The Breast 23 (2014) 591e596

Contents lists available at ScienceDirect

The Breast
journal homepage: www.elsevier.com/brst

Original article

Is breast cancer from Sub Saharan Africa truly receptor poor?

Prevalence of ER/PR/HER2 in breast cancer from Kenya
Shahin Sayed a, *, Zahir Moloo a, 1, Ronald Wasike b, 2, Peter Bird c, 3, Raymond Oigara d, 4,
Dhirendra Govender e, 5, Joshua Kibera a, 6, Henri Carrara f, 7, Mansoor Saleh g, 8
a
Department of Pathology, Aga Khan University Hospital, P.O. Box 30270-00100 GPO, Nairobi, Kenya
b
Department of Surgery, Aga Khan University Hospital, P.O. Box 30270-00100 GPO, Nairobi, Kenya
c
Department of Surgery, African Inland Church, Kijabe Mission Hospital, P.O. Box 20, KIjabe 00220, Kenya
d
Department of Surgery, St. Mary's Mission Hospital, P.O. Box 3409, Nairobi, Kenya
e
Division of Anatomical Pathology, Faculty of Health Sciences, University of Cape Town and National Health Laboratory Service e Groote Schuur Hospital,
Falmouth Building, Medical School Anzio Road, Observatory, South Africa
f
Department of Public Health and Family Medicine, Faculty of Health Sciences, Medical School, University of Cape Town, Anzio Road,
Observatory, South Africa
g
University of Alabama at Birmingham Comprehensive Cancer Center, 703 South 19th Street, Birmingham, 3529 AL, USA

a r t i c l e i n f o a b s t r a c t

Article history: Objectives: Studies on ER/PR/HER2 in breast cancer from Sub Saharan Africa (SSA) are fraught with in-
Received 25 January 2014 consistencies in the prevalence of hormone receptor status. In Kenya, ER/PR/HER2 for breast cancers is
Received in revised form not part of routine assessment and available in only three to four centers across the country. Variability in
24 May 2014
methodology and interpretation makes comparison between data difficult. Our aim was to accurately
Accepted 5 June 2014
determine the prevalence of ER/PR/HER2 using standardized techniques and double reporting. Prog-
Available online 7 July 2014
nostic tumor parameters were also correlated with clinical features and receptor status.
Materials and methods: Consecutive invasive breast cancers (IBC) accrued between September 2011 and
Keywords:
Breast cancer
December 2012 were analyzed at Aga Khan University Hospital, Nairobi (AKUHN). Tumor blocks were
Sub Saharan Africa stained for ER/PR/HER2 on an automated platform. Double reporting of ER/PR/HER2 was done using the
Hormone receptors Allred system and the ASCO/CAP guidelines respectively.
Results: A total of 301 cases of IBC were analyzed for pathology and ER/PR/HER2. The age range of pa-
tients was 19e94 years with a median of 47.5 years. Invasive ductal carcinoma (NOS) was the most
common histologic type (84.2%). ER positivity was seen in 72.8%, PR in 64.8% and HER2 in 17.6% of all
cases. Triple negative breast cancers (TNBC) constituted 20.2% of the cases. There was a significant as-
sociation between receptor status and histologic grade (p < 0.001) and statistically significant trend of
increasing pathological stage of tumor (pT) associated with TNBC (p ¼ 0.020).
Conclusions: We present a definitive prospective analysis of ER/PR/HER2 from a single center and
demonstrate that prevalence of receptor status from SSA is comparable with that in the West.
© 2014 Elsevier Ltd. All rights reserved.

Introduction

* Corresponding author. Tel.: þ254 (0) 733 719 045.

