amorphoustissue

Leftfaceskin

rednessand

blisters

♀HA

Hyper

Nasolabialfolds

No

No

No

♀HA

SHyper

Hypo

Nasolabialfolds

Lips

Right

IB,B

ilateralIIB

No

No

♀HA

PAIG

Hyper

Lips

BilateralIB

No

No

♂PLLA

HA

Hypo

Hyper

Glabella

Nasolabialfolds

BilateralIB

No

No

♀HA

PAAG

Unknown

Hyper

Hyper

Hypo

Cheeks

Lips

Zygom

as

BilateralIIA

,IIB

Yes

Inflam

matorygranulom

aRight

skin

ulcer

Lupus

andlichen

♀PA

AG

Hyper

Lips

BilateralIIA

,IIB

Yes

Inflam

matorygranulom

aLeftfrontalmeningiom

a

♀PA

AG

Hyper

Glabella

No

No

Amorphoustissue

Post-traumaticmalacia

frontalarea

♀PA

AG

S HA

Hyper

Hypo

Hyper

Zigom

asx,lips

Zigom

adx,chin

Zygom

as

BilateralIB,IIA

,IIB

No

No

Dentalimplan

tMRartefacts

♀CaH

AHA

HypoHyper

Zygom

asNo

No

No

Dentalimplan

tMRartefacts

♀PA

IGC

Hyper

Hyper

Zygom

asandcheeks

Lips

BilateralIIA

No

Fibroticreactio

nDentalimplan

tMRartefacts

Lefttem

poralarachnoid

cyst

♀PA

IGHypo

Zygom

asandcheeks

No

Yes

Inflam

matorygranulom

aRight

dentalapicalgranulom

a

♀HA

PAIG

Hyper

Hyper

Glabella

andNasolabialfolds

Zygom

as,L

ipsandCheeks

BilateralIIA

,IIB

No

Fibroticreactio

nDentalimplan

tMRartefacts

♀S

Hypo

Lips

No

No

No

♀S HA

Hypo

Hyper

Lips

Zygom

asandNasolabialfolds

No

No

No

♀HA

AH

Hyper

Hypo

Nasolabialfolds

LipsandGlabella

BilateralIIA

Yes

Inflam

matorygranulom

a

♀Bioplastiq

ueAH

Unknown

Hyper

Hypo

Hypo

Nasolabialfolds

Zygom

aLow

erlip

BilateralIA,IB,IIA

,IIB

(26mm)

No

Yes

No

Inflam

matorygranulom

a

♀AH

Hypo

Nasolabialfolds,L

ipsandGlabella

BilateralIIA

,IIB

Yes

Inflam

matorygranulom

a

♀AH

Hypo

Lips

No

No

No

♀Unknown

Unknown

Hyper

Hyper

Lips

BilateralIIA

,IIB

No

No

♂Unknown

Hyper

Zygom

asNo

Yes

Inflam

matorygranulom

a

♂Unknown

Hyper

Nasolabialfolds

BilateralIIA

Yes

Inflam

matorygranulom

a

Eur Radiol

fat saturation is an inversion-recovery pulse to nullify thesignal from fat. TIRM sequences are specific T2-weightedinversion-recovery acquisitions that are more efficient fornullifying the signal from fat and that allow optimal detectionof fillers within subcutaneous fat tissue. This acquisition isalso used to nullify the signal from silicone, with an inversiontime of 150 ms [35].

Reports have appeared sporadically in the literature inwhich MRI has been proposed for the evaluation ofcosmetic filler [7, 33, 36, 37] but the increase in theuse of these cosmetic procedures and subsequent facialcomplications necessitate improvements in the applica-tion of this modality.

Our study confirmed that HA appeared with a hyperintensesignal on T2-weighted images, as reported by previous studies[36, 37] in which MRI was performed four times on eachpatient. In these studies, the signal intensity of HA on T2-weighted images decreased progressively with time due toproduct reabsorption. In our study, the MR examination wasperformed just once on each patient, and therefore it wasimpossible to demonstrate the signal intensity variations onT2-weighted images.

Bello et al. [7] reported an MRI assessment of polyacryl-amide gel (PAAG) implanted in rabbit ears, which showed nofiller dislocation and absence of inflammatory reaction after afollow-up of seven months. OurMRI findings showed region-al lymph node enlargement in three of four patients withcomplications after PAAG injections and showed subcutane-ous contrast enhancement in two of those patients (Table 2).MRI detected dislocation of PAAG filler in the only patientwith no other MRI pathological finding.

