at SciVerse ScienceDirect

Clinical Radiology xxx (2013) e1ee7

Contents lists available

Clinical Radiology

journal homepage: www.cl inicalradiologyonl ine.net

Adiposis dolorosa (Dercum’s disease): MRI and ultrasoundappearancesB.J. Tins a,*, C. Matthews b, M. Haddaway a, V.N. Cassar-Pullicino a, R. Lalam a, J. Singh a,P.N.M. Tyrrell aaDepartment of Radiology, Robert Jones and Agnes Hunt Orthopaedic Hospital, NHS Trust, Oswestry, UKbDepartment of Metabolic Medicine, Robert Jones and Agnes Hunt Orthopaedic Hospital, NHS Trust, Oswestry, UK

article information

Article history:Received 2 April 2013Received in revised form3 May 2013Accepted 15 May 2013

* Guarantor and correspondent: B.J. Tins, DRobert Jones and Agnes Hunt Orthopaedic HospLane, Oswestry SY10 7AG, UK. Tel.: þ44 1691 4041

E-mail address: Bernhard.Tins@rjah.nhs.uk (B

0009-9260/$ e see front matter � 2013 The Royal Cohttp://dx.doi.org/10.1016/j.crad.2013.05.004

Please cite this article in press as: Tins BJ, e(2013), http://dx.doi.org/10.1016/j.crad.2013

AIM: To describe ultrasound and magnetic resonance imaging (MRI) features of adiposisdolorosa, Dercum’s disease, and to evaluate the MRI features prospectively against a largenumber of MRI examinations.MATERIALS AND METHODS: Institutional review board approval for this study was obtained.

The imaging features at MRI and ultrasound of 13 cases of adiposis dolorosa (nine female, fourmale; age range 32e72 years) were reviewed. MRI findings typical for adiposis dolorosa wereproposed and prospectively evaluated on 6247 MRI examinations performed over a period of 8months.RESULTS: Adiposis dolorosa demonstrates multiple, oblong, fatty lesions in the superficial

subcutaneous fatty tissue. They are mostly <2 cm in long axis diameter. They demonstratenodular (“blush-like”) increased fluid signal at unenhanced MRI and are markedly hyperechoicat ultrasound. There is no contrast medium enhancement at MRI and no increased Dopplersignal at ultrasound. Most lesions were clinically asymptomatic, some were painful/tender.There was no imaging evidence of oedema or inflammation. During prospective validation ofthese MRI features on 6247 MRI examinations, two cases with typical imaging features wereencountered; both were diagnosed as adiposis dolorosa on clinical review. All cases of adiposisdolorosa showed these imaging findings. This results in a very low likelihood that a nodular,blush-like appearance of subcutaneous fat on MRI is not due to adiposis dolorosa.DISCUSSION: Adiposis dolorosa, Dercum’s disease, should be suggested in the presence of

multiple (many) small, oblong, fatty lesions in the subcutaneous fatty tissue in adult patients ifthey are hyperechoic on ultrasound imaging or blush-like at unenhanced MRI; typically a smallnumber of these lesions are tender/painful. Imaging does not demonstrate inflammation oroedema in relation to these lesions. These MRI features should suggest the diagnosis and arelikely to be pathognomonic. The radiologist is often the first to suggest the diagnosis based onthe imaging features.

� 2013 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.

epartments of Radiology,ital, NHS Trust, Twmpath89; fax: þ44 1691 404057..J. Tins).

llege of Radiologists. Published by

t al., Adiposis dolorosa (Der.05.004

Introduction

Adiposis dolorosa, Dercum’s disease, is a complex,incompletely understood disorder of subcutaneous fat. TheWorld Health Organization has recognized it as a distinctdisease entity and classes it as lipomatosis not elsewhereclassified (ICD-10, IV, E88.2). Adiposis dolorosa is also

Elsevier Ltd. All rights reserved.

cum’s disease): MRI and ultrasound appearances, Clinical Radiology

B.J. Tins et al. / Clinical Radiology xxx (2013) e1ee7e2

referred to as Dercum’s syndrome, reflecting the lack ofclear definition and aetiology.

