Proliferation in African Breast Cancer: Biology and
Prognostication in Nigerian Breast Cancer Material
O. F. Ikpatt, M.D., T. Kuopio, M.D., Y. Collan, M.D.
Department of Pathology (OFI, University of Calabar Teaching Hospital, Nigeria; Department of
Pathology, Jyväskylä Central Hospital (TK), Finland; and Department of Pathology, University of Turku
Three hundred cases of invasive breast carcinoma
from the University of Calabar Teaching Hospital,
Nigeria were subjected to evaluation of proliferative
activity by mitotic counts. The prognostic signifi-
cance and association with other prognostic factors
were evaluated. The mitotic activity was expressed
as mitotic activity index (MAI), and standardized
mitotic index (SMI). Pearson’s correlation and uni-
The mean follow-up time was 25.9 months. The
mean values of SMI and MAI were 42.6 mitotic fig-
ures per square millimeter and 30.5 mitotic figures
per 10 high-power fields, respectively, and these
were much higher than values reported for Europe
or other Western countries. The SMI had a positive
correlation with tumor size (r ? 0.31, P < .0001),
histologic grade (r ? 0.68, P < .0001), nuclear
area (r ? 0.45, P < .0001), and negative correlation
with fraction of fields with tubular differentiation
(FTD; r ? ?0.56, P ? <0.0001). There was no sta-
tistically significant difference in the mitotic activity
between the postmenopausal and the premeno-
pausal patients. Also, lymph node–positive patients
had higher counts than did lymph node–negative
patients. Earlier determined grading associated de-
cision thresholds divided the patients into groups of
favorable and unfavorable prognosis. However, the
statistically optimal thresholds for Nigerian mate-
rial were different (32 and 92 mitotic figures per
square millimeter for SMI). Tumor size of 5 cm,
SMI, and MAI were independent prognostic factors.
Nigerian breast cancers are high-grade, high-stage,
and high-proliferating cancers occurring in a
younger population than those of the Western
countries. Proliferation is also more active. Evalua-
tion of SMI or MAI can improve the distinction
between aggressive and less aggressive variants of
KEY WORDS: Africa, Breast cancer, Mitotic activity,
Mitotic index, Nigeria, Prognostication, Prolifera-
Mod Pathol 2002;15(8):783–789
Earlier studies suggest epidemiological differences
between breast cancers in the Caucasian (Europe-
an) and African populations (1–3). In a recent clin-
icopathological comparison of breast cancer in Ni-
especially caused by high mitotic rates, character-
ized a significant proportion of the Nigerian breast
cancers, suggesting a more aggressive nature (4).
The proliferation markers are strong and reproduc-
ible prognosticators in invasive breast cancer (5–7).
To study the situation further, this study assesses
the proliferative activity of breast cancer in Nige-
rian breast cancers using the mitotic activity index
(MAI; 8) and volume fraction–corrected standard-
ized mitotic index (SMI; 9). Attempts are made at
evaluating the relevant thresholds for a grading sys-
tem that could be applied in Nigeria. The correla-
tion of mitotic indices with other known prognostic
factors and the importance of mitotic counts as
prognosticators in different patient fractions are
MATERIALS AND METHODS
A previously described group (4) of 300 patients
with histologically confirmed invasive breast can-
cers diagnosed at the University of Calabar Teach-
ing Hospital, Nigeria, between 1983 and 1999 was
examined. Patients with intraductal carcinoma
were excluded. Those with either a recurrent or
bilateral tumor were counted as one. Metastases
were detected clinically with the assistance of ra-
Copyright © 2002 by The United States and Canadian Academy of
VOL. 15, NO. 8, P. 783, 2002 Printed in the U.S.A.
Date of acceptance: April 19, 2002.
Address reprint requests to: O.F.R. Ikpatt, M.D., Department of Pathology,
University of Turku, Kiinamyllynkatu 10, Turku 20520, Finland; e-mail:
email@example.com; fax: 358-2-2613965.
diological, laboratory, and histological examina-
tions as appropriate. None of the patients had pre-
operative radiotherapy or any other form of
preoperative adjuvant treatment.
Follow-up information was obtained from the
hospital medical records. Follow-up history was
available for 129 patients. A surgical team with a
fairly uniform treatment protocol managed these
patients. The follow-up period ranged from 2.0 to
60.0 months (mean, 25.9; median, 24.0 mo). These
129 patients were, therefore, available for survival
All patients were treated with either simple or
radical mastectomy with axillary evacuation. Che-
motherapy was given after this to 85 (65.9%); 20
(15.5%) had endocrine therapy; and 19 (6.3%) had
Twenty-four (18.6%) had radiotherapy in addition
to the surgery.
The end points of the follow-up were the pres-
ence or absence of recurrence, metastasis recorded,
and survival status. This study concentrates on sur-
vival analysis, which keenly followed the pattern of
disease recurrence. Deaths from other causes than
cancer were recorded when information was avail-
able, and only cancer-associated survival was eval-
uated as an end point in the survival analysis.
