Oncocytic carcinoma of the breast: frequency, morphology
Moira Ragazzi MDa, Dario de Biase BSc PhDa, Christine M. Betts BSc MDb,
Anna Farnedi BSc PhDa, Saime Sezgin Ramadan MDc, Giovanni Tallini MDa,
Jorge S. Reis-Filho MD PhD FRCPathd,1, Vincenzo Eusebi MD FRCPatha,⁎
aDepartment of Haematology and Oncological Sciences “L. and A. Seragnoli”, Section of Anatomic Pathology at
Bellaria Hospital, University of Bologna, 40139 Bologna, Italy
bDepartment of Experimental Pathology, University of Bologna, 40126 Bologna, Italy
cDepartment of Pathology, Anadolu Medical Center, 41001 Kocaeli, Turkey
dThe Breakthrough Breast Cancer Research Centre, Institute of Cancer Research, SW3 6JB London, UK
Received 14 May 2010; revised 7 July 2010; accepted 8 July 2010
Summary Oncocytic breast carcinomas are tumors composed of no fewer than 70% of oncocytic cells
(World Health Organization). The purpose of this study was to determine the frequency, morphologic,
immunohistochemical, and clinical features of invasive oncocytic carcinoma in a large series. Twenty-
eight cases of putative oncocytic breast carcinoma (selected cases group) and 76 consecutive cases of
invasive breast carcinoma (consecutive cases group) were analyzed. Immunohistochemistry for
mitochondria, gross cystic disease fluid protein 15, chromogranin, estrogen receptor, progesterone
receptor, androgen receptor, HER2/Neu, cytokeratin 7, cytokeratin 14, epithelial membrane antigen, and
differentiation cluster 68 was performed. Score for mitochondria was based on intensity and percentage of
immunopositive cells. Classes were as follows: (1) oncocytic carcinoma: at least 70%, 3+; (2)
mitochondrion-rich carcinoma: 50% to 70%, 3+, or more than 50%, 2+; and (3) all the other cases were
referred to as invasive breast carcinoma. Ultrastructural examination was available for 6 cases of oncocytic
carcinoma. Morphologic and immunohistochemical features of the 3 groups were compared using Fisher
exact test (P b .05). For overall survival analysis, Kaplan-Maier curves were compared using log-rank and
Wilcoxon tests (P b .05). Our results suggest that oncocytic breast carcinoma is a morphologic entity with
distinctive histologic and ultrastructural features. Mitochondrion-rich carcinomas are histologically similar
to oncocytic carcinomas and constitute 19.7% of all invasive carcinomas, indicating that cytoplasmic
eosinophilia in breast cancer cells is often due to accumulation of mitochondria. Oncocytic carcinomas and
mitochondrion-rich carcinomas are more often grade III tumors and show human epidermal growth factor
receptor 2 overexpression. Clinical features and overall survival of oncocytic carcinomas are not
distinctive because they are similar to those of the other cases when matched for grade and stage.
© 2011 Elsevier Inc. All rights reserved.
Abbreviations: OC, oncocytic carcinoma; mt-rich, mitochondrion-rich; BTPTC, breast tumor resembling papillary thyroid carcinoma;
IBC, invasive breast carcinoma.
☆This paper was financed with funds from RFO ex 60% University of Bologna, Bologna, Italy.
⁎Corresponding author. Sezione di Anatomia Patologica, Ospedale Bellaria, Via Altura, 3, 40139, Bologna – Italy.
E-mail address: email@example.com (V. Eusebi).
1This author is funded in part by Breakthrough Breast Cancer.
0046-8177/$ – see front matter © 2011 Elsevier Inc. All rights reserved.
Human Pathology (2011) 42, 166–175
Oncocyte means “swollen cell” from ancient Greek
“Ογχοΰσθαι.” The term was coined in salivary glands by
Hamperl  in 1931 to describe a special type of epithelial
cell showing, at light microscopy, large size and deeply
acidophilic granular cytoplasm. In the thyroid, the same cell
was first described by Askanazy , whereas the designation
“Hürthle cell” is a misrepresentation because what Hürthle
described  were the ultimobranchial body-derived paraf-
ollicular calcitonin-producing cells .