Patients typically present with painful, subcutaneouslumps; further investigations show these to be fatty lesions.The pain can be severe and there is often associated obesity,generalized aches and fatigability, and a wide range ofmental disturbance. There is symptomatic overlap to thefibromyalgia syndrome, although the aetiology is thought tobe different.1e3 It is unclear whether psychological prob-lems are primary symptoms of this disorder or secondary tochronic pain and obesity.3e5

It affects women more frequently than men and oftenoccurs perimenopausally. There is a definite hereditaryelement, which is presumed to be autosomal dominantwith incomplete penetrance, although often the presenta-tion is thought to be due to a new mutation, and therefore,there is no relevant family history.1e4,6e9

It is unclear whether the painful lesions in adiposisdolorosa have defining histological features. Some studiesfind the lesions histologically identical to “normal” li-pomas.8 Other studies suggest inflammatory changes andangiolipoma-like change.4 There is evidence to suggest thatalthough there is mild inflammation, this is no more pro-nounced in obese controls and, therefore, not specific.10

Table 1Cases of clinically confirmed adiposis dolorosa listing the areas investigated with

Age (years)at presentation

Examinationtype

Examination area Number oflesions

Maxsize

37 MRI Lower calf NoneMRI Upper arm Few 8US Neck, arm, abdomen/

chest wall, legMany 12

48 MRI Lspine retro Few 8MRI Pelvis retro One 13MRI Upper back One 20US One 20

38 US Thighs, arm Few ineach area

18

65 MRI Shoulder 18

32 US Abdomen/chest wall Many 2542 MRI Whole spine Several 11

MRI Thighs Many 8US Thighs Many 30US Thighs, arms Many

66 US Shoulder, thighs Many 1172 US Back Several 20

MRI Spine Several 1172 US Groin, flank Many 17

MRI Abdomen, pelvis Few 462 MRI Hip Many 1747 US Chest, abdomen wall,

back, arms, legsMany 20

32 US Thigh, arm Many 20MRI Back, arm, thigh, pelvis Many 40

36 MRI Lower back Two 12US Lower back Several 10

In the number of lesions section few ¼ 2e5, several ¼ 5e15.FS (fat saturation) stands for proton density (PD) or T2-weighted FS or short tau inthe PD FS or T2 FS sequence this was indicated. The lesion with histological resuMRI, magnetic resonance imaging; ultrasound ultrasound; Hyper, hyperechogen

Please cite this article in press as: Tins BJ, et al., Adiposis dolorosa (Der(2013), http://dx.doi.org/10.1016/j.crad.2013.05.004

The painful lesions have been reported anywhere in thesubcutaneous fat, although there is said to be a prepon-derance for the lower extremities and the lower area of thetrunk of the body. Nevertheless, presentation even in theskull vault has been described.

The literature on adiposis dolorosa is, in parts, contra-dictory, and this disorder plays little to no role in the im-aging literature. However, as the authors’ awareness of thisdisorder grew, it became clear that it is often the radiologistwho first raises the possibility of this disorder, in particularwhen assessing patients with painful soft-tissue lumps. Theauthors have collected cases of proven Dercum’s disease toreview the imaging features of this disorder. In a secondstep, these imaging features were prospectively evaluatedon all magnetic resonance imaging (MRI) examinationsperformed at the authors’ institution to assess whetherthese findings are specific for Dercum’s disease.

Materials and methods

For the first part of the study the authors reviewed im-aging features of known cases of Dercum’s disease. Thesehad been collected over a period of 3 years, from May 2008

imaging.

imum(mm)

Blush-likeappearanceon MRI

US signalechogenicity

Best MRIsequence

Histology

NA Iso, hyperYes T1, FS

FS

Yes T1Yes PD FSYes PD FS, T1

HyperHyper

Yes fortypicallesions

Atypical large lesion,spindle cell lipoma

HyperYes StirYes FS

HyperHyperHyperHyper

Yes StirHyper

Yes FSYes FS

Hyper

HyperYes FSYes Stir Angiolipoma, spindle

cell lipomaHyper

version recovery (STIR) sequence. Where the STIR sequence was better thanlt was larger and more atypical in appearance than the more typical lesions.icity.

