Histopathology 1998, 32, 568–578
Correspondence
Foam cells and histiocytes in endometrial
stromal tumours
Sir : The excellent paper of McCluggage et al.1 reports
one case of low-grade endometrial stromal sarcoma
with trabecular sex cord-like areas. In these areas the
tumour cells had a rhabdoid morphology. Among the
tumour cells there were nests of foam cells, which were
considered by authors to be histiocytes. Both ultra-
structural and immunohistochemical studies concen-
trated on the cells of the stromal sarcoma, including
those of the sex cord-like areas. However neither the
immunophenotype nor the ultrastructure of the foam
cells were determined.
Recently we have studied a very similar case in a 52-
year-old woman with a 120 × 140 mm solid mass in her
left ovary. A hysterectomy, including both tubes and
ovaries, was performed. The ovarian tumour was a
fibrothecoma. In the uterus several intramural leio-
myomas were found. A submucosal, white-yellow
nodule, 25 mm in diameter was also present, which
corresponded to a low-grade stromal sarcoma with
frequent sex cord-like areas and others composed of
epithelioid cells in a trabecular pattern. Some of these
epithelioid cells showed intracytoplasmic inclusions and
rhabdoid phenotype. Contiguous to the tumour cells,
there were numerous groups of cells with large and
multivacuolated cytoplasm (Figure 1). Moreover aggre-
gates of histiocytic cells, with clear, not foamy,
cytoplasm were also present. In the immunohistochem-
ical study, stromal sarcoma cells and sex cord-like cells
were strongly positive for vimentin (Biogenex), AE1-
AE3 (Biogenex) (Figure 2) and 8,18-anti-cytokeratin
(Medac) antibodies. Focally, the cells of the stromal
sarcoma, were also positive with anti-smooth muscle
actin (Biogenex) and anti-desmin (Biogenex) antibodies,
Figure 1. Groups of foam cells close to sex cord-like trabeculae.
q 1998 Blackwell Science Limited.
Figure 2. Stromal sarcoma cells (left) and foam cells (right) positive
with anticytokeratin antibody (AE1-AE3).
but negative were the rhabdoid cells of the trabecular
areas. Likewise, the foam cells showed cytokeratins
(Figure 2) and vimentin in their cytoplasm. However,
only some isolated foamy cells were positive for smooth
muscle actin. The clear, non-vacuolated cells had an
immunophenotype according with its histiocytic mor-
phology, being stained with anti-vimentin and KP1
(Dako) antibodies. Staining for EMA (Dako), CEA (Dako)
and S100 protein (Biogenex) antigens were negative in
all these cellular types.
The presence of foam cells in endometrial stromal
tumours and sex cord-like uterine tumours has been
reported2,3. They may be also present in hyperplasias
and carcinomas of the endometrium4. On electron
microscopy abundant lipid cytoplasmic vacuoles and
membrane junctions were observed, being partially
surrounded by basal membrane and containing scanty
lysosomes3,5. Based on these findings, the foam cells
have been classified as specialized, transformed stromal
cells instead of true histiocytes. The foam cells in our
case showed immunohistochemical features similar to
other stromal tumour cells and were negative with the
KP1 antibody. So endometrial stromal sarcomas can
present tumour foam (non-histiocytic) cells, true
histiocytes or, like our case, both cellular types.
D.Suarez Vilela
F.M.Izquierdo Garcia
Servicio de Anatomı́a Patológica
Hospital de Leon
Altos de la Nava s/n
León
Spain
1. McCluggage WG, Date A, Bharucha H, Toner PG. Endometrial
stromal sarcoma with sex cord-like areas and focal rhabdoid
differentiation. Histopathology 1996; 29; 369–374.
 Correspondence 569

2. Norris HJ, Taylor HB. I. A clinical and pathological study of 53

endometrial stromal tumors. Cancer; 1976; 19; 755–766.
3. Fekete PS, Vellios F, Patterson BD. Uterine tumor resembling an
ovarian sex-cord tumor: Report of a case of an endometrial stromal
tumor with foam cells and ultrastructural evidence of epithelial
differentiation. Int. J. Gynecol. Pathol. 1985; 4; 378–387.
4. Dawagne MP, Silverberg SG. Foam cells in endometrial carcinoma.
Gynecol. Oncol. 1982; 13; 67–75.
5. Fechner RE, Bossart MI, Spjut HJ. Ultrastructure of endometrial
stromal foam cells. Am. J. Clin. Pathol. 1979; 72; 628–633.