Electron microscopy (EM) subsequently evidenced that
in the number and size of mitochondria, which also imparted
the cytoplasmic granularity . Askanazy cells (oncocytes)
thyroid follicles, as the former were replete with numerous
mitochondria evenly distributed throughout the cytoplasm,
whereas thelatter contained mitochondria mainly localizedat
the poles of follicular cells . The proposal by Ghadially 
of 60% of the total cytoplasmic volume loaded by
mitochondria has become the generally accepted cutoff
value for the designation of oncocyte. Even thought
oncocytic tumors may occur in several different anatomical
and gastrointestinal tract , they tend more frequently to
affect endocrine organs and glandular epithelia such as
thyroid, salivary glands, and renal tubules  that have high
metabolic activity. In the definition of oncocytoma, the
proportion of neoplastic cells exhibiting oncocytic features
varies in different organs, being 70% in breast , 75% in
thyroid neoplasm [14,15], and 100% in renal tumors .
The clinicaloutcome isquite variablein different tumors; it
is favorable in renal and parathyroid oncocytomas and 50% of
oncocytic fibrolamellar carcinoma of the liver , whereas
prognosis of oncocytic tumors of thyroid , salivary gland
oncocytic carcinomas (OCs) , oncocytic pituitary adeno-
mas [20,21], and oncocytic meningiomas [8,22] seems to be
less favorable than that of related nononcocytic neoplasms.
Antibodies used for tumors characterization
Antibodies Clone (Ventana
Abbreviation: CD, differentiation cluster.
aVentana-Benchmark, Tucson, AZ.
bDetected by “Universal alkaline phosphatase red detection
cDetected by “Ultraview universal DAB detection system.”
glands are strongly stained (score 3+), striated ducts moderately
stained (score 2+), and lymphocytes are negative.
Dilution test of the antibody in Warthin tumor. Oncocytic
tumor cells from a case of OC of the consecutive cases group. In the
inset (B), 3+ cytoplasmic positivity is shown. C, Moderate
positivity in more than 50% of cells in a mt-rich case from the
consecutive cases group.
A, Strong positivity observed in no fewer than 70% of the
167 Oncocytic carcinoma of the breast
In breast, primary OCs were defined as tumors composed
of no fewer than 70% of oncocytic cells . In spite of rare
reports [13,23], the incidence of OCs of the breast was felt to
be underestimated  probably because of histologic
similarities between oncocytes and apocrine cells . The
aim of this study was to define the frequency, morphologic,
and clinical features of invasive oncocytic breast carcinoma
2. Materials and methods
2.1. Case selection
All cases were retrieved from the files of the Section of
Anatomic Pathology “M. Malpighi” of the University of
Bologna at Bellaria Hospital.
2.1.1. Selected cases
Eight cases of invasive breast carcinomas (IBCs) labeled
as oncocytic (OC) were selected from the consultation files
of 1 of us (V.E.), and 20 cases similar to OC were obtained
from the routine files in the period from 2006 to 2009.
2.1.2. Consecutive cases
Seventy-six consecutive primary IBCs not treated with
any form of adjuvant therapy were retrieved from the routine
files throughout 2 years (1997-1998) to determine the
incidence of OCs in a consecutive series and to have a
reasonably adequate follow-up (FU) period (6-144 months;
mean, 95.89 months). The clinical notes were obtained from
the hospital charts. Recurrent lesions were excluded. FU was
obtained in 74 cases.
Sections obtained from formalin-fixed paraffin-embedded
tissue were stained with hematoxylin and eosin. Slides of all
cases were reviewed by 2 of the authors (M.R. and V.E.).
The World Health Organization histologic criteria for the
diagnosis of OC were used . Histologic grading was
carried out as described by Elston and Ellis . Tumor
extent (pT) and regional lymph node status (pN) were
assesed for each case, according to pathologic cancer staging
system TNM .
Serial, 3-μm-thick, paraffin sections mounted on pre-
coated slides were processed using standardized automated
procedures using prediluted antibodies (Ventana-Bench-
mark, Tucson, AZ), with the exception of the antimitochon-
drion antibody (Ab-2, clone MTC02; LabVision, Freemont,
CA). Immunohistochemistry was performed using Polymer
as detection system (Ultravision LP Detection System HRP
Polymer; LabVision). Source, antigen retrieval, and detec-
tion systems for all prediluted antibodies are reported in
CasesAge, mean (y)Sex DiagnosisGrade pT
A. Selected cases group, 28 tumors
OC17 26-85 (66.5)17 FIDC: 13
Mt-Rich 1049-87 (68)10 FIDC: 9
B. Consecutive cases group, 76 tumors
1 72 1FIDC-st: 1 G2: 1
15 49-94 (70.3)14 F IDC: 12G1: 1
1 MIDC-st: 3
Mt-Rich15 40-77 (63.9) 15 FIDC: 11
IBC 4632-96 (65.8) 44 F
Abbreviations: St, special type; F, female; M, male.
