Between February 1999 and October 2002, three children presented to the
orthopaedic clinic at our institution for evaluation of thigh dimples; each
was found to have a lipoblastoma. The families of our three patients were
informed that data concerning the cases would be submitted for
publication.
Case 1. A thirty-nine-month-old girl presented with a three to
four-month history of an enlarging dimple on the proximal part of the lateral
aspect of the left thigh. The defect was noticed when it was approximately 7
mm in size, and it had grown to 25 mm in size by the time of presentation to
the orthopaedic clinic. No history of trauma or pain was reported. Gestation,
birth, development, and the medical history were unremarkable. Physical
examination revealed a 2.5-cm, nontender deficit of the subcutaneous fat
overlying the left greater trochanter. No swelling or mass was visible or
palpable. No other skin or musculoskeletal abnormality was found. The
diagnosis of fat atrophy was considered.
Plain radiographs demonstrated no calcifications in the soft tissues and no
osseous abnormalities. Magnetic resonance imaging scans were indeterminate,
showing a 1 cm × 2 cm × 1.5 cm area in the subcutaneous tissues
that was hypointense on T1-weighted images and hyperintense on T2-weighted
images, with no changes in the fascia lata or underlying muscles (Figs.
1-A and
1-B). It was specifically noted
that the signal characteristics of the lesion were not characteristic of
mature scar tissue.
The parents reported continued enlargement of the dimple, and excisional
biopsy was planned. Six weeks after the initial presentation, the child
underwent an excisional biopsy that included the underlying fascia as well as
a cuff of normal subcutaneous tissue. Histological analysis showed a
diffuse-type lipoblastoma with irregular lobules of mature and immature fat
cells in a loose myxoid stroma, separated by connective-tissue septa
(Fig. 1-C). Scar tissue was
excluded as a diagnosis because of the presence of intact epidermis and dermal
adnexal structures. Fat atrophy was excluded because of the complete absence
of histiocytic and inflammatory infiltrates. No hyperchromatism, nuclear
atypia, or pleomorphisms were noted. The margins were free of tumor.
One week after surgery, the wound spontaneously dehisced from what appeared
to have been a small seroma. The wound was dressed and was allowed to close by
secondary intention. It was completely healed by the six-week postoperative
evaluation. Thirty-two months after surgery, the patient was doing well
without palpable or visible recurrence. Near-normal adipose tissue had
reconstituted the field of resection.
Case 2. The parents of a twenty-one-month-old girl noticed a
dimple on the anterolateral aspect of the child's left thigh two weeks before
presenting to the orthopaedic clinic. The parents stated that the child did
not have a limp, did not seem to be in pain, and had exhibited no changes in
behavior. The medical history included a breech presentation, a cesarean
delivery, and tympanoplasties for recurrent otitis media. The physical
examination demonstrated a 4-cm-diameter dimple on the anterior aspect of the
thigh that was nontender and was nonadherent to the underlying tissues. A
slight bluish hue was noted. Lipoblastoma, fibrosis, and hemangioma were all
considered as possible diagnoses.
Magnetic resonance imaging showed a plate-like nondiscrete area of
increased T2 signal, reduced T1 signal (relative to fat), and gadolinium
enhancement that did not involve the underlying muscle (Figs.
2-A,
2-B, and
2-C.) An absence of
subcutaneous fat in the area was also noted. No diagnosis was offered.
Two weeks after presentation, the patient underwent an excisional biopsy of
what was found to be a diffuse-type lipoblastoma. Histological analysis of the
biopsy sample revealed white-tan lobules of fatty tissue that contained
lipoblasts and more mature adipocytes in a loose stroma of spindle-shaped
cells (Figs. 2-D and
2-E). Scar tissue and fat
atrophy were excluded as possible diagnoses because of the presence of intact
dermal adnexal structures and the lack of histiocytic and inflammatory
infiltrates. There was focal involvement of the deep margins.
At the 4.5-month postoperative evaluation, the patient was doing well, the
incision had healed, and there was no evidence of recurrence.
Case 3. A thirty-one-month-old girl presented with a six-week
history of an enlarging "hole" in the anteromedial aspect of the
left thigh. The parents had noted asymmetric limb use but no specific limp.
The depression in the thigh had grown from 1 to 2 cm within four weeks, had a
slight gray-purple discoloration, and was not causing any pain. No
constitutional symptoms were reported. The patient had already been examined
by a pediatrician, a rheumatologist, and a dermatologist, the last two of whom
had tentatively suggested the diagnosis of lipoatrophy and fat necrosis. The
medical history revealed only neonatal jaundice and recent tympanoplasties.
There was no history of trauma or vaccination to the child's left thigh.
Physical examination demonstrated no limp or limb-favoring, no firm mass, no
adherence of tissue layers, no tenderness, and no skin changes other than a
slight gray-purple hue overlying the dimple, which measured 2.5 cm in diameter
(Fig. 3-A).