MRI—without exposing patients to ionizing radiation—offers the plastic surgeon spatial visualization on axial andcoronal scans of the site and extent of the filler as well as theanatomical landmarks, a result that is not possible withultrasound.

MRI has another important role as a diagnostic tool inpatients with complications. As demonstrated by Paajaneenet al., this diagnostic modality can accurately identify thepresence of foreign material in soft tissues and can detect softtissue inflammation in vivo using IV administration of gado-linium [29].

The present study, with the use of Fischer’s exact test,statistically demonstrated that gadolinium enhancement ofsoft tissue observed in complicated cases was related to agranulomatous reaction, excluding all other complications.This is the first time that this correlation has been reportedand this finding is fundamental for the treatment of fillercomplications (Fig. 5). A previous study by Kransdorf et al.[38] reported a diagnosis of granulomatous inflammatoryreaction based on soft tissue enhancement after IV adminis-tration of a paramagnetic contrast agent in a case of subcuta-neous granuloma annulare.T

able2

(contin

ued)

PTS

Type

offiller

T2-w.signal

intensity

Sitesof

injection

Lymph

node

enlargem

ent

andsite

Contrastenhancement

afterGadolinium

BiopsyandHistology

Occasionalfindings

♀Unknown

Hypo

Lips

BilateralIB,IIA

No

No

♀Unknown

Hypo

Nasolabialfolds

BilateralIIA

,IIB

Yes

Inflam

matorygranulom

a

♀UnknownUnknown

Hyper

Hypo

Nasolabialfolds

No

No

Fibroticreactio

n

Legend

HA

Hyaluronicacid,C

Collagen,

SSilicone,PA

AG

Polyacrylam

idegel,PLLA

Poly-L-lactic

acid,PA

IGPolyalkylim

idegel,PMMA

Poly(methylmethacrylate)

microspheres,

CaH

ACalcium

hydroxylapatite,B

ioplastiq

ueSiliconeandpolyvinilpirrolid

one,AHAcrylichydrogelparticles,copolymer

of40

%hydroxyl-ethyl-m

ethacrylate(H

EMA)andethyl-methacrylate(EMA)andHA60

%

Eur Radiol

As HFUS is not capable of diagnosing certain granulomatousinflammatory responses after filler injection, the higher costs ofMRI are justified. The diagnosis of severe granulomatous reac-tions after filler injections is critical for subsequent treatmentplanning and provides evidence, if needed, for legal action.

MRI allows filler characterization and evaluation of adjacenttissue modification after in vivo injection, particularly with theuse of T2-weighted acquisition, as shown by Gensanne et al.[36]. The signal intensity of dermal fillers on T2-weighted im-ages may be different within the same subject as a result of theinjection of different products in different periods and possiblegranulomatous inflammatory or fibrotic reactions. In such situa-tions, the use of paramagnetic contrast media is required in orderto differentiate fibrosis from granulomatous reaction, since

fibrosis does not show significant enhancement after IV admin-istration of paramagnetic contrast media.

A report by Feeney [39] demonstrated an inflammatoryreaction related to hydroxylapatite dermal fillers using FDGPET/CT and MRI in neoplastic patients who received fillerinjections after treatment. Feeney found that intense facialsubcutaneous FDG uptake was associated with the presenceof high-attenuation material (600-700 HU) in five patientsevaluated by PET/CT, with mild enhancement on thepost-gadolinium sequence in only one patient evaluatedby MRI, describing inflammatory reactions in the site ofinjection.

Scintigraphy with radiolabelled white blood cells (WBC)was recently found to be the most accurate method for

Fig. 2 (A, B, C, D): Group 2 patient with temporary dermal fillers (HA)injected 7 months previously, with lumps at the level of superior lip andglabella (A). The lumps show hyperintense signal on TIRM acquisitions(TR: 9000 ms, TE: 67 ms, TI: 150 ms, ETL: 11) (B) (white arrows) andstrong contrast enhancement on TSE T1-weighted fat-saturatedacquisition (TR: 487 ms, TE: 11 ms, ETL: 5, TI: 150 ms) after IV

administration of Gd-DOTA (C) (white arrows) on the same axial planeat the level of the glabella. Histological specimens stained with H&E (D)show multiple spaces, optically empty or with material, surrounded bymultinucleated giant cells (small white arrows) and by histiocytes,forming some foreign-body granulomas

Eur Radiol

diagnosing infection in patients with long-term dermal fillercomplications [30]. The main limitation ofMRI is difficulty indiagnosing a subcutaneous infection. In cases of clinical sus-picion of infection, or to exclude an infective process, scintig-raphy with radiolabelled WBC could be included in the man-agement of such patients [40].