Composite early carcinoma (ordinary

adenocarcinoma, carcinoid,
microglandular-goblet cell carcinoid,
neuroendocrine mucinous carcinoma)
of the stomach
Sir : Mixed or composite gastric carcinomas are rare
tumours, characterized by an intimate admixture of
glandular and endocrine components with frequent
histologic transitions1. We have observed a case of
composite early gastric carcinoma that showed areas of
tubular adenocarcinoma, carcinoid, microglandular-
goblet cell carcinoma and areas of mucinous adeno-
carcinoma with neuroendocrine differentiation.
The patient was a 60-year-old woman. Approxi-
mately 2 months prior to admission she developed
epigastric pain. Gastric endoscopy revealed a small
protuberant lesion with ulcer in the posterior wall of the
gastric body, and a subsequent biopsy showed adeno-
carcinoma. A partial gastrectomy with perigastric
lymph node dissection was performed. Postoperative
recovery was uneventful, and the patient has been
asymptomatic without evidence of recurrence or
metastasis for the following 36 months. The resected
stomach had a protuberant lesion with ulceration
measuring 20 × 15 mm in the posterior wall of the
body. On histological examination the tumour appeared
to occupy the mucosa and to infiltrate deeply in the
submucosa without extension to the muscularis pro-
pria. It was diagnosed as early gastric cancer. Four
distinct components of the tumour were readily
recognizable showing features of (intestinal type)
tubular adenocarcinoma, carcinoid, microglandular-
goblet cell carcinoid and mucinous adenocarcinoma. In
the deeper part of the mucosa the tumour cells were
arranged in a classic carcinoid pattern of solid nests and
trabeculae, composed of polyhedral cells with uniform
round-to-oval nuclei and finely granular eosinophilic Figure 1 a, Carcinoid tumour with solid nests and trabeculae
cytoplasm (Figure 1a). The adenocarcinomatous glands (haematoxylin and eosin). b, Histological appearance of the
concurrent adenocarcinoma and carcinoid. Chromogranin A
were more abundant in the upper half of the mucosa
immunoreactivity is intense in the middle and deep lamina propria of
(Figure 1b), and showed no intracytoplasmic mucin. the mucosal component of a carcinoid, but is absent from
The carcinoid pattern merged in many areas with adenocarcinoma (immunoperoxidase–chromogranin A with
the adenocarcinomatous pattern (Figure 1b). The haematoxylin counterstain).

q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.

570 Correspondence
microglandular-goblet cell carcinoid was mainly situ-
ated in the submucosal layer (Figure 2a). The tumour
cells were arranged in small clusters or trabecular cords
composed of between three and 12 cells. The nuclei
were generally thin and crescent shaped. They were
compressed to the rim of the cells by abundant
cytoplasm mucin. A peripheral orientation of the
nuclei predominated within the cellular clusters
(Figure 2a). The cells resembled normal goblet cells of
Figure 2. a, Trabeculae of neoplastic goblet cells within the
submucosa (haematoxylin and eosin). b, Microglandular-goblet
cell carcinoid tumour within submucosal lymphatics. Cluster of
serotonin-positive neuroendocrine cells in intimate association
with goblet cells (immunoperoxidase–serotonin with haematoxylin
counterstain). c, Mucinous adenocarcinoma in the submucosa.
Numerous chromagranin A-positive neuroendocrine cells, floating
in the mucus lakes (immunoperoxidase–chromogranin A with
haematoxylin counterstain).
the gastrointestinal tract except for nuclear
compression. Tumour cells invaded submucosal lym-
phatics (Figure 2b). Mucinous adenocarcinoma was
found in the submucosa (Figure 2c). It showed
extracellular lakes of mucin in which tumour cells
were present in small clusters together with neuroendo-
crine cells (Figure 2c). The mucin stained mainly with
Alcian blue, but some traces of diastase–periodic acid–
Schiff positive material were also present. The acid
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.