E-mail addresses: shaheen.sayed@aku.edu (S. Sayed), Zahir.moloo@aku.edu
According to the latest data estimates [1], breast cancer is the
(Z. Moloo), Ronald.wasike@aku.edu (R. Wasike), petersuebird@gmail.com (P. Bird), second most common malignancy in Kenya, with the incidence
oigara@yahoo.com (R. Oigara), Dhiren.govender@uct.ac.za (D. Govender), Joshua. almost doubling in 2012 compared to 2008 (4465 vs. 2260 cases
chege@aku.edu (J. Kibera), Henri.carrara@gmail.com (H. Carrara), mns@uab.edu respectively).
(M. Saleh).
1 The incidence of breast cancer in Uganda has also doubled in
Tel.: þ254 (0) 731 337 099.
2
Tel.: þ254 (0) 722 747 601. three decades (11/100,000 to 22/100,000 as of 1995) [2]. A similar
3
Tel.: þ254 (0) 735 410 635. increase in incidence has been reported from Nigeria 15.3/100,000
4
Tel.: þ254 (0)725 400 081. in 1976 to 33.6/100,000 in 1992 [3]. Published data from Zambia
5
Tel.: þ27 (0) 214 066 162. show that at 8.6%, breast cancer was the third most common ma-
6
Tel.: þ254 (0) 733 719 045.
7 lignancy in females after cancers of the cervix and conjunctiva
Tel.: þ27 (0) 214 066 724.
8
Tel.: þ1 (404) 272 7222. (41.5% and 9% respectively) [4].