cum’s disease): MRI and ultrasound appearances, Clinical Radiology

B.J. Tins et al. / Clinical Radiology xxx (2013) e1ee7 e3

to June 2011. The diagnosis was based on assessment by aphysician with experience in diagnosis and treatment ofthis disorder. It is based partly on typical clinical features, inparticular, the presence of multiple, painful, fatty soft-tissuelesions and partly on exclusion of other disease entities. Thepresence of multiple but pain-free lipomas was not regar-ded as presentation of adiposis dolorosa.

The imaging features, and if present histological features,of the cases of Dercum’s disease were reviewed. Ultrasoundexaminations were performed on a Toshiba Aplio with a12 MHz compound transducer. MRI was performed on aSiemens Avanto MRI system using surface coils. Routinely,T1-weighted sequences were performed in two planes, andshort-tau inversion recovery (STIR) and proton-density fat-saturated (PD-FS) sequences in one plane each.

All imaging cases were initially collected prospectively,and for the study analysed retrospectively by the firstauthor. The examinations were first reported by thefollowing authors B.T., V.C.P., R.L., J.S., P.T., who are all sub-specialty trained musculoskeletal radiologists.

For the prospective part of the study, the authors set outto test the hypothesis that the presence of multiple, largelyround or oblong areas of increased water signal in thesubcutaneous fat is diagnostic of Dercum’s disease by pro-spectively evaluating all MRI images performed at theirinstitution for the presence of this feature. All MRIs per-formed in the authors’ institution are reviewed by aconsultant radiologist. Before the beginning of the study, allconsultant radiologists were briefed about the MRI featuresof notable lesions.

The presence of multiple areas of largely round areas ofincreased water signal in the subcutaneous fat was to benoted. These resemble an ill-defined nodular lesion onunenhanced imaging. The signal is increased in water-sensitive sequences, such as PD-FS, T2-weighted FS, orSTIR images, and the signal is decreased in T1-weightedsequences.

The observations were structured in two distinct parts. Inpart 1 of the prospective part of the study, all MRIs per-formed from 1 August 2011 to 31 March 2012 performed forwhatever indication and of whatever area were reviewedfor the presence of the MRI features described above. Incases where the imaging features were suggestive of adi-posis dolorosa, clinical review by a clinician experienced inthis disorder was undertaken.

In part 2 of the prospective part of the study, all caseswith an already known diagnosis of adiposis dolorosa pre-senting in the study period were to be reviewed for thepresence of the previously described imaging features.

Institutional review board approval was obtained for thisstudy.

Figure 1 Ultrasound images of a 37-year-old woman with adiposisdolorosa. Power Doppler (a) and grey-scale images (b) show thetypical sonographic features: superficial location, hyperechogenicity,absence of increased vascularity or oedema, and relatively small size.

Results

The results for part 1 of the study, the imaging featuresseen in known cases of Dercum’s disease are as follows: 13patients with Dercum’s disease were identified in the studyperiod. The findings are summarized in Table 1. Nine

Please cite this article in press as: Tins BJ, et al., Adiposis dolorosa (Der(2013), http://dx.doi.org/10.1016/j.crad.2013.05.004

patients were female, four male. The age range was 32e72years. The older patients (older than 40 years) had symp-toms for many years, whereas the younger patients pre-sented with newly symptomatic painful lumps. In 11 casesthe diagnosis was first suggested by a radiologist. In allthese cases, the patient presented with a painful soft-tissuelesion. In two cases, the diagnosis of adiposis dolorosa wasknown already, and the patients were sent for the assess-ment of painful lesions.

Nine patients underwent MRI examinations and 11 pa-tients underwent ultrasound examinations. Two patientsunderwent only MRI imaging and four patients underwentonly ultrasound. The other seven patients underwent MRIand ultrasound.