mucin was of non-sulphated sialomucin type, blue
staining with high iron diamine–Alcian blue. The
Grimelius silver staining demonstrated that the major-
ity of neuroendocrine cells were argyrophilic, whereas
the argentaffin reaction of Fontana–Masson was
negative. The six lymph nodes found were free of
tumour deposit. Very strong immunoreactivity for
chromogranin A was identified in carcinoid nests as
well as in neuroendocrine cells of the microglandular-
goblet cell carcinoid and the mucinous carcinoma
(Figures 1b and 2c). The results of an immunohisto-
chemical search for the presence of specific endocrine
products revealed a moderate number of serotonin-
immunoreactive cells within the neuroendocrine por-
tion of the tumour (Figure 2b), but no evidence of
gastrin, somatostatin, pancreatic polypeptide, VIP,
ACTH, glucagon, or calcitonin. In the non-malignant
gastric mucosa there was a nonerosive, non-specific
gastritis without evidence of Helicobacter pylori infec-
tion. Chromogranin-positive cells were singly scattered
and did not form nodular aggregates.
Our case was not associated with atrophic gastritis,
metaplastic epithelium, dysplasia or with multiple
proliferative lesions, such as intramucosal microcarci-
noid and endocrine cell proliferations of the micro-
nodular and linear type, which are currently regarded
as carcinoid precursor changes2,3. Therefore, we
suggest that the composite early gastric carcinoma
arises de novo from pluripotential neck stem cells, which
are capable of endocrine and epithelial differentiation.
The intimate admixture of the different cell types and
patterns within the tumour is also compatible with this
hypothesis.
In 1991, Yang and Rotterdam4 reported one case
of mixed carcinoid-adenocarcinoma of the stomach
and reviewed 20 cases from the literature, approxi-
mately evenly divided between well differentiated and
poorly differentiated adenoendocrine cell carcinoma.
Recently, a composite carcinoma of the stomach con-
sisting of endocrine and mucous epithelial cells with
interspersed amphicrine cells was reported5. The
present case differs from that reported previously in
its range of histopathological patterns including areas
of carcinoid, microglandular-goblet cell carcinoid,
tubular adenocarcinoma and neuroendocrine mucinous
adenocarcinoma.
R.A.Caruso
M.F.Heyman*
L.Rigoli†
C.Inferrera
Department of Human Pathology and
†Medical Genetics, University of Messina,
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.
Correspondence 571
Italy and
*Laboratoire D’Anatomie Pathologique B,
Centre Hospitalier
Regional et Universitaire de Nantes,
France
1. Lewin K. Carcinoid tumors and the mixed (composite) glandular-
endocrine cell carcinomas. Am. J. Surg. Pathol. 1987; 11 (Suppl. 1);
71–86.
2. Bordi C, Yu JY, Baggi MT et al. Gastric carcinoids and their
precursor lesions. A histologic and immunohistochemical study of
23 cases. Cancer 1991; 67; 663–672.
3. Capella C, Hetz PU, Hofler H, Solcia E, Kloppel G. Revised
classification of neuroendocrine tumors of the lung, pancreas and
gut. Virchows Archiv 1995; 425; 547–560.
4. Yang GCH, Rotterdam H. Mixed (composite) glandular-endocrine
cell carcinoma of the stomach. Report of a case and review of
literature. Am. J. Surg. Pathol. 1991; 15; 592–598.
5. Pasquinelli G, Santini D, Preda P, Cariani G, Bonora G, Martinelli
GN. Composite gastric carcinoma and precursor lesions with
amphicrine features in chronic atrophic gastritis. Ultrastruct.
Pathol. 1993; 17; 9–24.
Solitary fibrous tumour of the submandibular
gland
Sir : We describe a solitary fibrous tumour (SFT) within
the submandibular salivary gland, a rare location.
Although most evidence favours an origin from
mesenchymal cells1,2, the histogenesis of these tumours
is still uncertain. An immunohistochemical analysis of
cellular markers has been performed and, to further
investigate the nature of the lesion, the extracellular
matrix has also been examined.
A 46-year-old woman noticed a painless swelling in
the right submandibular region, that had appeared
about 1 year before. Physical examination revealed a
well circumscribed, nontender nodule in the right
submandibular gland, and the patient underwent
surgical excision of the affected gland. The resected
submandibular gland contained a well circumscribed,
firm, grey-pinkish 23 × 19 mm nodule with a central
brownish cystic cavity (Figure 1). Microscopically, the