http://dx.doi.org/10.1016/j.breast.2014.06.006
0960-9776/© 2014 Elsevier Ltd. All rights reserved.
592 S. Sayed et al. / The Breast 23 (2014) 591e596
In their thought provoking commentary, Akarolo-Anthony et al.
review the evidence for a rising incidence of breast cancer and call
it the “emerging epidemic” in Africa [5]. Despite this reported in-
crease and the importance of key prognostic breast cancer markers,
diagnostic breast pathology and hormone receptor analysis is yet to
be offered as part of routine service in most institutions in SSA.
Thus, treatment at most centers remains empiric.
Limited local expertise in IHC and the perceived expense of
testing for hormone receptors are some of the contributing factors.
Scarce local histopathology services mean that specimens are
transported to distant central laboratories in ill-fitting containers
with insufficient formalin, resulting in extensive autolytic changes,
and thus precluding accurate histologic and IHC analysis. In addition,
the recommended CAP guidelines for histopathology reporting are
not followed in some centers, hence vital prognostic histologic de-
tails are often missing from reports. Multi-disciplinary tumor boards
do not exist at many institutions, limiting interaction between the
clinician and the pathologist, which impacts on optimal patient care.
Most studies published from SSA are small and predominantly
retrospective. Furthermore, appreciation of the importance of
prognostic features and hormone receptor status appears to have
been grossly overlooked. For example, the Uganda revised 2008
Breast cancer guideline is designed to be a reference document for
health workers involved in breast cancer care, yet only provides an
abbreviated outline of pathology reporting and makes only a
cursory mention of hormone receptors [2].
In the most recent, retrospective review from South Africa [6],
electronic data of 1216 patients with IBC from 1st Oct 2006 to 4th
July 2012 was retrieved and reviewed for age, sex, race, tumor size,
tumor stage, LN status, receptor status, tumor grade and HIV status.
Data for ER was available in 1072 cases of which 64.9% were re-
ported as ER positive. Data for HER2 status was available in 1029
cases of which 26% were HER2 positive. Contrary to conventional
guidelines, all equivocal HER2 results (HER2 2þ score) were
considered to be HER2 negative.
Data from West Africa have shown conflicting ER/PR receptor
status with ER positive tumor rates ranging between 25% and 70%
[7e9]. The few published studies from East Africa that have
analyzed receptor status did not utilize standard methodology and
interpretation according to ASCO/CAP guidelines. For example, in
the largest series to date from Kenya, Nyagol et al. [10] describe a
10% cut off point for ER and PR and a 10% cut off for HER2 with a
FISH ratio of >2 considered to be positive for HER2. In this partic-
ular series, in which 120 consecutive breast cancers were analyzed
for receptors, ER positive tumors comprised 37.3% and HER2 status
was reportedly positive in only 12.66% of the cases.
A retrospective study of 101 consecutive mastectomy cases from
our centre [11] reported an ER positive rate of 46% and a HER2
positive rate of 23%. Once again a 10% cut off point was used for both
ER and HER2. Other studies from Kenya and Nigeria report
comparatively lower rates of ER positive tumors at approximately
25% [12,13]. Smaller data series from Uganda [14] show a positive ER
rate of 51% (18/35 cases) and HER2 positive rate of 11% (5/35 cases).
All HER2 equivocal were considered negative for the receptor. As
such, most publications have not used standard reporting criteria,
thus contributing to the highly variable results noted in the literature.
We thus embarked on a study to determine, as accurately as
possible, the prevalence of ER/PR/HER2 at AKUHN, using stan-
dardized methodology and where possible to correlate the receptor
status with clinical and pathologic features.
Methods