168M. Ragazzi et al.
The optimal dilution of the antimitochondrion antibody
(1:2000) was obtained after several tests on a case of Warthin
tumor. Dilution was optimized when staining of cytoplasm
of lymphocytes had faded, while the glandular pink
epithelium, notoriously constituted by oncocytes, was still
intensely positive (Fig. 1).
Immunostained cytoplasm for mitochondrion antibody
was scored as follows:
• 3+ (strong): when positivity was as strong as that seen
in oncocytic cells of Warthin tumor (This score was
considered the criterion standard of oncocytic cells in
the present series.)
• 2+ (moderate): when positivity was as intense as cells
of striated ducts that are rich in mitochondria.
• 1+/Neg: when staining was weak or absent as seen in
lymphocytes (Fig. 1).
Accordingly, lesions were classified into 3 different
• OC regarded cases in which strong positivity (3+) was
observed at least 70% of the tumor cells (Fig. 2A, B).
• Mt-rich carcinomas were those cases that displayed
moderate positivity (2+) more than 50% of the tumor
cells or cases showed strong positivity (3+) between
30% and 70% of neoplastic cells (Fig. 2C).
• All the other cases were regarded as IBCs not enriched
Immunohistochemical stains for estrogen receptor (ER),
progesterone receptor (PR), androgen receptor (AR), human
epidermal growth factor receptor 2 (HER2/Neu), gross cystic
disease fluid protein 15 (GCDFP-15), chromogranin,
epithelial membrane antigen (EMA), cytokeratin (CK) 7,
CK14, and differentiation cluster 68 were additionally
obtained in all cases.
The results of all the immunohistochemical stains were
assessed independently by 3 observers (G.T., M.R., and
V.E.). When discrepancies occurred, the case was discussed
on a multiheaded microscope, and a consensus was reached.
For ER and PR, the Saint Gallen  score of 10% of
positive cells was taken as the cutoff point. The same values
were followed for AR. For HER2/Neu, the American Society
of Clinical Oncology/College of American Pathologists 
guidelines were followed. All other antibodies were
OC, n = 32 (%)Mt-rich, n = 25 (%) IBC, n = 47 (%)OC vs mt-rich
P = .8597a
OC vs IBC
P = .9930a
Mt-rich vs IBC
P = .9021a
Age (y), mean (range)
Histologic subtypes (%)
Mixed (ILC + IDC)
Histologic grade (%)
Tumor size (%)b
Lymph node metastasis pN (%)c
Ductal carcinoma in situ
66.46 (26-94)65.7 (40-87) 66.29 (32-96)
P = 1.0000
P = .2728
P = .0786
P = .6046
P = .3585
P = .0759
P = .5118
P = .4151
P = .7052
P = 1.0000
P = .5403
P = 1.0000
P = 1.0000
P = .7937
P = .2312
P = .0651
P = .0043⁎⁎
P = .4091
P = .1361
P = .0192⁎
P = .0560
P = .2877
P = 1.0000
P = .2097
P = .0954
P = .2421
P = .5595
P = 1.0000
P = .8053
P = 1.0000
P = .2750
P = .2476
P = .3961P = .6154P = .8000
P = .5762
P = .3805
P = .5851
P = 1.000
P = .1332
P = .4503
P = .1871
P = .3521
P = .4819
P = 1.0000
P = .0309⁎
P = .3079
Abbreviation: DIN, ductal intraepithelial neoplasia.
aStudent t test.
bpT was not available for 1 OC.
cpN was not available for 7 OC, 1 mt-rich, and 4 IBC.
dComprising 1 LCIS/LIN (lobular carcinoma in situ/ lobular intraepithelial neoplasia).
⁎P b .05.
⁎⁎P b .01.
169Oncocytic carcinoma of the breast
considered positive when more than 10% of the total
neoplastic proliferation was stained.