lesion comprised a patternless proliferation of spindled-
to-ovoid cells associated with a fibrous matrix (Figure
1). Vascularity was prominent, and in some fields
suggested a haemangiopericytic pattern. Tumour cells
exhibited a diffuse, strong immunoreactivity for vimen-
tin and CD34 antigen, and a moderate cytoplasmic
staining for bcl-2 protein. Apart from vascular base-
ment membranes, the tumour tissue resulted negative
for laminin and type IV collagen. However, in focal
areas, a patchy pericellular staining for both of these
substances was seen in the tumour (Figure 2). The
abundant fibrous matrix interposed among tumour cells
572 Correspondence
Figure 1. Submandibular gland prenchyma (top) is seen at the
periphery of the tumour (bottom). Tumour cells have elongated
nuclei and poorly defined cytoplasm (inset) (haematoxylin & eosin).
was immunoreactive mainly for type III collagen (Figure
2) and, to a lesser degree, for fibronectin and tenascin.
Scant CD68-positive macrophages were present,
whereas antibody to Factor XIIIa stained many
histiocytes interspersed among tumour cells.
The location of SFT within the salivary glands is very
unusual, and only occasional examples have been
documented in the literature2. On ultrastructural
examination, presence of basement membrane material
around neoplastic elements has been described in some
but not all SFTs1. We found in the present case a focal
patchy immunostaining for both basement membrane
Figure 2. Immunostaining for extracellular matrix components
discloses a patchy pericellular reactivity for type IV collagen. The
intercellular matrix shows widespread staining with anti-type III
collagen antibody (inset) (ABC method on paraffin sections).
components type IV collagen and laminin around
tumour cells. This finding agrees with the ultrastruc-
tural evidence of basement membrane in some SFTs
reported in literature and, in turn, could be in keeping
with the postulated differentiation of the tumour cells
towards myofibroblasts1 which are known to be partly
enveloped by a basement membrane3. Most of extra-
cellular matrix was immunoreactive for type III
collagen, fibronectin and tenascin, all of which are
mesenchymal-type matrix substances that have been
found in lesions with fibroblastic/myofibroblastic differ-
entiation4. In our case the only immunohistochemically
detectable cytoskeletal filament was vimentin that, in
absence of any ultrastructural evidence, is not specific to
substantiate a myofibroblastic character of the tumour
cells. Due to their capability to undergo morpho-
functional modulation in response to environmental
stimuli, fibroblasts seem to represent an extraordinarily
heterogeneous cell population, among which the myofi-
broblastic phenotype is only one type3. This view could
be in line with the somewhat changeable immunophe-
notype observed in SFT which, in addition to variably
express fibroblastic/myofibroblastic markers, has some-
times been found to label for totally unrelated substances
such as neurofilaments and neuron-specific enolase2.
As observed in other SFTs1,2,5 our tumour exhibited a
vimentin-CD34-bcl-2-positive phenotype and, in addi-
tion, comprised a population of factor XIIIa-immuno-
reactive cells. The latter have often been found in CD34-
positive tumours6. Occurrence of vimentin-CD34-posi-
tive mesenchymal cells has been observed in both adult
and fetal normal submandibular glands (M.Guarino,
F.Giordano, and F.Pallotti, unpublished observations).
Whether these fibroblast-like cells are the origin of the
SFT we described, or whether the vimentin-CD34-
positive immunophenotype has emerged during tumour
development from relatively more undifferentiated
mesenchymal cells, is speculative.
M.Guarino
F.Giordano
F.Pallotti
S.Ponzi*
Departments of Anatomical Pathology and
*Otorhinolaryngology,
Hospital of Vimercate,
Milan, Italy
1. Steinez C, Clarke R, Jacobs GH, Abdul-Karim FW, Petrelli M,
Tomashefski JF. Localized fibrous tumors of the pleura: correlation
of histopathological, immunohistochemical and ultrastructural
features. Pathol. Res. Pract. 1990; 186; 344–357.
2. Hanau CA, Miettinen M. Solitary fibrous tumour: histological and
immunohistochemical spectrum of benign and malignant variants
presenting at different sites. Hum. Pathol. 1995; 26; 440–449.
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.
 Correspondence 573