The Research and Ethics Committee of AKUHN approved the
study. AKUHN is a 300-bed university hospital that provides
tertiary level health care services and pathology service to forty-
five satellite centers located in various parts of Kenya. The pathol-
ogy laboratory attained and adopted the ISO 15189 International
laboratory accreditation and subscribes to the UK-NEQAS for its
external QA program for ER/PR and HER2.
A total of 304 consecutive invasive breast cancers over a 13-
month period (September 1, 2011 to September 30, 2012) were
analyzed for histopathology prognostic parameters that included
grading, presence/absence of LVI, lymph node metastases and
extranodal extension. Grading was limited to mastectomies,
lumpectomies and a few wedge biopsies (where there was suffi-
cient tissue for analysis of the three parameters used for grading). A
total of 301 cases were analyzed for ER/PR/HER2 in-house.
Appropriate tumor blocks were selected and sectioned at 4 mm.
Heat induced Epitope retrieval (HIER) was performed for 20 min at
95  C on the Dako PT Link Pre-treatment System using the high pH
Target retrieval solution; pH9 diluted 1 in 50 as per manufacturer's
instructions.
The sections were stained for ER (FLEX RTU Monoclonal
Rabbit Anti-Human ER a, Clone EP1) PR, (FLEX RTU Monoclonal
Rabbit Anti-Human PgR 636 Antibody clone) and HER2/neu
(Polyclonal Rabbit Anti-Human c-erbB-2 Oncoprotein diluted
1:200 with the EnVision™ FLEX antibody diluent) on the Dako
Autostainer Link 48 platforms using the EnVision™ FLEX Kit
detection kits. Diaminobenzidine (DAB) was used as the
chromogen.
Interpretation of ER/PR was done according to the Allred
scoring system [15]. A composite score of 3 and above for in-
tensity and proportion of cells staining was considered to be
positive. HER2 scoring was done according to the ASCO/CAP
guidelines [16]. HER2 was scored positive if 30% or more of the
tumor cells showed complete and strong membrane positivity. A
score of 0 or 1þ were considered HER2 negative. Cases which
were scored as HER2 2þ (equivocal) were subjected to FISH to
assess gene amplification, at our reference laboratory. A FISH test
result was considered positive if ratio of the LSI HER2/neu/CEP17
signals was >2.2.
Normal breast tissue was used as an inbuilt positive control
for ER/PR. Additionally, a section of uterine cervix was used as an
on slide positive control: All the stromal, columnar epithelial
and squamous epithelial cells showed a moderate to strong
nuclear staining reaction. For HER2, an on slide control section
from a composite tissue block of HER2 score of 1þ, 2þ and 3þ
was used.
We ensured that only morphologically well-fixed tumor blocks
were selected for receptor analysis. In addition, our major outreach
centers are supplied with 10%Neutral Buffered Formalin (NBF) so
that specimens coming from distant health facilities are immedi-
ately immersed in sufficient amounts of fixative.
To minimize errors in interpretation, double reporting of all
histologic parameters that included grading, presence/absence
of LVI, lymph node metastases and extranodal extension and the
interpretation of ER/PR/HER2 was undertaken by SS and ZM.
Discrepant results were discussed to arrive at a consensus. It
was noted that ER/PR results tended to have a bimodal staining
pattern (strong positive or negative) and hence there was 100%
concordance between the two pathologists. For HER2, wherever
there was a discrepancy between a HER2 equivocal result and a
positive score, caution was exercised and the HER2 result re-
ported as equivocal and subjected to FISH for confirmation.
Statistical analysis
Continuous numeric data were examined for normality using
histograms and the Shapiro Wilk test, if found to be skewed the
 S. Sayed et al. / The Breast 23 (2014) 591e596 593

medians and ranges were presented. Categorical data were sum- Table 1
marized as proportions. Chi-squared or Fisher's exact test, as Clinical characteristics of study population.