2.4. Ultrastructural study
EM was obtained on 6 cases of OCs. Small samples were
microdissected from paraffin-embedded blocks brought to
buffered glutaraldehyde, postfixed in OsO4and embedded in
Epon 812 (Fort Washington, PA). Thin sections were stained
with uranyl acetate and Reynold lead citrate, and observed in
a Philips CM 10 transmission EM (Eindhoven, the
2.5. Statistical analysis
Overall survival from surgery was calculated by Graph-
Pad Prism tool. Kaplan-Maier curves were compared using
log-rank and Wilcoxon tests. A P value less than .05 was
considered statistically significant.
Morphologic and immunohistochemical features of the
3 groups were compared using Fisher exact test (P b .05).
Mean age was compared using Student t test (P b .05).
All P values were 2-tailed; 95% confidence intervals
3.1. Selected cases
All 28 patients were female, ranging in age from 26 to 87
years (mean, 65.3 years). Histologically, these tumors were
regarded as oncocytic and displayed patterns consistent with
invasive duct carcinoma (IDC) in 22 cases, neuroendocrine
carcinoma in 1 case, apocrine carcinoma in 3 cases, breast
tumor resembling papillary thyroid carcinoma (BTPTC) 
in 1 case, and invasive lobular carcinoma (ILC) in 1 case.
Fifty-seven percent of the cases (16/28) were grade III;
39.3% (11/28), grade II; and 3.6% (1/28), grade I.
3.2. Consecutive cases
Seventy-six consecutive IBCs were from 73 women and 3
men, ranging in age from 32 to 96 years (mean, 66.1 years).
Cases included 56 IDC, 8 ILC, 2 mixed (ductal/lobular)
invasive carcinomas, 6 mucinous carcinomas, 1 apocrine car-
cinoma, 1 neuroendocrine carcinoma, 1 invasive papillary
carcinoma, and 1 atypical medullary carcinoma. The distribu-
tion of histologic grade was as follows: 34.20% (26/76), grade
III; 55.30% (42/76), grade II; and 10.50% (8/76), grade I.
n = 32 (%)
n = 25 (%)
n = 47 (%)
OC vs IBC mt-rich
Patterns of growth Solid
Granular and strong
P = .4195
P = 1.000
P = .7779
P = .0745
P = .1783
P = .0001⁎⁎⁎ P = .0156⁎
P = .0006⁎⁎⁎ P = .0008⁎⁎⁎
P = .7927
P = .0376⁎
P = .0004⁎⁎⁎ P = .1275
P = .5751
P = .5044
Large cytoplasmP = .0470⁎ P b .0001⁎⁎⁎ P = .0011⁎⁎
P = .7345
P = .0352⁎ P = .4051
P = .0959
P = .1126P = .0450⁎
P = .0011⁎⁎
P b .0001⁎⁎⁎ P b .0001⁎⁎⁎
P = .0002⁎⁎⁎ P = .1290
NucleoliP = .0176⁎ P b .0001⁎⁎⁎ P = .0863
Cytoplasmic bordersaEvidentP = .0514P = .0006⁎⁎⁎ P = .3789
Lymphoid stromaP = .1075P b .0001⁎⁎⁎ P = .0123⁎
NOTE. “Other features” are micropapillary, plexiform, papillary, mixed.
Abbreviation: DIN, ductal intraepithelial neoplasia.
aCytoplasmic borders were not evaluable for 3 mt-rich and 5 IBC ILC.
⁎P b .05.
⁎⁎P b .01.
⁎⁎⁎P b .001.
170M. Ragazzi et al.
3.3.1. Antimitochondrion antibody
According to positivity to antimitochondrion antibody in
the selected cases group, 17 cases showed 3+ positivity at
least 70% of cells. These cases, which also included 1
carcinoma showing neuroendocrine differentiation, 1
BTPTC, and 2 carcinomas with diffuse apocrine differenti-
ation, were classified as “OC.” Ten cases, which also
included 1 ILC, were considered “mt-rich.” One case,
showing consistent apocrine differentiation, was classified
as “IBC” (Table 2A).
In the consecutive cases group of 76 tumors, antimito-
chondrion antibody revealed 15 “OC” (19.7%), 15 cases
(19.7%) were “mt-rich,” and 46 cases (60.6%) were
reclassified as “IBC” (Table 2B).
Considering together the 2 cohorts, 32 cases with OC
features were obtained. Pathologic and clinical findings of
these cases are summarized in Table 3.