We found neck lymph node enlargement in 17 of 26(65 %) patients with complications and in 6 of 19(31.5 %) normal volunteers with facial dermal fillers.Lymph node enlargement was noted at the IA, IB, IIA,and IIB cervical levels in complicated cases. These lymphnodes were all oval in shape (longitudinal and axial axis

ratio greater than 2), with a longitudinal diameter up to10 mm and a smooth surface, confirming their reactiveaspect [32]. Lymph node enlargement is not specific todermal filler complications, and could also be related toprevious ENT (ear, nose, throat) infections, but the higherpercentage of enlarged lymph nodes characteristicallydraining the treated facial areas in pathological versusnormal cases [41] is highly suggestive of a positive corre-lation with facial dermal filler complications. After everyepisode of lymph node functional hyperplasia following animmune response, the cellular components (macrophages,dendritic and Langerhans cells, and B and T lymphocytes)

Fig. 3 (A, B, C, D): Group 2patient with cutaneous fistula inright zygomatic region (whitecircle) (A). MRI shows a diffusedeep fat tissue hypointense signal(white arrow) on TSE T1-weighted axial scan (TR: 600–800 ms, TE: 11 ms, ETL: 5) (B),while TSE T2-weighted axialscan (TR: 2750 ms, TE: 108 ms,ETL: 19) shows bothhyperintense and hypointensespots (white arrow) (C) in rightzygomatic region. TSE fat-saturated T1-weighted axial scanon the same plane (TR: 487 ms,TE: 11 ms, ETL: 5, TI: 150 ms)(D) performed after IV injectionof Gd-DOTA shows widespreadcontrast enhancement of thesubcutaneous fat tissue of theright zygomatic region (D) and asubsequent skin biopsydemonstrates a diffusegranulomatous inflammatoryreaction

Eur Radiol

return to normal proportions, while the stromal hyperplasiaremains [42].

The survey of different signal intensities between enlargedlymph nodes and injected subcutaneous fillers suggests theabsence of foreign-body reaction and the absence of lymphnode contrast enhancement excludes lymphadenitis. In thisstudy, patients were not selected for surgical removal of en-larged lymph nodes, as it was not necessary for clinicalpurposes, nor was it allowed by the local ethics committeeauthorization.

Cooperation between radiologists and plastic surgeons isimperative for achieving the best diagnostic MRI results andoffering optimal and customized care to patients presentingcomplications after filler injections for cosmetic purposes.

MRI distinguishes among complications (migration, granulo-matous reaction and fibrosis), providing valuable informationfor treatment planning. MR multiplanar acquisition definingprecise anatomical landmarks offers the best imaging modal-ity to evaluate filler migration. The IVadministration of para-magnetic contrast media allows the differential diagnosis be-tween inflammatory granulomatous reaction and subcutane-ous fibrosis, which does not show significant contrast en-hancement (Fig. 5). Based on diagnostic imaging results,when dislocation occurs, surgery is planned. In cases of gran-ulomas, medical therapy is our first choice, and if no improve-ment occurs, surgical removal is the next option. In cases offibrosis, a lipofilling procedure and/or surgical removal areindicated, depending on the patient’s wishes.

Fig. 4 (A, B, C, D):Group 2 patient with permanent dermal filler (PAIG)on both nasolabial folds and zygomatic region evaluated with TSE T1-weighted (TR: 600–800 ms, TE: 11 ms, ETL: 5) (A) and TSE T2-weighted fat saturated (TR: 2750 ms, TE: 108 ms, ETL: 19) (B) on thesame axial scan (white arrows). After IVadministration of Gd-DOTA, no

areas of contrast enhancement are detected on TSE T1-weighted fat-saturated axial scan (TR: 487 ms, TE: 11 ms, ETL: 5, TI: 150 ms) (C)(white arrows). A subsequent skin biopsy and histological specimenstained with H&E (D) shows an amorphous material with collagen fibres,with no inflammatory infiltration

Eur Radiol

In conclusion, regardless of its high cost, contrast-enhanced MRI should be recommended as a diagnostic toolfor patients with severe late side effects after injection oftemporary or permanent facial dermal fillers, particularlywhen an inflammatory granulomatous reaction is suspected.