appropriate, were used to examine the association between re- n

ceptors and key demographic variables. Risk factors for triple Age (median, range) (N ¼ 304) 304 47.5 (19e94)
negative versus ERþ and/or PRþ receptor status was analyzed us- Age category
ing generalized linear modeling to estimate the Prevalence Ratios <50 164 54.0
(PR) and associated 95% CIs. All analyses were performed using 50þ 140 46.0
Sex (N ¼ 304)
Stata version 13.1 (Stata Corp LP, 4905 Lakeway Drive, College
Female 292 96.0
Station, TX 77845, USA). Male 12 4.0
Menopausal status (N ¼ 109)
Results Pre 68 62.4
Post 41 37.6
Clinical stage (N ¼ 99)
A total of 304 consecutive cases were accrued during the 13 I 12 12.1
month study period and analyzed for pathologic prognostic fea- II 26 26.3
tures and 301 of these cases tested for receptor status (ER/PR and III 43 43.4
HER2). Male breast cancer constituted 4% of the total cases and IV 18 18.2
Procedure (%) (N ¼ 304)
these too were included in the analysis.
Mastectomy 162 53.3
Age range of patients was 19e94 years with a median age of 47.5 Core biopsy 65 21.4
years. The tumors were equally distributed between women less Lumpectomy 62 20.4
than and more than 50 years of age. Other (wedge biopsy, incisional biopsy) 15 4.9
Menopausal status was only available for 109 patients, of whom Laterality (%) (N ¼ 304)
Bilateral 1 0.3
the majority (62.4%) were premenopausal. Left 135 44.4
Clinical stage was available for 99 patients of whom the majority Right 112 36.8
(61.6%) were Stage 3 (43.4%) and Stage 4 (18.2%) combined. Only Not specified 56 18.4
12.1% of patients presented in Stage 1.
Of the total specimens received, 53.3% were mastectomies and
41.8% were core biopsies and lumpectomies and the remainder tumors, but this elevated risk was not statistically significant. There
were wedge resections/incisional biopsies. Both the right and left was a statistically non-significant 20% higher chance of tumors
breast were equally involved with the central and upper outer more than 5 cm in greatest dimension being TNBC when compared
quadrants more commonly affected (Table 1). to tumors less than 2 cm.
The median tumor size was 4 cm (range 1 mme18 cm). Of the Although not statistically significant, TNBC tumors were 2.2
224 tumors, 65 tumors (29.1%) were more than 5 cm in greatest times more likely to be grade 3 rather than grade 1 when compared
dimension. with ER and/or PR positive tumors. The presence of LVI and lymph
Invasive ductal carcinoma (NOS) was the most common histo- node metastases was less likely in TNBC when compared to ER and/
logic type at 84.2% followed distantly by invasive lobular, invasive or PR positive tumors, although this too was not statistically
mucinous and metaplastic carcinomas at 2.6% each. Grade 3 tumors significant.
constituted 53.7% of the cases. Although none of the prognostic factors for pT status was indi-
Lymphovascular invasion was present in 70.4% of the 294 cases vidually statistically significant, there was a significant trend
in which it was unequivocally possible to determine the same. (p ¼ 0.02), between pathologic tumor stage (pT) and TNBCs; the
Of the 150 cases that either had sentinel node examination, proportion of TNBC tumors increased with advancing stage, stage 3
axillary sampling or axillary dissection, the number of nodes (32.5%) and stage 4 (45.4%) (Table 4).
ranged from 1 to 31 with a median of 9 nodes for axillary dissection
specimens. In the 85 cases with nodal involvement, 95.3% showed Discussion
extranodal tumor extension.
Of the 301 cases on which ER/PR was performed, ER positivity Our study, which is the largest reported prospective single
was seen in 72.8% of cases and PR positivity was at 64.8%. Luminal a center series of ER/PR/HER2 testing from East Africa conducted
tumors (ERþ/ PR positive and HER2 negative) constituted 61.2% of locally, and also the only series from SSA, where double reporting of
all ER and/or PR positive cases and luminal B (ERþ/ PR positive histologic parameters and ER/PR/HER2 has been done, raises
and HER2 positive) made up 10.8% of cases. questions whether breast cancer from SSA are truly hormone re-
In the 301 cases on which HER2 was performed, 47 cases were ceptor poor.
HER2 3þ positive and 28 were equivocal (2þ). Of the equivocal Our findings of 72.4% ER positive and 64% PR positive rates is
cases only 13 had enough tissue available for further testing by similar to the prospective study from Nigeria [9] where an ER
FISH. Of these 13 cases, 6 were positive for HER2, giving a total positive rate of 70.6% and a PR positive rate of 64% was reported.
HER2 positive number of 53 cases (17.6%). Triple negative breast However, in that study interpretation of ER/PR did not follow ASCO/
cancers (TNBC) constituted 20.2% of the total cases (58 out of 286) CAP guidelines [17], and it is notable that a 10% cut off was used for
(Table 2). a positive ER/PR score thus raising the possibility of an underesti-
There was no association between patient age, presence of mation of the ER positive rate. There was no association between
lymphovascular invasion and the receptor status; however, there age and ER/PR/HER2 status and no statistical difference in ER status
was a significant association between receptor status and histologic between patients who were <50 and more than 50 years, similar to
grade. (pT) (p < 0.001) but no statistical significance between re- our findings. The authors attributed their findings to the de-
ceptor status and pathological stage of tumor (p ¼ 0.08) (Table 3). ficiencies in the pre-analytic phase of specimen handling which
Further analysis of the subset of data including the TNBC and the contribute to falsely elevated receptor poor breast cancers in pa-
ER and/or PR positive breast cancers is shown in (Table 4). Patients tients of African ancestry [9]. This could account for the data re-
with triple negative breast cancers were 1.4 times more likely to be ported by Gukas et al. of an ER/PR positive rate of only 27% (10 out
older than 50 years when compared to ER and/or PR positive of 36 cases) [13].
594 S. Sayed et al. / The Breast 23 (2014) 591e596