3. Schmitt-Graff, Desmoulière A, Gabbiani G. Heterogeneity of

myofibroblast phenotypic features: an example of fibroblastic cell
plasticity. Virchows Arch. 1994; 425; 3–24.
4. Berndt A, Kosmehl H, Katenkamp D, Tauchmann V. Appearance of
the myofibroblastic phenotype in Dupuytren’s disease is associated
with a fibronectin, laminin, collagen type IV and tenascin
extracellular matrix. Pathobiology 1994; 62; 55–58.
5. Chilosi M, Facchetti F, Dei Tos AP et al. bcl-2 expression in pleural
and extrapleural solitary fibrous tumours. J. Pathol. 1997; 181:
362–367.
6. Silverman JS, Tamsen A. Mammary fibroadenoma and some
phyllodes tumour stroma are composed of CD34þ fibroblasts and
factor XIIIaþ dendrophages. Histopathology 1996; 29; 411–419.

Thymoma arising with a thymolipoma

Sir : On review of a series of thymomas from the personal
consultation files of one of the authors (JR), we
encountered a unique case of a hitherto undescribed
occurrence: a thymoma arising within a thymolipoma.
The patient was a 67-year-old, otherwise healthy
female who was found to have an anterior mediastinal
mass on radiographic examination. No symptoms of
myasthenia gravis were reported. At thoracotomy, a
100 mm, well circumscribed ovoid mass was found
attached to the pericardium. Following resection, the
patient was free of tumour recurrence for 10 years, after
which she was lost to follow-up.
On cut section, most of the 100 × 60 × 40 mm lesion
was lobulated, soft and yellow, consistent with adipose
tissue. At one pole, a 25 mm firm, circumscribed while
nodule was identified (Figure 1).
Microscopically, the larger lesion comprised an even
admixture of approximately 40% unremarkable mature
adipose tissue and approximately 60% normal thymic
parencyma. Both thymic cortex and medulla containing
Hassal’s corpuscles were present (Figure 2a). In
contrast, the 25 mm nodule was a cellular lesion
divided by thin fibrous bands emanating from a fibrous
capsule (Figure 2b). It comprised plump epithelioid cells
having pale chromatin, prominent nucleoli and pink ill-
defined cytoplasm, along with equal numbers of
interspersed mature lymphocytes. Mitoses were scarce.
Perivascular serum lakes and areas representing
medullary differentiation were identified. We categor-
ized this encapsulated thymoma as predominantly
mixed under the Lattes–Bernatz classification and
cortical under the Müller–Hermelink classification. A
thin rim of normal lobulated thymic tissue uninvolved
by either thymolipoma or thymoma was microscopically
identified at the periphery.
Thymolipoma is usually asymptomatic and inciden-
tally identified as a large, well circumscribed anterior
mediastinal mass. Reported cases have reached 16 kg
and 360 mm1. Because the mass conforms to the shape
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.
Figure 1. Cut section.
of the adjacent pericardium, it can simulate cardiome-
galy radiographically2. Thymolipoma is distinguished
from the more common mediastinal lipoma by the even
admixture of thymic parenchyma with adipose tissue
throughout and the increase in thymic parenchyma
relative to normal for the patient’s age.
While several theories have been proposed, the
pathogenesis of thymolipoma remains controversial.
One theory, citing the haphazard mixture of thymic
parenchyma and fat, holds that it is a thymic
hamartoma1. A malformative theory3 is supported by
a reported case in which parathyroid tissue was
admixed with a tumour resembling thymolipoma.
This case suggests possible origin from aberrant
development of the third pharyngeal pouch. Other
theories suggest that thymolipoma represents fatty
regression of a thymoma or of a previously hyperplastic
thymus. Perhaps the most intriguing theory is of
thymolipoma representing a benign tumour of specia-
lized thymic stroma4 which maintains its relationship
with the thymic epithelium as it grows. In this scenario,
thymolipoma would be analogous to other tumours of
specialized stroma such as fibroadenoma of the breast
and adenofibroma of the uterus. To further support this
analogy, cases of a malignant tumour of specialized
thymic stroma, thymoliposarcoma, have been reported5.
In these, liposarcoma expands the stroma between
lobules of thymic parenchyma, again preserving the
relationship of thymic stroma to epithelium.
574 Correspondence

Department of Pathology,
Memorial Sloan-Kettering Cancer Center, New York, NYand
*Laboratory of Pathology,
Buenos Aires,
Argentina

1. Moran CA, Rosado-de-Christenson M, Suster S. Thymolipoma:

clinicopathologic review of 33 cases. Mod. Pathol. 1995; 8; 741–
744.
2. Rosado-de-Christenson ML, Pugatch RD, Moran CA, Galobardes J.
Thymolipoma: analysis of 27 cases. Radiology 1994; 193; 121–
126.
3. van Hoeven KH, Brennan MF. Lipothymoadenoma of the para-
thyroid. Arch. Pathol. Lab. Med. 1993; 117; 312–314.
4. Hall GFM. A case of thymolipoma with observations on a possible
relationship to intrathoracic lipomata. Br. J. Surg. 1948; 36; 321–
324.
5. Havlicek F, Rosai J. A sarcoma of thymic stroma with features of
liposarcoma. Am. J. Clin. Pathol. 1984; 82; 217–224.
6. Hull MT, Warfel KA, Kotylo P, Goheen MP, Brown JW. Proliferating
thymolipoma: ultrastructural, immunohistochemical, and flow
cytometric study. Ultrastruct. Pathol. 1995; 19; 75–81.

Jejunal hamartoma as a rare cause of

gastrointestinal haemorrhage
Sir : Magnus Alsleben first described myoepithelial hamar-
tomas or adenomyomas in the submucosa of the stomach
in 19031. These tumours are very rare. They predomi-
nantly occur in the antrum, duodenum and then in
decreasing numbers in the remainder of the small bowel2.
Figure 2. Microscopy shows, a, large lesion to be composed of unre-
Adenomyomas are considered to be hamartomas since
markable mature adipose tissue and normal thymic parenchyma, they contain muscular, glandular and fatty elements. We
and, b, nodule to consist of thin fibrous bands emanating from the report the second case of a small bowel myoepithelial
fibrous capsule. hamartoma causing gastrointestinal haemorrhage.
A 65-year-old man was referred to our hospital with
a 4-day history of melaena. There were no other
This unique case of thymoma arising within thymo-
lipoma may be interpreted in several ways. One can
suggest that it strengthens the link between thymoma
and thymolipoma that is supported by their common
association with myasthenia gravis, though myasthenia
gravis is far more commonly associated with thymic
hyperplasia than with either of these two. If one views
thymolipoma as a benign tumour of thymic stroma
analogous to fibroadenoma of the breast, a thymoma
arising within it would be somewhat analogous to the
development of epithelial neoplasia in fibroadenoma,
such as lobular carcinoma in situ. Of note, a case of
thymolipoma showing diffuse thymic epithelial
proliferation6 has been reported.
P.Argani
I.C.K.de Chiocca* Figure 1. The submucosal hamartoma clearly visible in the small
J.Rosai bowel wall.

q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.