In all patients the lesions were centred in the superficialsubcutaneous fat and not in the deep subcutaneous fat(Figs 1e4). The lesions were far more numerable than couldbe appreciated clinically, most of the lesions visible at im-aging were clinically asymptomatic. Most lesions measured<2 cm in diameter and were oblong with the long axisparallel to the skin. All lesions were strongly hyperechoic atultrasound and did not demonstrate flow on power orcolour Doppler imaging (Fig 1). The lesions demonstrated ablush-like appearance on unenhanced MRI with decreasedT1-weighted signal and increased signal in water-sensitivesequences such as STIR, T2-FS, or PD-FS-weighted imaging

cum’s disease): MRI and ultrasound appearances, Clinical Radiology

Figure 2 MRI images of the left thigh of a 42-year-old woman with adiposis dolorosa. Axial PD-FS (a, left) and T1-weighted (a, right) imagesshow faint blush-like signal changes (arrows) of the superficial subcutaneous fat in the superficial subcutaneous tissue, better appreciated on thePD-FS images. Coronal STIR (b, left) and T1-weighted images (b, right) again show scattered lesions (arrows) in the subcutaneous fat, the lesionsare better appreciated on the water-sensitive sequences.

B.J. Tins et al. / Clinical Radiology xxx (2013) e1ee7e4

(Fig 2). Larger lesions were more inhomogeneous inappearance at MRI (Fig 3), and generally, symptomatic le-sions were larger than the majority of lesions seen. Largerand heterogeneous lesions were only seen in patients withvery extensive disease, most lesions were small and moretypical in appearance (Fig 3). There were no other imagingdifferences between symptomatic and asymptomatic le-sions on ultrasound or MRI.

There was no oedema in or around any of the lesions(symptomatic or not) seen on ultrasound. In at least onecase ultrasound demonstrated more lesions than MRI(Fig 4).

One lesion was surgically removed; the histologicaldiagnosis was that of an angiolipoma. One lesion was bio-psied; the histological diagnosis was spindle cell lipoma.

Based on this the presence of multiple lesions in thesubcutaneous fat with blush-like nodular increased fluidsignal at MRI was defined as suggestive of Dercum’s disease.For the prospective part of the study, all MRI examinationsperformed at the authors’ institution were assessed forthese features over a period of 8months. During this period,

Please cite this article in press as: Tins BJ, et al., Adiposis dolorosa (Der(2013), http://dx.doi.org/10.1016/j.crad.2013.05.004

6247 MRI examinations were assessed; most examinationswere performed for musculoskeletal indications. Two casesof imaging features suggestive of Dercum’s disease wereencountered. In both cases the diagnosis of adiposis dolo-rosa was first suggested by the radiologist.

Patient 1 was a 31-year-old man with multiple painfullumps in the subcutaneous fat of his arms, trunk, and lowerlimbs, which had typical MRI features. There was a familyhistory of similar lesions in the patient’s mother and threebrothers.

Patient 2 was a 74-year-old woman without a familyhistory, but with multiple painful subcutaneous lesions.MRI demonstrated scattered, small, ill-defined, nodular orblush-like lesions in the subcutaneous fat of the frontal anddorsal abdominal wall. Both patients were classed as havingadiposis dolorosa after clinical review.

There was no other case where the ill-defined nodular/blush-like features of the subcutaneous fat were seen. Therewere no cases of new clinical diagnosis of adiposis dolorosa;therefore, there were no cases in part 2 of the prospectivepart of the study. In combination, this makes it very likely

cum’s disease): MRI and ultrasound appearances, Clinical Radiology

Figure 3 A 32-year-old man with multiple lesions of the arms, legs, and trunk. Serial axial images of the upper arm (a, top: PD-FS, bottom: T1-weighted) and the thigh (b, top: PD-FS, bottom: T1-weighted) show multiple subcutaneous lesions, the larger lesions are more inhomogeneous inappearance but still positioned in the superficial subcutaneous fat. A chest wall lesion is noted on the axial images of the arm (a, arrow, top leftimage). Larger lesions are more likely to be clinically symptomatic. Sagittal STIR images help to appreciate the multitude of lesions (c, arm), (d,thigh, STIR and T1-weighted imaging). Very extensive disease, such as in this case, is less common than the more scattered lesions as seen in Fig 2.