Table 2 Table 4
Pathologic characteristics of study population. Prognostic factors for triple negative versus ER and/or PRþ tumors.

n % N % TNBC PR (95%CI)
Age category (in years)
Morphologic type (N ¼ 304)
<50 116 20.7 1.0 (Referent)
Invasive ductal carcinoma, nos 255 84.2
50þ 117 29.1 1.4 (0.9e2.2)
Invasive lobular carcinoma 8 2.6
Size category (cm)
Invasive micropapillary carcinoma 7 2.3
<2 20 30.0 1.0 (referent)
Invasive mucinous carcinoma 8 2.6
2e5 105 18.1 0.6 (0.3e1.3)
Invasive papillary carcinoma 5 1.6
>5 54 37.0 1.2 (0.6e2.6)
Metaplastic carcinoma 8 2.6
Grade
Other (adenoid cystic, apocrine, 13 4.3
1 15 20.0 1.0 (referent)
lymphoepithelioma-like, microinvasive)
2 78 5.1 0.3 (0.1e1.03)
Grade (N ¼ 229)
3 86 44.0 2.2 (0.8e6.3)
I 16 7.0
LVI
II 90 39.3
Absent 71 32.4 1.0 (referent)
III 123 53.7
Present 154 20.8 0.6 (0.4e1.01)
Tumor size in cm (median, range) 224 4.0 (0.1e18.0)
Extranodal extension
<2 24 10.7
No 4 0.0 1.0 (referent)
2e5 135 60.3
Yes 54 25.9 ∞
>5 65 29.0
pT
LVI (N ¼ 294)
1 21 23.8 1.0 (referent)
Present 207 70.4
2 91 18.7 0.8 (0.3e1.9) Trend p ¼ 0.02
Extranodal extension (N ¼ 85)
3 40 32.5 1.4 (0.6e3.3)
Present 81 95.3
4 22 45.4 1.9 (0.8e4.7)
ER (N ¼ 301)
Positive 219 72.8
PR (N ¼ 301)
Positive 195 64.8 studies. It is well established that about 20% of HER2 equivocal
HER2 (N ¼ 301)
Negative (0 and 1þ) 233 77.4
results are gene amplified and thus eligible for Trastuzumab ther-
Positive (3þ) 53 17.6 apy [18]. We should note that of the 28 samples that were HER2
Equivocal (2þ) 15 5.0 equivocal by IHC (2þ) in our study, only 13 had sufficient tissue for
Receptor combinations (N ¼ 286) FISH analysis and 6/13 (46%) were deemed FISH amplified e a
ERþ and/or PRþ and HER2 negative 175 61.2
higher conversion rate than described in the literature [18]. How-
Triple negative 58 20.2
ERþ and/or PRþ and HER2 positive 31 10.8 ever, given the small number of residual HER2 equivocal tumors in
HER2þ 22 7.7 our overall sample size (5%), this would not significantly alter our
pT stage (N ¼ 224) HER2 receptor prevalence results.
1 26 11.6 Although our subset analysis based on receptor status only
2 112 50.0
3 50 22.3
showed a statistically significant trend in pathologic tumor stage
4 36 16.1 and TNBC, it is clinically significant that TNBC tumors tended to be
larger, of higher grade and less likely to demonstrate LVI and lymph
Interestingly, in both the Nigerian study [9], and the electronic node metastases when compared with ER and/or PR positive tu-
data review from South Africa [6], all HER2 equivocal scores were mors. The small size of the subset could account for the lack of
considered to be negative which is contrary to guideline recom- statistical significance.
mendations, thus affecting the TNBC rate reported from the two As suggested in our earlier publication, the previously reported
higher proportion of observed TNBC and basal-like disease from
Table 3 SSA could be explained by a lower life expectancy and selection
Association of receptor classification with clinicopathologic features. bias. Those women with rapidly progressive tumors may be seeking
ERþ and/or PRþ ERþ and/or PRþ TNBC HER2 þve p-Value medical attention and thus undergo biopsy and pathologic analysis,
HER2 ve and HER2þ compared to women with hormone receptor positive tumors
n % n % n % n % whose tumors tend to be less aggressive [19]. We suggest that this
paradigm is likely to shift as access to health services in SSA im-
Age category (N ¼ 263)
<50 92 52.6 6 75.0 24 41.4 12 54.6 proves and Ministries of Health like those in Kenya have established
50þ 83 47.4 2 25.0 34 58.6 10 45.4 0.2 departments of Non Communicable diseases that emphasize the
Size category (N ¼ 199) role of screening and early diagnosis in cancer.
<2 cm 14 10.4 0 0.0 6 13.3 3 23.1
Citing two systematic studies [9,20] from West Africa, Akarolo-
2e5 cm 86 64.2 3 60.0 19 42.2 8 53.8
>5 cm 34 25.4 2 40.0 20 44.4 4 23.1 0.3
Anthony et al. in their review, also provide evidence against per-
Grade (N ¼ 202) ceptions that breast cancer in women of African descent is indeed
1 12 9.0 1 20.0 3 6.7 0 0.0 hormone receptor poor. The authors stress the importance of
2 74 52.2 0 0.0 4 8.9 3 16.7 <0.001 educating clinicians, and technologists in pre-analytic aspects of
3 48 35.8 4 80.0 38 84.4 15 83.3
specimen handling [5].
LVI (N ¼ 255)
Absent 48 28.271.8 1 12.5 23 41.8 6 27.3 Another important consideration accounting for the difference
Present 122 7 87.5 32 58.2 16 72.7 0.2 in receptor results from SSA is the level of IHC standardization and
Extranodal extension (N ¼ 68) whether instrumentation as opposed to the manual method of
