clarified, this particular pattern seems to be a useful
hallmark for predicting PDGFRA gene status. By
analogy, perinuclear dot-like immunodecoration has
been described recently for other useful diagnostic
markers in a number of tumours.8–10
In conclusion, strong PDGFRA immunostaining
(either dot-like, cytoplasmic or membranous) in com-
bination with weak or absent CD117 staining (but
strong DOG1 positivity) appears to predict the pres-
ence of a PDGFRA mutation in the overwhelming
approach is expected to be time-saving, either where
the analysis is complemented by sequencing proce-
dures starting from the PDGFRA gene, or in providing
additional useful information where there is a lack of
available molecular analysis support.
Angelo P Dei Tos2
1Department of Pathology, Fondazione IRCCS Istituto
Nazionale dei Tumori, Milan, Italy and2Department of
Pathology General Hospital of Treviso, Italy
1. Novelli M, Rossi S, Rodriguez-Justo M et al. DOG1 and CD117 are
the antibodies of choice in the diagnosis of gastrointestinal stromal
tumours. Histopathology 2010; 57; 259–270.
2. Corless CL, Schroeder A, Griffith D et al. PDGFRA mutations in
gastrointestinal stromal tumors: frequency, spectrum and in vitro
sensitivity to imatinib. J. Clin. Oncol. 2005; 23; 5357–5364.
3. Miettinen M, Wang ZF, Lasota J. DOG1 antibody in the differential
diagnosis of gastrointestinal stromal tumors. A study of 1840
cases. Am. J. Surg. Pathol. 2009; 33; 1401–1408.
4. Miselli F, Millefanti C, Conca E et al. PDGFRA immunostaining can
help in the diagnosis of gastrointestinal stromal tumors. Am. J.
Surg. Pathol. 2008; 32; 738–743.
5. Negri T, Bozzi F, Conca E. Oncogenic and ligand-dependent
activation of KIT ⁄ PDGFRA in surgical samples of imatinib-treated
gastrointestinal stromal tumours (GISTs). J. Pathol. 2009; 217;
6. Wardelmamn E, Hrychyk A, Merckelbach-Brusse S et al. Associ-
ation of platelet-derived growth factor receptor alpha mutations
with gastric primary site and epithelioid or mixed cell morphology
in gastrointestinal stromal tumors. J. Mol. Diagn. 2004; 6;
7. Yamamoto H, Kojima A, Nagata S et al. KIT-negative gastrointestinal
stromal tumor of the abdominal soft tissue: a clinicopathologic and
genetic study of 10 cases. Am. J. Surg. Pathol. 2011; 35; 1287–1295.
8. Guo Y, Yuan F, Deng H, Wang HF, Jin XL, Xia JC. Paranuclear dot-
like immunostaining for CD99: a unique staining pattern for
diagnosing solid-pseudopapillary neoplasms of the pancreas. Am. J.
Surg. Pathol. 2011; 35; 799–806.
9. Gerald WL, Rosai J. Desmoplastic round cell tumor with
multi-phenotypic differentiation. Zentralbl. Pathol. 1993; 193;
10. Guo Y, Yuan F, Deng H. Paranuclear dot-like immunostaining for
CD99: a unique staining pattern for diagnosing solid-pseudopap-
illary neoplasm of the pancreas. Am. J. Surg. Pathol. 2011; 35;
Assessment of integrase interactor 1 (INI-1)
expression in primary tumours of bone
? 2012 Blackwell Publishing Limited.
Sir: Epithelioid sarcoma (ES) is a distinctive high-grade
soft-tissue tumour affecting predominantly the distal
extremities of young people. The differential diagnosis
includes a locally arising and metastatic carcinoma,
because in addition to the epithelioid morphology it is
all cases. Reaching a diagnosis is now helped by the
absence of expression of integrase interactor 1 (INI-1) in
more than 90% of ES, whereas this rarely occurs in
carcinomas, apart from renal medullary carcinomas.1
The INI-1 gene is implicated in chromatin remodel-
ling and transcriptional regulation of cells, and is
expressed in all cell types.2Loss of INI-1 immunoreac-
tivity is also seen in malignant rhabdoid tumour, in
addition to other neoplasms including 50% of malig-
nant peripheral nerve sheath tumours, and in a
minority of extraskeletal myxoid chondrosarcomas,
myoepithelial carcinomas and chordomas.