The patients were 31 female (96.87%) and 1 male
(3.13%); their age ranged from 26 and 94 years (mean, 66.46
years). Tumor size ranged from 0.8 to 9 cm (mean, 2.94 cm).
When age, tumor size, lymph node status, and histologic
subtype were matched in the 3 groups, no statistical
difference was observed. Only grade III carcinomas were
significantly more frequent in the OC and mt-rich groups
than in IBC cases (Table 3).
Tumors with oncocytic features were originally diag-
nosed as IDC NOS (not otherwise specified) in 25 patients
(78.1%) and “special-types” in 7 (21.9 %; 2 pure mucinous
carcinomas, 3 apocrine, 1 BTPTC, and 1 neuroendocrine).
In 21 cases (Table 4), the neoplastic cells predominantly
displayed a solid architecture, featuring nests, sheets
(Fig. 3A, B), and cords (Fig. 4A) of tumor cells. In the 2
cases of mucoid carcinomas, a plexiform architecture was
evident; in the case of BTPTC, papillary architecture was
prominent; and finally, 2 cases displayed glandular struc-
tures. In 6 cases, 2 or more patterns coexisted together.
Sixteen tumors displayed circumscribed borders
(Fig. 5A), and the stroma contained lymphoid elements
in 18 cases with prominent lymphoid infiltration in 7
The neoplastic cells were round to ovoid in most of
the cases and showed large and strongly eosinophilic to
clear granular cytoplasm. In 2 cases, signet ring cells
were numerous. Nuclei were irregular and pleomorphic in
24 cases with prominent nucleoli in 22. In 2 cases, nuclei
were clear of the type seen in papillary carcinoma of the
sheets of neoplastic cells (consecutive cases group).
OC shows solid structure displaying (A) nests and (B)
arranged in cords and show eosinophilic cytoplasm, and nuclei are
pleomorphic with prominent nucleoli. B, Most of the cells are
strongly immunoreactive with the antimitochondrion antibody.
OC from consecutive cases group. A, Neoplastic cells are
171Oncocytic carcinoma of the breast
thyroid. Cytoplasmic borders were evident in 20 cases
When the histologic features were matched between OC
and mt-rich cases, the only statistically significant differ-
ences were strongly eosinophilic cytoplasm and prominent
nucleoli (Table 4 and Fig. 4A, B).
More consistent statistical differences were observed
between OC and IBC cases (Tables 3, 4, and 5). Accord-
ingly, a solid growth pattern with pushing margins were
observed more frequently in OC, whereas IBC cases were
more frequently glandular or showed diffuse growth as seen
in ILC. Intense granular eosinophilia of the cytoplasm,
pleomorphic nuclei, prominent nucleoli, neat cytoplasmic
borders, and moderate lymphoid stroma were all features
statistically more frequent in OC.
Carcinoma in situ was associated with an invasive
component in 64.4% of cases. No statistical difference was
found among carcinoma in situ type and grade among the 3
groups (OC, mt-rich, and IBC; Table 3).
3.4. Additional immunohistochemical findings
Table 5 compares the immunohistochemical profile of
OC, mt-rich, and IBC. HER2/Neu was statistically more
frequently positive (3+) in OC in comparison with IBC cases.
No other statistically significant differences were found.
3.5. Ultrastructural Findings
On electron microscopic analysis, the 6 samples of OC
showed similar ultrastructural features. Numerous mitochon-
dria were present scattered throughout the cytoplasm,
without evidence of polar condensation. They varied from
round to ovoid in shape, sometimes swollen with flattened
cristae. No secretory granules were observed (Fig. 7). Rare
lysosomes were apparent. Nuclei were round to ovoid in
shape and displayed dusty scattered chromatin, 1 to 2 small
nucleoli and neat and linear nuclear membrane.
There was no statistically significant difference (log-rank
and Wilcoxon tests) in survival rate between OC, mt-rich,
and IBC in the consecutive cases group, although a trend for
shorter survival was apparent when OCs were matched with
IBC cases (Fig. 8). No differences were observed for grade
and tumor extent (pT).
To date only 4 cases of oncocytic breast carcinomas have
been reported in the literature [13,23]. They include 3
invasive carcinomas and 1 carcinoma in situ, all patients
were older than 70 years, and all cases showed remarkably
good prognosis. Histologically, the 4 cases were well-
differentiated tumors composed of large cells having well-
defined borders and abundant eosinophilic granular cyto-
plasm. The case of carcinoma in situ reported by Damiani
et al hadapapillaryappearance.Inallcases,theoncocytic
cases group). B, The stroma contains lymphoid elements (selected
A, Most OCs show circumscribed borders (consecutive
large eosinophilic granular cytoplasm with well-evident neat
borders. Nuclei are irregular and show prominent nucleoli.