Acknowledgments The authors would like to thankDr. Eugenio Pucci,MD, for evaluation of histological specimens; Dr. Ezio Nicodemi, MD,who referred some patients and performed some biopsy procedures; Dr.Chiara Briani for statistical analysis; and Prof. James Marsh Sternberg,MD, for valuable suggestions for this manuscript.

The scientific guarantor of this publication is Marco Di Girolamo, thelead author of this manuscript. The authors of this manuscript declare norelationships with any companies whose products or services may berelated to the subject matter of the article. The authors state that this workhas received no funding. Chiara Briani, MD, kindly provided statisticaladvice for this manuscript. Institutional Review Board approval wasobtained. Written informed consent was obtained from all subjects(patients) in this study. No study subjects or cohorts have been previouslyreported. Methodology: retrospective diagnostic or prognostic study,performed at one institution.

References

1. Eppley BL, Dadvand B (2006) Injectable soft tissue fillers: clinicaloverview. Plast Reconstr Surg 118:98–110

2. Broder KW, Cohen SR (2006) An overview of permanent andsemipermanent fillers. Plast Reconstr Surg 118:7–14

3. Smith KC (2008) Reversible vs. Nonreversible fillers in facial aes-thetics: concerns and considerations. Dermatol Online J 14:3

4. Sherman RN (2006) Sculptra: the new three dimensional filler. ClinPlast Surg 33:539–550

5. Shumaker PR, Sakas EL, Swann MH, Greenway HT Jr (2009)Calcium hydroxylapatite tissue filler discovered 6 years after implan-tation into the nasolabial fold: case report and review. Dermatol Surg35:375–379

6. Lemperle G, Romano JJ, Busso M (2003) Soft tissue augmentationwith Artecoll: 10-year history, indications, techniques and complica-tions. Dermatol Surg 29:573–587

7. Bello G, Jackson IT, Keskin M, Kelly C, Dajani K, StudingerR et al (2007) The use of polyacrylamide gel in soft tissueaugmentation: an experimental assessment. Plast ReconstrSurg 119:1326–1336

8. Lahiri A, Waters R (2007) Experience with Bio-Alcamid, a new softtissue endoprosthesis. J Plast Reconstr Aesthet Surg 60:663–667

Fig. 5 Description of theproposed diagnostic MRIpathway in patients with facialdermal filler in normal andcomplicated cases

Eur Radiol

9. Bergeret-Galley C, Latouche X, Illouz YG (2001) The value of a newfiller material in corrective and cosmetic surgery: dermalive andDermadeep. Aesthet Plast Surg 25:249–255

10. Ersek RA, Beisang AA (1992) Bioplastique: a new biphasic polymerfor minimally invasive injection implantation. Aesthet Plast Surg 16:59–65

11. Lemperle G, Morhenn V, Charrier U (2003) Human histology andpersistence of various injectable filler substances for soft tissue aug-mentation. Aesthet Plas Surg 27:354–366

12. Narins RS, Beer K (2006) Liquid Injectable Silicone: a review of itshistory, immunology, technical considerations, complications andpotential. Plast Reconstr Surg 118:77S–84S

13. Scaranelo AM, de Fátima Ribeiro Maia M (2006) Sonographic andmammographic findings of breast liquid silicone injection. J ClinUltrasound 34:273–277

14. Prather CL, Jones DH (2013) Liquid injectable silicone for soft tissueaugmentation. J Am Acad Dermatol 19:159–168

15. Chen F, Carruthers A, Humphrey S, Carruthers J (2013) HIV-associated facial lipoatrophy treated with injectable silicone oil: apilot study. Dermatol Surg 69:431–437

16. Gladstone HB, Cohen JL (2007) Adverse effects when injectingfacial fillers. Semin Cutan Med Surg 26:34–39

17. Requena C, Izquierdo MJ, Navarro M, Martínez A, Vilata JJ, BotellaR et al (2001) Adverse reactions to injectable aestheticmicroimplants. Am J Dermatopathol 23:197–220

18. Reisberger EM, Landthaler M, Wiest L, Schröder J, Stolz W (2003)Foreign body granulomas caused by polymetylmethacrilate micro-spheres. Arch Dermatol 139:17–20