Absent 4 9.1 0 0.0 0 0.0 0 0.0 analysis was used. It is noteworthy that in the recently published
Present 40 90.9 4 100.0 14 100.0 6 100.0 0.8
pT (N ¼ 196)
data from South Africa [6], an automated platform was used for ER/
1 16 12.3 0 0.0 5 11.1 3 18.8 PR/HER2 analysis, as was the case in our study. In those centers that
2 74 56.9 2 40.0 17 37.8 5 31.2 use the manual method of IHC, the rigor with which attention is
3 27 20.8 1 20.0 13 28.9 3 18.8 paid to the calibration of pipettes and pH meters, quality of water
4 12 9.2 2 40.0 10 22.2 5 31.2 0.08
used for buffer preparation, accurate dilution of concentrated
 S. Sayed et al. / The Breast 23 (2014) 591e596
antibodies, and proficiency testing of technical staff performing the
test could account for the variable ER/PR/HER2 rates from SSA.
Given this unreliability of hormone receptor rates in breast
cancer from Africa, one may question the need for testing for the
same, and propose that the practice of empiric therapy with
tamoxifen or aromatase inhibitors continue. However, as Bird et al.
argue [12], the economic savings from screening for ER receptor in
breast cancer is worth the initial expense in view of the fact that the
majority of the Kenyan rural population lives on less than 30 USD a
month. The initial testing would also identify patients who are
likely to respond to tamoxifen, avoiding toxicity and ineffective
hormone therapy for those who are receptor negative. To ensure
quality, the adoption of a hub and spoke model for ER/PR/HER2
testing would be the appropriate model in low resource settings.
While we are able to provide accurate pathologic and IHC
analysis, we lacked clinical information for a number of cases, thus
precluding multivariate analysis for some of the relevant variables.
The incomplete clinical data set for a number of the variables meant
less precise estimates for some of the subset analysis. This aspect
points to the vital importance of collecting both pathologic as well
as clinical data as part of a tumor registry.
The wide variation in reported rates of ER/PR/HER2 positivity
from SSA is clearly multifactorial. We suggest that large regional
and multicenter collaborative studies, emphasis on adoption of
standardization of test performance and interpretation and sub-
scription to quality assurance schemes would be a priority. When
following established tissue handling methodology and standard-
ized reporting criteria, it turns out that the prevalence of ER, PR and
Her2 receptors in patients in SSA may not be different from that
found in breast cancer patients in the West, where ER positive
breast cancers range from 70% to 78%,with higher rates in whites
and older patients and a slightly lower rate in nonwhite and
younger people [21,22]. The establishment of centers of excellence
for breast cancer care, national cancer registries, physician educa-
tion, multi-disciplinary tumor boards and pathologist involvement
in development of national breast cancer guidelines are some of the
issues that also need to be addressed to provide optimal care for
breast cancer patients. This would be an important step in raising
the standard of breast cancer care in SSA.
Conclusion
Our study provides a definitive prevalence rate for ER/PR/HER2
in SSA and it would appear that the prevalence rates in SSA are
comparable to those in the West. This has significant clinical as well
as pharmaco-economic implications and is a subject that deserves
attention on part of clinicians as well as health policy makers.
Ethical approval
The study was approved by the Research and Ethics Committee
of Aga Khan University Hospital, Nairobi.
Contributions
Drs Shahin Sayed and Mansoor Saleh designed the research
concept and study with input from Dr Zahir Moloo.
The manuscript was written by Drs Shahin Sayed and Mansoor
Saleh, with review and input by Drs Peter Bird, Zahir Moloo and
Prof. Dhirendra Govender.
Henri Carrara performed all the Statistical analysis and pre-
sentation of data in tables.
All the other authors enrolled patients on the study and
contributed to the data collection.
595
Conflict of interest statement
All authors declare that the answer to the questions on your
conflict of interest statement are all No and therefore have nothing
to declare.
Acknowledgments
The authors would like to acknowledge the contribution of the
following,
Funding of this study was enabled through the University
Research Council Grant of the Aga Khan University (111004EA). The
study sponsor had no role in the design, collection, analysis and