OC from the selected cases group. Neoplastic cells show
172 M. Ragazzi et al.
nature of the neoplastic cells was proven by ultrastructure and
also by immunohistochemistry in the 3 cases reported by
Damiani et al . All cases expressed ER with positivity
found in 10% to 90% of the neoplastic cells. PR was negative
in 1 case and positive in the remaining 3 cases, with
percentages of positive cells ranging from 5% to 20%
[13,23]. As stated by Damiani et al , the number of
reported cases was too limited for any conclusion, and a study
of more cases was advocated.
In the present series, 32 cases of OC were selected on
morphologic grounds using the World Health Organization
criteria  and validated using immunohistochemistry
(ie, ≥70% of neoplastic cells strongly positive for antimi-
tochondrion antiserum, consistent with the scoring system
proposed by Damiani et al ). To minimize possible cross-
reactivity with different substrates found in mitochondria
(eg, biotin), a polymer rather than avidin-biotin-peroxidase
complex was used. Optimal antibody dilution was retained
when lymphocyte-immunostained cytoplasm of a case of
Warthin tumor had faded, whereas the glandular pink
epithelium, notoriously constituted by oncocytes, remained
intensely positive. Oncocytic differentiation was further
confirmed by EM in 6 cases that showed cells with numerous
enlarged mitochondria scattered throughout the cytoplasm
without evident polarization. In addition, no secretory
granules were evident.
Three groups of tumors were highlighted using the antimi-
tochondrion antibody, namely, oncocytic, mt-rich, and IBC.
OCs showed some different features from IBC, all statis-
tically significant: (1) solid pattern of growth, with pushing
margins (P = .0004); (2) neoplastic cells characterized by
large granular and strongly eosinophilic cytoplasm (P b
.0001), and well-defined borders (P = .0006); (3) nuclei
mostly pleomorphic with prominent nucleoli (P b .0001);
and (4) mild to moderate lymphoid stroma (P b .0001). In
OC (n = 32)mt-rich (n = 25) IBC (n = 47) OC vs mt-rich OC vs IBCmt-rich vs IBC
P = .7579
P = 1.0000
P = .5203
P = .5203
P = 1.0000
P = .5762
P = .1880
P = .1412
P = 1.0000
P = 1.0000
P = .2164
P = 1.000
P = .0124⁎
P = .2654
P = 1.0000
P = .0720
P = .0571
P = 1.0000
P = .5531
P = .1811
P = .5530
P = .1734
P = .6591
P = .2750
P = 1.0000
P = 1.0000
P = 1.0000
Abbreviation: CHR, chromogranin.
⁎Statistically significant, P b .05.
Numerous large mitochondria are irregularly scattered in the cyto-
plasm. This is also seen in the cell illustrated in the inset.
Ultrastructural features of a representative case of OC.
trend for shorter survival in patients with OCs, although the dif-
ference is not statistically significant.
Survival data in the consecutive cases group. There is a
173Oncocytic carcinoma of the breast
general, OCs showed “luminal” phenotype with expression
of CK7, EMA, and consistent positivity for hormone recep-
tors. There was only a triple-negative case that was also
CK14 immunoreactive. This latter case was otherwise
indistinguishable from the other OCs. Thirteen cases showed
coexpression of apocrine or neuroendocrine differentiation.
HER2/Neu 3+ expression was more frequent and statistically
associated (P = .0124) with OCs when these tumors were
compared with IBC. No relevant differences were observed
with the remaining antibodies. OCs also displayed charac-
teristic EM features, very similar to those previously reported
by Damiani et al .
Tumors classified as “mitochondrion-rich” often shared
the morphologic features of OCs, and they were distin-
guished purely on immunohistochemical grounds. mt-rich
cells displayed weaker cytoplasmic eosinophilia and less
conspicuous nucleoli in comparison with oncocytic cells,
and these were the only statistically significant differences
(P = .0352 and P = .0176, respectively).
Our analysis of a series of consecutively accrued invasive
breast cancers revealed that 19.7% of the lesions were OCs,
suggesting that their prevalence may be substantially greater
than currently reported .