19. Alcalay J, Alkalay R, Gat A, Yorav S (2003) Late-onset granuloma-tous reaction to Artecoll. Dermatol Surg 29:859–862

20. Funt D, Pavicic T (2013) Dermal fillers in aesthetics: an overview ofadverse events and treatment approaches. Clin Cosmet InvestigDermatol 6:295–316

21. Hanke CW, Rohrich RJ, BussoM, Carruthers A, Carruthers J, FagienS et al (2011) Facial soft tissue fillers conference: assessing the stateof the science. JAAD 64:S66–S85

22. Lemperle G, Gauthier-Hazan N, Lemperle M (1998) PMMA-Microspheres (Artecoll) for long-lasting correction of wrinkles: re-finements and statistical results. Aesthet Plast Surg 22:356–365

23. Bailey SH, Cohen JL, Kenkel JM (2011) Etiology, prevention, andtreatment of dermal filler complications. Aesthet Surg J 31:110–121

24. Grippaudo FR, Mattei M (2011) The utility of high-frequency ultra-sound in dermal fillers evaluation. Ann Plast Surg 67:469–473

25. Grippaudo FR, Mattei M (2010) High-frequency sonography oftemporary and permanent dermal fillers. Skin Res Technol 16:265–269

26. Wortsman X, Wortsman J (2011) Sonographic outcomes of cosmeticprocedures. AJR 197:W910–W918

27. Schelke LW, Van Den Elzen HJ, Erkamp PPM, Neumann HAM(2010) Use of ultrasound to provide overall information on facialfillers and surrounding tissue. Dermatol Surg 36:1843–1851

28. Pooley RA (2005) AAPM/RSNA physics tutorial for residents.Fundamental physics of MR imaging. RadioGraphics 25:1087–1099

29. Paajaneen H, Brasch RC, Schmiedlu Ogan M (1987) Magneticresonance imaging of local soft tissue inflammation using gadolini-um-DTPA. Acta Radiol 28:79–83

30. Grippaudo FR, Pacilio M, Di Girolamo M, Dierckx RA, Signore A(2013) Radiolabelled white blood cell scintigraphy in the work-up ofdermal filler complications. Eur J Nucl Med Mol Imaging 40:418–425

31. Som PM, Curtin HD, Mancuso AA (1999) An imaging-based clas-sification for the cervical nodes designed as an adjunct to recentclinically based nodal classifications. Arch Otolaryngol Head NeckSurg 125:388–396

32. Steinkamp HJ, Hosten N, Richter C, Schedel H, Felix R (1994)Enlarged cervical lymph nodes at helical CT. Radiology 191:795–798

33. Ginat DT, Schatz CJ (2013) Imaging features of midface injectablefillers and associated complications. Am J Neuroradiol 34:1488–1495

34. Bitar R, Leung G, Perng R, Tadros S, Moody AR, Sarrazin J et al(2006) MR pulse sequences: what every radiologist wants to knowbut it is afraid to ask. RadioGraphics 26:513–537

35. Berg WA, Ngujen TK, Middleton MS, Soo MS, Pennello G, BrownSL (2002) MR imaging of extracapsular silicone from breast im-plants: diagnostic pitfalls. AJR 178:465–472

36. Gensanne D, Josse G, Schmitt AM, Lagarde M, Vincensini D (2007)In vivo visualization of hyaluronic acid injection by high spatialresolution T2 parametric magnetic resonance imaging. Skin ResTechnol 13:385–389

37. Josse G, Haftek M, Gensanne D, Turlier V, Mas A, Lagarde JM et al(2010) Follow up study of dermal hyaluronic acid injection by highfrequency ultrasound and magnetic resonance imaging. J DermatolSci 57:214–216

38. Kransdorf MJ, Murphey MD, Temple HT (1998) Subcutaneousgranuloma annulare: radiologic appearance. Skelet Radiol 27:266–270

39. Feeney JN, Fox JJ, Ackhurst T (2009) Radiological impact of the useof calcium hydroxylapatite dermal fillers. Clin Radiol 64:897–902

40. Park TH, Comment on Grippaudo et al (2013) Radiolabelled whiteblood cell scintigraphy in the work-up of dermal filler complications.Eur J Nucl Med Mol Imaging 40:418–425

41. Gray H (1918) Anatomy of the human body. Lea & Febiger,Philadelphia. Bartleby.com, 2000. www.bartleby.com/107/

42. Ghirardelli ML, Jemos V, Gobbi PG (1999) Diagnostic approach tolymph node enlargement. Haematologica 84:242–247

Eur Radiol