B.J. Tins et al. / Clinical Radiology xxx (2013) e1ee7 e5

that the imaging findings discussed were specific for adi-posis dolorosa.

Discussion

Adiposis dolorosa, Dercum’s disease, is probably under-appreciated by most doctors, including radiologists.

Please cite this article in press as: Tins BJ, et al., Adiposis dolorosa (Der(2013), http://dx.doi.org/10.1016/j.crad.2013.05.004

Although rare (the incidence is unclear), this disorder shouldbe considered in the differential diagnosis when painful ortender subcutaneous fatty lumps are encountered. Dercum’sdisease presents with multiple lesions, most of which areasymptomatic. The reason for the painful nature of the lipo-matous lesions is not clear. Pressure on adjacent nerves,oedema, inflammation and autoimmune problems all have

cum’s disease): MRI and ultrasound appearances, Clinical Radiology

Figure 4 A 42-year-old woman with multiple lesions of the thigh. Axial MRI (a, top: PD-FS, bottom: T1-weighted) show typical lesions.Extended-view ultrasound (b) demonstrates a much higher lesion number than seen at MRI.

B.J. Tins et al. / Clinical Radiology xxx (2013) e1ee7e6

been suggested thoughnonehas been conclusively proven. Itis unclear whether the painful lesions in adiposis dolorosahave defining histological features. Some studies found thatthe lesionswere indistinguishable from “normal” lipomas.8,10

Other studies suggest inflammatory changes andangiolipoma-like change,10 and it has also been suggested

Please cite this article in press as: Tins BJ, et al., Adiposis dolorosa (Der(2013), http://dx.doi.org/10.1016/j.crad.2013.05.004

that there are subtypes of adiposis dolorosa with differenthistopathological features.4 In the present study, there weretwo cases with histology demonstrating angiolipoma in oneand spindle cell lipoma in the other. A proliferation of vesselsand distortion of septa could explain the imaging features atMRI and ultrasound.

cum’s disease): MRI and ultrasound appearances, Clinical Radiology

B.J. Tins et al. / Clinical Radiology xxx (2013) e1ee7 e7

All cases of Dercum’s disease showed very similar im-aging features. The lesions were immediately subcutaneous,oblong rather than round, with the long axis parallel to theskin surface. Most lesions were <2 cm in maximum diam-eter. At ultrasound they were hyperechoic and did not showincreased flow on Doppler imaging. At MRI they showednodular increased fluid signal, better appreciated on water-sensitive sequences such as STIR, T2-FS or PD-FS-weightedimaging rather than on T1-weighted sequences. Larger le-sions were more inhomogeneous at MRI. The appearancesresembled a vascular “blush” on unenhanced images. Nocontrast medium enhancement was seen.

The sonographic findings in adiposis dolorosa differ fromthose of simple lipomas, which are often larger and not ashyperechoic, and although they can be multiple, they are notas numerous as seen in typical cases of Dercum’s disease.Simple lipomas are also not restricted to the superficial fatlayer but occur anywhere in the subcutaneous fat. Simple li-pomasper searenot tenderorpainful, although theycancauselocal discomfort. The sonographic appearance of the lesions inDercum’s disease have been recently described in histologi-cally proven cases of angiolipomas.11 In the study of Banget al.11 most lesions were isolated. Simple lipomas demon-strate only fat signal at MRI, unlike the lesions in adiposisdolorosa disease. The MRI appearance is consistent with thatof angiolipomas. Contrastmediumenhancement is not part ofthe routine imaging protocol for soft-tissue lesions in the au-thors’ institution.However, contrastmediumwas given inonecase, but the lesions did not show significant enhancement.

Only a small number of the visible lesions are clinicallysymptomatic. In more severe cases, patients may havemorethan 100 lesions, most are clinically asymptomatic. Symp-tomatic lesions were larger than the majority of lesions onultrasound, and larger lesions were more heterogeneous onMRI. There was no oedema in or around any of the lesions(symptomatic or not) seen on ultrasound.