Abundant eosinophilic cytoplasm in breast carcinoma
cells is usually interpreted as a sign of apocrine or neuro-
endocrine differentiation. In our series of consecutive cases
(the consecutive cases group), abundant mitochondria,
verified by immunohistochemistry, were actually more
frequent than the expression of apocrine markers (eg,
GCDFP-15) and substantially more frequent than that of
neuroendocrine markers (eg, chromogranin). In addition,
both apocrine and neuroendocrine differentiation often
coexisted with an mt-rich or oncocytic phenotype.
The remarkable indolent course of the 4 cases initially
described as oncocytomas of the breast  has led to the
belief that these tumors are of low-grade malignancy. This
is probably due to the fact that all the lesions were well-
In the present study, the overall survival of patients with
OCs was similar to that of patients with IBCs, with a trend for
decreased survival in OCs that possibly reflects the signifi-
cantly higher proportion of grade 3 and positive cases of
HER2/Neu 3+. Similar results were obtained when oncocytic
tumors were compared with grade- and size-matched IBCs.
In conclusion, the present data indicate that OC of the
breast is the most frequent special type mammary carcinoma.
Therefore, carcinomas showing oncocytic features have to be
added to the list of tumor types that manifest with cytoplasmic
granular eosinophilia such as neuroendocrine and apocrine
carcinomas from which they have to be distinguished.
At the present, OC is a morphologic entity, and there is
no practical reason for including OC into everyday reports,
since no significant clinical features were found. However,
more cases of OC of the breast have to be studied in view
of the fact that oncocytic tumors of colon , thyroid ,
and meninges  seem to be resistant to radiation therapy,
a finding that might be taken in consideration when plan-
ning the treatment in breast carcinomas. Preliminary data
indicate that oncocytic and mt-rich carcinomas have distinct
patterns of genetic aberrations when compared with grade
and ER-matched invasive ductal carcinomas of no special
type . This suggests that accumulation of mitochondria,
resulting in either oncocytic or mt-rich tumors, may be a
marker of a subtype of breast cancer with distinctive mole-
In conclusion, it appears that:
• oncocytic breast carcinoma is a morphologic entity
with distinctive histologic and ultrastructural features
compared with IBC;
• oncocytic and mt-rich carcinomas share similar mor-
phologic features, and OC constitute about 19.7% of all
invasive carcinoma, indicating that cytoplasmic eosin-
ophilia in breast cancer cells is often due to accumu-
lation of mitochondria;
• OC and mt-rich carcinomas are more often grade III
tumors (HER2 overexpression is more frequent in OC
and mt-rich carcinomas than in IBCs); and
• clinical features are not distinctive, and there are no
statistically significant differences regarding overall
survival compared with IBC, even when matched for
grade and stage.
The authors wish to thank Drs Eugenio Marinaro and Jeff
Brown for having contributed 2 cases and provided clinical
information on them. Dr Dorotea Orsoni from Ospedale di
Budrio, the Medical Direction of Ospedale di Bazzano, and
Dr Vincenza Perlangeli of the tumor registry of AUSL
(Azienda Unità Sanitaria Locale) of Bologna have helped us
with the FU data.
 Hamperl H. Beiträge zur normalen und pathologischen Histologie
menschlieher Speicheldriisen. Z Mikrosk Anat Forsch 1931;27:1-55.
 Askanazy M. Pathologisch-anatomische Beiträge zur Kenntniss des
morbus basedowii, insbesondere uber die dabei auftretende Muske-
lerkrankkung. Deutsch Arch Klin Med 1898;61:118-86.
 Hurthle K. Beitrage zur kenntnis des sekretionsvorganges in der
schilddruse. Arch F D Ges Physiol 1894;56:1-44.
 Asa SL. My approach to oncocytic tumours of the thyroid. J Clin
 Roth SI, Olen E, Hansen LS. The eosinophilic cells of the parathyroid
Light and electron microscopic observations. Lab Invest 1962;11:
 Tremblay G, Pearse AG. Histochemistry of oxidative enzyme systems
in the human thyroid, with special reference to Askanazy cells. J Pathol
 Ghadially FN. Diagnostic electron microscopy of tumors. 2nd ed.
London: Butterworth & Company; 1985.
174 M. Ragazzi et al.
 Roncaroli F, Riccioni L, Cerati M, et al. Oncocytic meningioma. Am J Download full-text
Surg Pathol 1997;21:375-82.