Figure 2. Detail of the hamartoma.
complaints. Laboratory tests showed only an anaemia.
Oesophago-gastro-duodenoscopy and colon-radiography
revealed no abnormalities. A small-bowel transit clearly
showed a side standing mass in the distal part of the
jejunum. At laparotomy, an intraluminal tumour was
found at about 1 m from the ileo-caecal valve. A segment of
small bowel containing the tumour was resected and a
primary end-to-end anastomosis performed. Postoperative
recovery and follow-up was uneventful; anaemia disap-
peared. Histologically the lesion was submucosal and
composed of glands surrounded by muscular and con-
nective tissue. The epithelium of the glands was cylindric in
shape and cystically dilated with no signs of cytological
atypia—a jejunal hamartoma (Figures 1 and 2).
Macroscopically the adenomyoma presents as a
sessile polyp with a diameter ranging from 5 to
15 mm. It is a submucosal lesion containing glands
lined by a columnar epithelium and smooth muscle
fibres. Diagnosis is always made postoperatively, simply
because most patients have no symptoms. Intussuscep-
tion is the first complication in the majority of cases.
High quality small-bowel transit radiography can only
confirm a lesion in the small bowel. But that gives us a
broad differential diagnosis. Even if the tumour can be
reached by endoscopy, it is still situated submucosally. So
in practice only resection and pathological investigation
gives a definitive diagnosis.
On reviewing the literature we could only find 13
case reports describing small-bowel myoepithelial
hamartomas: eight in the ileum, four in the jejunum
and one in a Meckel’s diverticulum3,4. Only one of these
patients had anaemia as presenting symptom5.
S.H.van Helden
G.Jutten
H.Van Hoey
A.M.Dierick*
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.
Correspondence 575
Departments of Surgery and
*Pathology, KLINA Brasschaat,
Heerlen,
The Netherlands
1. Vandelli A, Cariani G, Bonora G, Padovani F, Saragoni L, Dell
Amore D. Adenomyoma of the stomach. Surg. Endosc. 1993; 7;
185–187.
2. Olmsted WW, Ros PR, Hjermstad BM, McCarthy MJ, Dachman AH.
Tumours of the small intestine with little or no malignant
predisposition. A review of the literature and report of 56 cases.
Gastrointest. Radiol. 1997; 17; 231–239.
3. Serour F, Gorenstein A, Lipnitzky V, Zaidel L. Adenomyoma of the
small bowel: a rare cause of intussusception in childhood. J. Pediatr.
Gastroenterol. Nutr. 1994; 18; 247–249.
4. Gonzalvez J, Marco A, Andujar M, Iniguez L. Myoepithelial
hamartoma of the ileum: a rare cause of intestinal intussusception
in children. Eur. J. Ped. Surg. 1995; 5; 303–304.
5. Gourtsoyiannis NC, Bays D, Papaioannou N, Theotokas J, Barouxis
G, Karabelas T. Benign tumours ofthe small intestine: preoperative
evaluation with barium infusion technique. Eur. J. Radiol. 1993;
16; 115–125.
Genital carcinoma secondary to pagetoid
spread from a pagetoid urothelial carcinoma
in-situ
Sir : Pagetoid spread of urothelial carcinoma to the
genital tract is a rare phenomenon producing a pattern
resembling that of Paget’s disease. We report a case of
pagetoid urothelial carcinoma in-situ of the bladder
associated with pagetoid spread to the vulva and the
cervix. This was associated with bilateral ureteric and
urethral pagetoid extension.
A 71-year-old woman presented in 1995 with
painless haematuria. Following the diagnosis of carcin-
oma in-situ, she was treated for 3 months with BCG.
Eighteen months later, a routine cervical/vaginal smear
demonstrated malignant urothelial cells, although
clinical examination was negative at this time and the
patient was asymptomatic. As the haematuria and the
pelvic pain dramatically increased, a radical cystectomy
with hysterectomy and resection of the vagina was
performed. The patient is well 3 months after surgery.
Typical carcinoma in-situ was seen in the original
biopsies and the cystectomy specimen (Figure 1a). In
the bladder, pagetoid changes was extensively present.
Similar cells were present in the urothelium of both
ureters and the urethra (Figure 1b). They were also
identifiable in atrophic epithelium of the vaginal and
cervix (Figure 2). There was no evidence of invasive
disease. Staining for Alcian blue and PAS confirmed the
absence of mucins. The malignant cells were positive for
epithelial membrane antigen only and negative for CEA,
S100, HMB45 and PSA.
576 Correspondence
Pagetoid changes in urothelial carcinoma in-situ are
rare and only 0.6% of all the bladder carcinomas had
focal pagetoid features1. These changes occur either in
pure CIS or in CIS areas accompanied with papillary
and/or invasive carcinomas and these changes are often
present in patients first treated with radiotherapy2,3,
with chemotherapy1, by diathermy1,4 or by bacillus
Calmette–Guerin1. Pagetoid changes do not seem to
have clinical or prognostic implications1. The histo-
genesis of pagetoid changes in urothelial carcinoma
remains unknown. However, non-specific local reaction
of the neoplastic cells to injury1 or loss of E-cadherin
expression5 have been proposed.
The association of genital Paget’s disease with bladder
carcinoma has been previously described. Thus, there
have been reported cases of Paget’s disease of the glans
penis due to a urothelial carcinoma2,4, but there were
pagetoid changes in only one case of these bladder
carcinomas4, and only one case of these tumours was
an in-situ carcinoma3. Association of Paget’s disease of
the vulva with bladder carcinoma has also been
Figure 1. a, Bladder specimen: urothelial carcinoma in-situ with
large atypical cells dispersed in a pagetoid fashion. b, Pagetoid
urothelial cells dispersed along the urethral mucosa (haematoxylin
& eosin).
documented3,4,6. In the study published by Breen6,
Paget’s disease represented only 5% of the carcinomas
of the vulva. In this study of 98 cases, only 25% of the
cases of vulvar Paget’s disease were associated with
other malignancies which were mainly gynaecological
carcinomas, and only two cases were related to
transitional bladder carcinomas. Three situations can
explain the association of a Paget’s disease and a bladder
tumour: the coincidental presence of two unrelated
malignancies with no continuity between the two
lesions4; an extension of the Paget’s disease into the
urinary tract4; and, as in our case, a pagetoid extension
to the vulva of the bladder carcinoma3,4.
Our case is unusual because of the extent of the
gynaecologic involvement. Pagetoid extension of an in-
situ urothelial carcinoma into the vagina up to the
cervix has not been previously described. This gynae-
cologic extension did not give rise to any symptoms and
was detected by chance on a routine cervico-vaginal
smear. Occult and asymptomatic pagetoid spread is yet
known in other localizations, and the pathologist must
be aware of the possibility of urethral, ureteral, or
genital extension in such pagetoid urothelial carcino-
mas in-situ or infiltrating. As in our case, frozen section
Figure 2. The carinomatous pagetoid cells involved the epithelium of
the vagina (a) and the epithelium of the cervix (b) (haematoxylin &
eosin).
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.
 Correspondence 577