A case of ES of the ilium in a 31-year-old female has
been described recently.3The tumour exhibited the
classical histology and immunoprofile, including the
loss of INI-1 protein. However, the authors also
commented on the presence of intercellular hyalinized
and partly calcified matrix, which made us consider the
diagnoses of osteosarcoma and chondroblastomas,
particularly as both may also express cytokeratins.4,5
The article prompted us to assess the immunoreactivity
of INI-1 in a series of bone tumours known to express
cytokeratin, as little is known about the expression of
INI-1 in these. For completeness, we also included
conventional chondrosarcomas which do not express
cytokeratins, as they represent the second most com-
mon malignant bone tumour.
Seven hundred and twenty-four tumours, including
280 osteosarcomas, 140 chordomas, 21 chondrobl-
astomas, 17 metastatic cancers of unknown primary
(CUP), 16 CK-positive Ewing sarcomas, 15 CK-
positive spindle cell sarcoma of bone, six CK-positive
Histopathology, 61, 1225–1248.
dedifferentiated chondrosarcoma, six adamantinomas,
two angiosarcomas of bone, two desmoplastic small
round cell tumours of bone, one malignant mixed
tumour of bone and 218 chondrosarcomas, were
included. Immunoreactivity in osteosarcomas, chor-
domas, chondrosarcomas and chondroblastomas was
assessed on our tissue microarrays; the others were
analysed on full sections. A conventional ES of soft
tissue was used as a control. The BAF47 antibody
(clone 25; 1 ⁄ 400; Becton-Dickinson, Franklin Lakes,
NJ, USA) was employed.
Only five of 724 (0.6%) tumours showed loss of INI-1
expression, all of which were confirmed using full tissue
sections: they included two osteosarcomas (two of 280),
two chordomas (two of 140) and one CUP. The INI-1
negative CUP presented in the femur of a 25-year-old
focally showed rhabdoid features (Figure 1). There were
areas of tumour necrosis, but mitoses were scarce. There
CD3, CD20, CD30, CD31, CD34, CD45, CD99, CK5 ⁄ 6,
CK7, CK20, calretinin, WT1, alpha-feto protein, S100,
HMB45, SM-actin, desmin and myogenin, nor did it
reveal EWSR1 or SS18 gene rearrangements as assessed
by fluorescence in-situ hybridization (FISH). Despite
extensive clinicoradiological investigations, an obvious
primary visceral or soft tissue site was not found. In view
a primary epithelioid sarcoma of bone.
In conclusion, loss of INI-1 expression occurs rarely
in bone tumours, although it is observed very occa-
sionally in osteosarcoma and chordomas, and can also
rarely provide a diagnosis of a locally arising ES. This
study highlights the value of assessing INI-1 expression
Figure 1. A light photomicrograph of a haematoxylin and eosin section of a primary epithelioid sarcoma of bone. The tumour permeates host
bone (A). The cells have abundant eosinophilic cytoplasm: the nuclei were ovoid, with finely granular chromatin (B). The tumour is
immunoreactive for pankeratin (C) and there is no expression of INI-1, but note the immunoreactivity on lymphocytes (D).
Histopathology, 61, 1225–1248.
in all poorly differentiated cytokeratin-positive tumours Download full-text
in bone, particularly when a diagnosis of a CUP is being
considered, especially in young individuals.
declarations of interest
No conflicts of interest are declared.
This work was funded by Bone Cancer Research Trust,
and supported by the infrastructure of the Stanmore
Musculoskeletal Research Programme and Biobank.