interpretation of data, writing of the manuscript or submission of
the manuscript for publication. This work was supported in part by
the 2011 GSK-ERI grant award to Dr. Shahin Sayed (PI).
Nasra Gathoni, librarian at AKUHN for her support in retrieving
articles for the literature review.
Innocent Abayo, Research Assistant, Department of Pathology,
AKUHN, for his assistance with the formatting of this manuscript.
Prof Zul Premji, Chief of Pathology, AKUHN for kindly reviewing
the manuscript. AKU, Karachi, for providing the FISH testing for
HER2 analysis.
References
[1] GLOBOCON. IARC 2013; 2012. http://www-dep.iarc.fr/.
[2] Gakwaya A, Galukande M, Luwaga A, Jombwe J, Fualal J, Baguna P, et al. Breast
cancer guidelines for Uganda (2nd edition 2008). Afr Health Sci 2008;8(2):
126e32.
[3] Ihekwaba FN. Breast cancer in Nigerian women. Br J Surg 1992;79:771e5.
[4] Bowa K, Wood C, Chao A, Chintu C, Mudenda V, Chikwenya M. A review of the
epidemiology of cancers at the University Teaching Hospital, Lusaka, Zambia.
Trop Doct 2009;39(1):5e7.
[5] Sally NA, Temidayo OO, Adebamowo CA. Emerging breast cancer epidemic:
evidence form Africa. Breast Cancer Res 2010;12(Suppl. 4):S8.
[6] Valarie AM, Maureen J, Eunice VB. Breast cancer receptor status diagnosis in
over 1200 consecutive public hospital patients in Soweto South Africa: a cases
series. Breast Cancer Res 2013;15. R84:1e13.
[7] Ohene YM, Adjei E. Breast cancer in Kumasi, Ghana. Ghana Med J 2012;46(1):
8e13.
[8] Stark A, Celina GK, Iman M, Baffour A, Anthony NA, Valarie T, et al. African
ancestry and higher prevalence of triple negative breast cancer. Cancer
2010;116(21):4926e32.
[9] Adebamowo CA, Famooto A, Ogundian TO, Aniagwu T, Nkwodiammah C,
Akang EE. Immunohistochemical and molecular subtypes of breast cancer in
Nigeria. Breast Cancer Res Treat 2008;110(1):183e8.
[10] Nyagol J, Nyong'o A, Byakika B, Muchiri L, Cocco M, Spina D, et al. Routine
assessment of hormonal receptors and her-2/neu status underscores the need
for more therapeutic targets in Kenyan women with breast cancer. Anal Quant
Cytol Histol 2006 Apr;28(2):97e103.
[11] Gakinya SM, Sayed S, Chauhan R. Breast cancer molecular subtypes and their
clinicopathological characteristics amongst patients at the Aga Khan Univer-
sity Hospital (Nairobi). Ann Afr Surg 2010;5:19e24.
[12] Bird PA, Hill GA, Housami N. Poor hormone receptor expression in East African
breast cancer: evidence of a biologically different disease? Ann Surg Oncol
2008;15(7):1983e8.
[13] Gukas ID, Jennings BA, Leinster SJ, Mandong BM, Igun GO, Girling AC, et al.
Clinicopathological features and molecular markers of breast cancer in Jos,
Nigeria. West Afr J Med 2005;24(3):209e13.
[14] Indrojit R, Othieno E. Breast cancer in Uganda microscopic study and receptor
profile of 45 cases. Arch Pathol Lab Med 2011 Feb;135(2):194e9.
[15] Allred DC, Harvey JM, Bernado M, Clark B. Prognostic and predictive factors in
breast cancer by immunohistochemical analysis. Mod Pathol 1998;11:
155e68.
[16] Wolff AC, Hammond ME, Hicks DG, Dowsett M, McShane LM, Allison KH, et al.
Recommendations for human epidermal growth factor receptor 2 testing in
breast cancer: American Society of Clinical Oncology/College of American
Pathologists clinical practice guideline update. J Clin Oncol 2013 Nov
1;31(31):3997e4013.
[17] Hammond MEH, Hayes DF, Dowsett M, Allred DC, Hagerty KL, Badve S, et al.
American Society of Clinical Oncology/College of American Pathologists
guideline recommendations for immunohistochemical testing of estrogen and
progesterone receptors in breast cancer. Arch Pathol Lab Med 2010;134:
907e22.
596 S. Sayed et al. / The Breast 23 (2014) 591e596
[18] Meijer SL, Wesseling J, Smit VT, Nederlof PM, Hooijer GKJ, Ruijter H, et al.
HER2 gene amplification in patients with breast cancer with equivocal IHC
results. J Clin Pathol 2011;64(12):1069e72.
[19] Mateya T, Sayed S, Mukono S, Moloo Z, Yasmin R. Basal-like breast cancer in
East Africa. Breast J 2011;17:438e40.
[20] Traore A, Traore C, Dembele BT, Kante L, Diakite I, Maiga A, et al. Cancer du
sein dans deux centres hospitaliers de Bamako (Mali): aspects diagnostiques
rapeutiques. J Afr Cancer 2010;2:88e91.
et the
[21] Anderson WF, Katki HA, Rosenberg PS. Incidence of breast cancer in the United
States: current and future trends. J Natl Cancer Inst 2011;103:1397e402.
[22] Parise CA, Bauer KR, Brown MM, Caggiano V. Breast cancer subtypes as defined
by the estrogen receptor (ER), progesterone receptor (PR), and the human
epidermal growth factor receptor 2 (HER2) among women with invasive breast
cancer in California, 1999e2004. Breast J 2009;15(6):593e602.