during surgery can help to verify the status of the visualized but showing no abnormality. No further
surgical margins. investigations were performed.
Macroscopically the specimen weighed 320 g and
L.Arnould*
was unremarkable except for a 40 mm diameter fundal
L.Chalabreysse†
fibroid mass. Histological examination of one section
C.Belichard‡
showed a typical leiomyoma. However, another section
J.Cuisenier‡
showed an infiltrate of mitotically-inactive small
C.Billerey§
rounded cells with mildly pleomorphic nuclei and a
F.Collin*
small amount of eosinophilic cytoplasm. These were
Departments of *Pathology and orientated in single or ‘Indian’-file between the smooth
‡Surgery, Centre G.F.Leclerc, Dijon; muscle cells (Figure 1a). Further sections failed to reveal
†Department of Pathology, a similar infiltrate. The infiltrating round cells were
University Hospital, Strasbourg; and positive for cytokeratin, CAM5.2 (Figure 1b) and
§Department of Pathology, negative for desmin on immunohistochemical staining.
University Hospital, The small cell size, ‘Indian’-file pattern and immuno-
Besançon, France histochemical profile in this case allowed a diagnosis of
leiomyoma with metastatic adenocarcinoma, probably
lobular from the breast, to be made. This finding
1. Orozco RE, Zwaag RV, Murphy WM. The pagetoid variant of resulted, on re-examination of the patient, in the
urothelial carcinoma in situ. Hum. Pathol. 1993; 24; 1199–1202. discovery of marked right-sided nipple and skin retrac-
2. Tomaszewski JE, Korat OC, Livolsi VA, Connor AM, Wein A. Paget’s
disease of the urethral meatus following transitional cell carcinoma
tion and a 40 mm mass behind the nipple. Fine
of the bladder. J. Urol. 1986; 135; 368–370. needle aspiration cytology, mammography and breast
3. Turner AG. Pagetoid lesions associated with carcinoma of the
bladder. J. Urol. 1980; 123; 124–126.
4. Powell FC, Bjornsson J, Doyle JA, Cooper AJ. Genital Paget’s disease
and urinary tract malignancy. J. Am. Acad. Dermatol. 1985; 13;
84–90.
5. Rebel JMJ, Thijssen CDEM, Vermey M. E-Cadherin expression
determines the mode of replacement of normal urothelium by
human bladder carcinoma cells. Cancer Res. 1994; 54; 5488–
5492.
6. Breen JL, Smith CI, Gregori CA. Extramammary Paget’s disease.
Clin. Obstet. Gynecol. 1978; 21; 1107–1115.