Adrienne M Flanagan1,3
1Department of Histopathology, Royal National
Orthopaedic Hospital, Stanmore, Middlesex,2Department
of Histopathology, Great Ormond Street Hospital, London
and3UCL Cancer Institute, London, UK
1. Hornick JL, Dal Cin P, Fletcher CDM. Loss of INI1 expression is
characteristic of both conventional and proximal-type epithelioid
sarcoma. Am. J. Surg. Pathol. 2009; 33; 542–550.
2. Muchardt C, Yaniv M. When the SWI ⁄ SNF complex remodels…
the cell cycle. Oncogene 2001; 20; 3067–3075.
3. Raoux D, Peoc’h M, Pedeutour F et al. Primary epithelioid sarcoma
of bone. Report of a unique case, with immunohistochemical and
fluorescent in situ hybridization confirmation of INI1 deletion. Am.
J. Surg. Pathol. 2009; 33; 954–958.
4. Okada K, Hasegawa T, Yokoyama R et al. Osteosarcoma with
cytokeratin expression: a clinicopathological study of six cases
with an emphasis on differential diagnosis from metastatic cancer.
J. Clin. Pathol. 2003; 56; 742–746.
5. Bousdras K, O’Donnell P, Vujovic S et al. Chondroblastomas but
not chondromyxoid fibromas express cytokeratins: an unusual
presentation of a chondroblastoma in the metaphyseal cortex of
the tibia. Histopathology 2007; 51; 414–416.
Hepatic sinusoidal obstruction syndrome
reduces the effect of oxaliplatin in colorectal
? 2012 Blackwell Publishing Limited.
Sir: We read with interest the above paper by Vreuls
et al., which demonstrates that there is a statistically
significant correlation between high-grade sinusoidal
obstruction syndrome (SOS) and a failure to respond to
oxaliplatin-based chemotherapy. This important find-
ing strengthens the results of Tamandl et al.,1who
reported that the presence of SOS was associated with
poorer disease-specific survival, although it was diffi-
cult to draw any useful conclusions from this study
based on the small numbers reported.
Whilst no attempt is made in the current study to
oxaliplatin, the authors propose two hypotheses related
to SOS-induced hepatic hypoperfusion. The first of these
is that hypoperfusion results in activation of hypoxia-
in tumour angiogenesis and invasion. This theory is an
extrapolation of the findings of Rubbia-Brandt et al.,2
stress response genes [e.g. Jun, superoxide dismutase 2
(SOD2)] and angiogenic factors [e.g. vascular endothe-
from a series of patients with SOS. It should be
highlighted that there is no evidence that SOS is
associated with similar changes occurring within
tumour tissue itself over and above those occurring as
a consequence of chemotherapy alone.
The second theory proposed by the authors is that
hepatic hypoperfusion leads to decreased oxaliplatin
delivery to the tumour. Given that SOS occurs uni-
formly throughout the liver, it would seem to suggest
that there is a reasonable delivery of drug to the tissue
despite the developing SOS. While this theory may have
some merit, further investigation is needed before it can
An alternative explanation may be that SOS is
associated with biochemical changes in the liver
microenvironment which favour the progression of
liver metastases. In their gene expression study, Rub-
bia-Brandt et al.2reported that there was an 11-fold
increase in expression of the chemokine CCL20 within
the liver of patients with SOS. Rubie et al.3have
demonstrated that FOLFOX chemotherapy (folinic acid,
fluorouracil, oxaliplatin) is also associated with in-
creased expression of the CCL20 receptor, CCR6, within
colorectal liver metastases. It has been demonstrated
previously that by binding the CCR6 receptor, CCL20
can lead to increased migration and proliferation of
colorectal cancer cells.4Furthermore, it appears that
CCL20 produced by the liver is a chemoattractant to
CCR6-expressing colorectal cancer cells, thereby pro-
moting the development of colorectal liver metastases.5
It may therefore be that SOS leads to the development
of an autocrine loop favouring the growth and devel-
opment of colorectal liver metastases; this is worthy of
Perhaps the most important issue is to identify those
patients at particular risk of developing SOS following
oxaliplatin exposure, and thereby modify their treat-
Histopathology, 61, 1225–1248.