Magnetic resonance imaging demonstrated a region of reduced T1 signal,
increased T2 signal, and gadolinium enhancement involving the subcutaneous
tissue, with apparent volume reduction of the subcutaneous tissue but no
infiltration of the underlying musculature (Figs.
3-B,
3-C, and
3-D). The diagnosis of
panniculitis was suggested.
One month after presentation, the patient was taken to the operating room
for an excisional biopsy because of continued slow expansion of the defect.
The diagnosis of lipoblastoma, suspected because of our previous two cases,
was confirmed histopathologically (Figs.
3-E and
3-F). Histopathologic
examination revealed immature adipose tissue in a lobular pattern, with
lipoblasts of varying maturity and interspersed myxoid matrix. There was no
evidence of fat necrosis, fat atrophy, or scar formation. The deep margins
were positive for lobules of lipoblasts.
At the one-month follow-up evaluation, the incision was noted to be well
healed. The parents were satisfied with the results of the treatment.
Additional Cases. With use of the SNOMED (Systemized
Nomenclature of Human and Veterinary Medicine) diagnostic retrieval system
(SNOMED International, a division of the College of American Pathologists,
Northfield, Illinois), we searched the pathology records at our institution
and identified nine cases of lipoblastoma/lipoblastomatosis that occurred
within the period from 1976 to 2002, including the three cases already
described (Table I). Of the six
patients not already described, one presented with a perianal lipoblastoma
associated with a variant of imperforate anus and the other five presented
with a painless, nontender lump. One patient had an ipsilateral mild limp,
another had a giant nevus flammeus overlying the mass, and a third had a
ganglion cyst in proximity to the lesion.
Addendum. Since the preparation of the original manuscript, the
senior author (F.R.D.) has seen two additional cases of skin dimpling in
toddlers. The first patient was a twenty-seven-month-old girl who had a
3-cm-diameter dimple on the lateral portion of the middle part of the thigh.
The parents reported that the patient had no history of local trauma,
injection, pain, or constitutional symptoms. The parents elected to observe
the lesion without further evaluation. Four months later, the lesion was
unchanged or perhaps was slightly smaller. The second patient was a
seventeen-month-old-boy (the first boy with this lesion whom we have seen) who
had a 4-cm-diameter dimple approximately 3 cm proximal to the patella. The
parents reported that the dimple had developed over the preceding month. There
was no explanatory history or additional history of any abnormalities, and the
findings on physical examination were normal. The parents agreed to allow the
patient to have a magnetic resonance imaging study, which demonstrated
findings that were identical to those described in the case reports presented
above. The diagnosis of lipoblastoma was made, and observation was
recommended. The long-term outcomes for these two patients remain unknown.
We describe three cases of what is, to our knowledge, a unique presentation
of lipoblastoma as a large dimple (i.e., a skin-retracting subcutaneous-tissue
defect). A soft-tissue neoplasm that presents in this manner is exceedingly
rare. Other than the dimples that are associated with occult spinal dysraphia,
the skin retraction that often abuts protuberant dermatofibrosis masses, the
sinus tracts of dermoid and epidermoid cysts, the ulcerated dimples of
basal-cell carcinoma, and the retracting dimples that are sometimes associated
with carcinoma of the breast, few lesions of any type are associated with skin
retraction. For this reason, scar and fat atrophy from unrevealed trauma or
previous injections figured prominently in each differential diagnosis
initially.
Most lipoblastomas present in children who are younger than five years of
age, and congenital cases occur
rarely6,7.
A slight predominance of males was described in the six largest series
(male-to-female ratio,
1.5:1.0)2,5-9.
Circumscribed lipoblastoma was approximately twice as common as diffuse
lipoblastomatosis in the largest
series2,5-9,
and both types have been reported in nearly every region of the body
(Table II). When lipoblastomas
arise in the extremities or the back, they present almost exclusively as a
painless, nontender mass or swelling. Dimpling and skin retraction have not
been reported
previously10.
According to the case reports in the literature, the typical appearance of
a lipoblastoma on magnetic resonance imaging varies from isointensity with
subcutaneous fat to relative hyperintensity on T2-weighted images and
hypointensity on T1-weighted
images11-13.
Most lesions include apparent lobules, septa, or other heterogeneous material.
The magnetic resonance images of the lipoblastomas in our three patients
differed from the results of previously reported cases in that they did not
show a discrete mass. All images showed high signal intensity on T2-weighted
sequences and relatively low signal intensity on T1-weighted sequences.
Contrast imaging, which was acquired for two lesions, demonstrated gadolinium
enhancement. We repeatedly considered other entities that might also retract
the skin, such as scar tissue from unreported trauma or injection. However,
only hypercellular, desmoid-like scars could have demonstrated T2
hyperintensity and gadolinium enhancement, and such scars would appear as a
lobulated mass rather than as a skin-retracting subcutaneous deficit.