Fortuitous diagnosis in a uterine leiomyoma

of metastatic lobular carcinoma of the breast
Sir : Tumour-to-tumour metastasis is a rare, but well-
described phenomenon. We present a case of metastasis,
from a previously undiagnosed breast carcinoma,
detected in a leiomyoma during routine microscopic
examination of a uterus removed for treatment of
menorrhagia.
A 54-year-old woman presented with irregular
vaginal bleeding which was unresponsive to norethis-
terone. An uncomplicated total abdominal hysterect-
omy and bilateral oophorectomy was performed and the
patient was discharged home on oestrogen patch
hormone replacement therapy. The patient’s past
medical history included investigations for breast
lumpiness and an FNAC diagnosis of fibroadenosis on
the left. In 1994 the patient had presented with
indentation of the skin of the right breast near the Figure 1. a, H & E. b, Immunohistochemical staining with
nipple. A mammogram was reported as poorly cytokeratin (CAM5.2).

q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.

578 Correspondence
ultrasound confirmed primary lobular breast carcinoma
and no other metastases were seen on bone scintogram
or abdominal ultrasound.
Histopathological examination of uterine leiomyomas
may result in the diagnosis of unsuspected leiomyo-
sarcoma, or atypical leiomyoma, the main differential
diagnosis in this case. In addition, a number of cases
have been reported of breast cancer metastases dis-
covered incidentally within leiomyomas1–4. However, in
these cases the diagnosis of tumour-to-tumour meta-
stasis was usually made at autopsy in the presence of
numerous other metastases or represented an unusual
site of recurrence of a known malignancy. Therefore the
phenomenon was interesting but did not affect the
patient’s management.
The histopathological and immunohistochemical
findings in this case were characteristic of metastatic
lobular carcinoma from the breast. Interestingly, the
atypical cells were present in only one of the six sections
examined. The primary lesion had previously remained
undiagnosed despite presentation to a breast clinic,
clinical examination on several occasions and admis-
sion to hospital for hysterectomy. In addition, the
patient had been prescribed hormonal treatment for
menorrhagia and, immediately following hysterectomy
and bilateral oophorectomy, had been commenced on
hormone replacement therapy. The histopathological
diagnosis of metastatic lobular breast carcinoma
enabled hormone replacement therapy to be discon-
tinued and prompted referral for appropriate further
investigation and management.
To our knowledge this is the first reported case in
which a primary breast cancer was only diagnosed
following the fortuitous finding of metastasis to a
uterine leiomyoma. This case clearly illustrates the
necessity for pathological investigation, with adequate
sampling and thorough microscopic examination by a
trained pathologist, of all surgical specimens, no matter
how apparently routine.
R.D.Liebmann
K.D.Jones*
R.Hamid*
M.Lapsley
Department of Histopathology and
*Obstetrics and Gynaecology,
St. Helier Hospital,
Carshalton,
Surrey, UK
1. Spiro RK. Breast cancer metastatic to a uterine leiomyoma. J. Med.
Soc. N. Jersey. 1979; 76; 285–287.
2. O’Brien TF, Gaber LW. Extragenital metastases to uterine leiomyo-
mata. A case report. J. Reprod. Med. 1992; 37; 476–478.
3. Kumar NB, Hart WR. Metastases to the uterine corpus from
extragenital cancers. Cancer 1982; 50; 2163–2169.
4. Beattie GJ, Duncan AJ, Paterson AJ, Williams ARW, Geirsson RT.
Breast carcinoma metastatic to uterine leiomyoma. Gynaecol. Oncol.
1993; 51; 255–257.
5. Banoni F., Labes J., Goodman A. A uterine leiomyoma containing
metastatic breast carcinoma. Am. J. Obstet. Gynaecol. 1971; 111;
427–430.
q 1998 Blackwell Science Ltd, Histopathology, 32, 568–578.