 Giangaspero F, Cenacchi G. Oncocytic and granular cell neoplasms of
the central nervous system and pituitary gland. Semin Diagn Pathol
 Polk P, Parker KM, Biggs PJ. Soft tissue oncocytoma. Hum Pathol
 Damiani S, Pasquinelli G, Eusebi V. GANT-like gastrointestinal
pacemaker cell tumours with oncocytic features. Virchows Arch
 Tallini G. Oncocytic tumours. Virchows Arch 1998;433:5-12.
 Damiani S, Eusebi V,Losi L, D'Adda T,Rosai J.Oncocytic carcinoma
(malignant oncocytoma) of the breast. Am J Surg Pathol 1998;22:
 Bronner MP, Li Volsi VA. Oxphilic (Askanazy/Hürtle cell) tumors of
the thyroid: microscopic features predict biologic behavior. Surg
 Tallini G, Carcangiu ML, Rosai J. Oncocytic neoplasms of the thyroid
gland. Acta Pathol Jpn 1992;42:305-15.
 Murphy WM, Grignon DJ, Perlman EJ. Tumors of the kidney, bladder
and related urinary structure. Washington. DC: American Register of
 Rosai J. Rosai and Ackerman's surgical pathology. 9th ed. New York:
 DeLellis RA, Lloyd RV, Heitz PU, Eng C, WHO. Pathology and
genetics of tumours of endocrine organs. Lyon: IARC; 2004.
 Ellis GL, Auclair PL. Tumors of the salivary glands. AFIP atlas of
tumor pathology, 4th series. Silver Spring MD: ARP Press; 2008.
 Breen P, Flickinger JC, Kondziolka D, Martinez AJ. Radiotherapy for
nonfunctional pituitary adenoma: analysis of long-term tumor control.
J Neurosurg 1998;89:933-8.
 Nishioka H, Ito H, Hirano A, Shibata N, Kobayashi M. Immunocyto-
chemical study of pituitary oncocytic adenomas. Acta Neuropathol
 Marucci G, Betts CM, Frank G, Foschini MP. Oncocytic meningioma:
report of a case with progression after radiosurgery. Int J Surg Pathol
 Costa MJ, Silverberg SG. Oncocytic carcinoma of the male breast.
Arch Pathol Lab Med 1989;113:1396-9.
 Tavassoli FA, Devilee P, WHO. Pathology and genetics of tumours of
the breast and female genital organs. Lyon: IARC Press; 2003.
 Tavassoli FA, Eusebi V. Tumors of mammary gland. Washington,
DC: AFIP Atlas of Tumor Pathology; 2009.
 Elston CW, Ellis IO. Pathological prognostic factors in breast cancer.
I. The value of histological grade in breast cancer: experience from a
large study with long-term follow-up. Histopathology 2002;41:
 American Joint Committee on Cancer. Cancer Staging Manual 7th
edition. New York: Springer; 2010.
 Goldhirsch A, Glick JH, Gelber RD, Coates AS, Thurlimann B, Senn
HJ. Meetings highlights: international expert Consensus on the
Primary Therapy of Early Breast Cancer 2005. Ann Oncol 2005;16:
 Wolff AC, Hammond ME, Schwartz JN, et al. American Society of
Clinical Oncology/College of American Pathologists guideline
recommendations for human epidermal growth factor receptor 2
testing in breast cancer. Arch Pathol Lab Med 2007;131:18-43.
 Eusebi V, Damiani S, Ellis IO, Azzopardi JG, Rosai J. Breast tumor
resembling the tall cell variant of papillary thyroid carcinoma: report of
5 cases. Am J Surg Pathol 2003;27:1114-8.
 Ambrosini-Spaltro A, Salvi F, Betts CM, et al. Oncocytic modifica-
tions in rectal adenocarcinomas after radio and chemotherapy.
Virchows Arch 2006;448:442-8.
 Sobrinho-Simoes M, Maximo V, Castro IV, et al. Hurthle (oncocytic)
cell tumors of thyroid: etiopathogenesis, diagnosis and clinical
significance. Int J Surg Pathol 2005;13:29-35.
 Reis-Filho JS, de Biase D, Lambros MBK, et al.  Genomic
profiling of mitochondria-rich breast cancer. Boston: USCAP; 2009.
175 Oncocytic carcinoma of the breast