Except for their overall shape, the three neoplasms were pathologically
typical lipoblastomas. Grossly, lipoblastomas are lobulated and yellow to tan
or brown. Circumscribed lipoblastomas have thick
capsules7. Diffuse
lipoblastomatosis may have remnants of a ruptured capsule or no capsular
tissue2. None of the
lesions in our three patients had identifiable capsular tissue. Each lesion
abutted normal dermis superficially and normal superficial fascia deeply. The
overlying normal epidermis and dermal adnexal structures eliminated the
possibility that the deficit was caused by scar tissue, and the lack of any
histiocytic and inflammatory infiltrates ruled out fat atrophy. Typical
lipoblastomas are characterized histopathologically by the proliferation of
mature and immature fat cells with mesenchymal precursors and lipoblasts at
varied stages of differentiation. These cells are interspersed between fibrous
trabeculae with or without a myxoid
stroma9. Hemorrhage
and focal areas of necrosis are possible, but the lack of atypical nuclei and
hyperchromatism can help to distinguish this tumor from its malignant
mimickers, such as myxoid
liposarcoma7.
Mitotic figures, when present, should be scant and organized in a normal
manner7.
The most important differential diagnosis for lipoblastoma is myxoid
liposarcoma. The distinction between the entities is begun epidemiologically:
liposarcomas are rare in children who are younger than ten years of
age14. Grossly,
myxoid liposarcomas are not as prominently lobulated as are lipoblastomas.
Histologically, lipoblastomas lack the characteristic microcystic spaces that
are often found in
liposarcomas6.
Additionally, the rare mitotic figures of lipoblastoma are found in
mesenchymal precursor cells rather than in the lipoblasts
themselves6.
As with many neoplasms of adipose tissue, chromosomal abnormalities have
been recognized in most of the cases of lipoblastoma in which such
abnormalities have been sought. PLAG1 has been isolated and cloned as the
affected gene15. In
lipoblastoma, PLAG1 transcription is upregulated by varied genomic
rearrangements in pluripotent mesenchymal stem cells, leading to
overproliferation and delayed
differentiation16.
While it is not yet clear how the expressed product of this gene accomplishes
these changes, its involvement is consistent enough that it can be directly
contrasted with translocations typical of myxoid liposarcoma, and a secure
diagnosis can be made when the histopathological results are
equivocal17.
While the understanding of the molecular pathogenesis continues to
progress, the natural history of an undisturbed lipoblastoma remains elusive.
The authors of every large report on this neoplasm have recommended complete
surgical excision with tumor-free
margins2,6-9.
Local recurrence can be a problem in up to 25% of
cases7,8,18.
Even with multiple local recurrences, however, no confirmed lipoblastoma has
ever been reported to metastasize. In fact, some recurrences appear to have
matured into typical lipomas or fibrolipomas by the time of repeat biopsy or
excision2,6.
Because of the clear cytogenetic abnormalities in lipoblastomas, which are
distinct from those found in most lipomas, it is not thought that undiagnosed
lipoblastomas contribute largely to lipomas in adulthood. Nonetheless, a few
lipoblastomas have been followed prospectively with serial biopsies or
multiple attempted excisions that have demonstrated continual maturation of
the lipoblast cells into mature
adipocytes19-21.
One case of spontaneous resolution has been
reported22. The
authors of that report argued that in cases of difficult-to-resect,
myofascially infiltrating lipoblastomatosis, the therapeutic options include
watchful waiting.
Our three patients were treated with marginal excision prior to definitive
diagnosis. In each case, the decision to proceed with excisional biopsy was
made either because of concern regarding the progressive enlargement of the
subcutaneous retraction or because of the lack of diagnostic confidence in the
results of magnetic resonance imaging. However, diffuse infiltration of the
lipoblastoma made two of the neoplasms difficult to resect without either
leaving residual tumor or severely disfiguring the child. In each instance,
the former option was chosen. Although metastasis is not a concern, the
possibility of local recurrence remains.
In light of the fact that three patients who had this condition presented
within two and one-half years to a pediatric orthopaedic surgeon who was not a
specialist in oncology, we believe that many of these neoplasms have
previously remained undiagnosed in our clinics and elsewhere. Such thoughts
lend additional strength to the argument for watchful waiting as the safest
option.
Given these therapeutic challenges and the suggestions of a possibly
nonaggressive natural history of lipoblastoma, these diffuse-type retracting
lesions present true clinical conundrums. Ongoing research to gain a better
understanding of the natural history of lipoblastoma may alter the way in
which we view this neoplasm in the future. For the present, an increased
awareness of this condition and all of its possible presentation scenarios is
critical for an appropriate diagnosis, treatment, and follow-up.