This is contrary to the hypothesis that the lesions areassociated with or the pain is due to inflammation andoedema; in the cases examined not a single lesion or per-ilesional tissue demonstrated oedema or increased vascu-larity or other evidence of inflammation. This is supportedby a recent histological study, which also found no undueinflammatory change.10

Publications showing imaging examples of Dercum’sdisease are very sparse and the features seen in a reportedcase report6 are in keeping with simple lipomas and arefundamentally different to the features seen in all cases ofadiposis dolorosa presented here.

Please cite this article in press as: Tins BJ, et al., Adiposis dolorosa (Der(2013), http://dx.doi.org/10.1016/j.crad.2013.05.004

In the prospective part of the study, two cases of Der-cum’s disease were suggested based on the imaging find-ings; the diagnosis was confirmed on clinical review. Nocase of Dercum’s disease not showing these imaging fea-tures was encountered. This indicates that the combinationof multiple subcutaneous fatty lesions, hyperechoic on ul-trasound, blush-like, nodular, increased fluid signal onunenhanced MRI, with most lesions being asymptomaticand some being tender, is very suggestive and possiblypathognomonic for adiposis dolorosa, Dercum’s disease.

In the published literature and the authors’ experience,including the cases presented here, the lesions occur mainlyin the superficial subcutaneous tissue of the trunk, theupper arms, and the thighs.

In summary, adiposis dolorosa, Dercum’s disease, pre-sents with multiple fatty subcutaneous lumps, only some ofwhich are painful or tender. On imaging there are manysmall, oblong lesions in the superficial subcutaneous fattytissue of trunk, arms, and legs. They are hyperechoic onultrasound and ill-defined nodular/blush-like lesions onunenhanced MRI. These lesions are not associated with anyinflammatory change on imaging. The radiologist is oftenthe first to suggest the diagnosis.

References

1. Herbst KL, Rutledge T. Pilot study: rapidly cycling hypobaric pressureimproves pain after 5 days in adiposis dolorosa. J Pain Res 2010;3:147e53.

2. Yousefi M. Adiposis Dolorosa. In: Elston D, editor. Medscape. New York:WebMD; 2010.

3. Hansson E, Svensson H, Brorson H. Review of Dercum’s disease andproposal of diagnostic criteria, diagnostic methods, classification andmanagement. Orphanet J Rare Dis 2012;7:23.

4. Lange U, Oelzner P, Uhlemann C. Dercum’s disease (lipomatosis dolo-rosa): successful therapy with pregabalin and manual lymphaticdrainage and a current overview. Rheumatol Int 2008;29:17e22.

5. Hansson E, Svensson H, Brorson H. Depression in Dercum’s disease andin obesity: a caseecontrol study. BMC Psychiatry 2012;12:74.

6. Amine B, Leguilchard F, Benhamou CL. Dercum’s disease (adiposisdolorosa): a new case report. Joint Bone Spine 2004;71:147e9.

7. Brodovsky S, Westreich M, Leibowitz A, Schwartz Y. Adiposis dolorosa(Dercum’s disease): 10-year follow-up. Ann Plast Surg 1994;33:664e8.

8. Campen R, Mankin H, Louis DN, Hirano M, Maccollin M. Familialoccurrence of adiposis dolorosa. J Am Acad Dermatol 2001;44:132e6.

9. Lynch HT, Harlan WL. Hereditary factors in adiposis dolorosa (Dercum’sDisease). Am J Hum Genet 1963;15:184e90.

10. Hansson E, Svensson H, Stenram U, Brorson H. Histology of adiposetissue inflammation in Dercum’s disease, obesity and normal weightcontrols: a case control study. J Inflamm (Lond) 2011;8:24.

11. Bang M, Kang BS, Hwang JC, et al. Ultrasonographic analysis of subcu-taneous angiolipoma. Skeletal Radiol 2012;41:1055e9.

cum’s disease): MRI and ultrasound appearances, Clinical Radiology