The characteristics of the patients are shown in
Perioperative biopsy specimens were fixed in
buffered formalin (pH 7.0), embedded in paraffin,
cut at 5 ?m, and stained with hematoxylin and
eosin. The histologic typing (10) and grading (11)
were done by one of the authors (OFRI), and un-
clear cases were reviewed after discussion and mi-
croscopy with another (YC). The criteria used in
identifying mitotic figures were those described by
Baak and Oort (12). Mitotic figures were character-
ized by an absent nuclear membrane with clear,
hairy extensions of nuclear material (condensed
metaphase), in a plane (metaphase/anaphase), or
in separate chromosomal aggregates (anaphase/te-
lophase). The basic idea was that at least one chro-
mosomal end was seen in a mitosis. Two parallel,
clearly separate chromosome clumps were counted
as one mitotic figure. The cytoplasm of the mitotic
cells was often larger during mitosis than in the
Counting was carried out in the most cellular
region at tumor periphery, avoiding areas of necro-
sis, inflammation, calcification, and in situ carci-
noma. If several areas met these criteria, the area
with the highest number of mitotic figures, assessed
subjectively, was chosen.
The first author carried mitotic count after a
training period (13), using a standard laboratory
microscope (objective, 40?; numerical aperture,
0.75; field diameter, 420 ?m).
The number of mitotic figures in 10 consecutive
fields from the most cellular area of the sample was
The volume fraction–corrected mitotic index or
SMI gives the mitotic count as the number of mi-
totic figures by the area of the neoplastic tissue in
the microscopic fields. This is the number of mito-
ses in 10 consecutive fields corrected for the vol-
ume fraction and field size. In this method, the area
fraction (as estimate of volume fraction) of neoplas-
tic tissue in the microscopic field is evaluated si-
multaneously with the mitotic count (9):
SMI ? k??MI)/(?Vv),
where k ? 100/r2, r is the radius of the field, and MI
? number of mitotic figures in the studied field. Vv
is the volume fraction (estimated by the area frac-
tion, as a percentage) of malignant epithelium in
the studied field.
The Finnish material was studied earlier (7, 14),
with the same methodology, and the results of
these studies will be compared with findings from
the Nigerian material in our discussion.
The SAS statistical package (SAS System for Win-
dows, Release 6.12; SAS Institute Inc., Cary, NC)
was used for analysis of both whole materials and
the subgroups according to variables like age, axil-
lary lymph node status, and tumor size at the time
TABLE 1. Characteristics of the Studied Patients and
Their Breast Cancer (n ? 300)
Age at diagnosis (y)
Menopausal status, n (%)
No. of premenopausal patients
Axillary lymph node status, n (%)
No. of positive patients
No. of negative patients
Tumor size (cm)
Follow-up time (mo)
Alive after follow-up
Causes of death during follow-up (n)
784 Modern Pathology
Previously determined (14) SMI threshold of 17
mitoses per square millimeter and MAI threshold of
10 mitotic figures per 10 high-power fields were
used to compare the outcomes between the group
of patients showing mitotic counts above and be-
low the cut points using Kaplan-Meier curves and
the log-rank test.
To obtain decision thresholds that could be ex-
pected to be more representative for the Nigerian
material, diagrams of the ?2of log-rank tests were
used to show variation of statistical significance
associated with each tested cut point. Cut points
yielding the most obvious rise in statistical signifi-
cance were the best at separating good and poor
prognostic groups and could therefore be used as
thresholds for classification of patients based on
To evaluate the prognostic significance of the
mitotic counts, univariate and multivariate analy-
ses based on Cox’s regression were applied (15).
The ratios indicating relative risk (RR) and their
95% confidence intervals (95% CI) showed associa-
tions between different prognostic factors and
breast cancer survival.
The characteristics of the Nigerian breast cancer
material are shown in Table 1. Descriptive statistics
were performed on all 300 patients, whereas sur-
vival analysis was restricted to the 129 with a
The mean age at diagnosis of breast cancers
among Nigerian patients was 42.7 (12.1) years. A
large fraction of the patients were premenopausal
(74.3%). A large tumor size with a mean (SD) of 4.8
(2.4) cm and of high frequency of lymph node in-
volvement (78.7%) characterized the material. The
average follow-up time was 25.9 months, and a
survival rate of 71.3% was observed.
The mean values of SMI and MAI were 42.6 mi-
totic figures per square millimeter and 30.5 mitotic
figures per 10 high-power fields, respectively, for
the whole material of 300 patients. These mean
values were not statistically different from those
obtained for the 129 patients that were subse-
quently used for survival analysis (SMI, 45.5 ? 32.5,
P ? .3437; MAI, 33.2 ? 28.7, P ? .3288).
The MAI and SMI values were higher in the post-
menopausal patients than in the premenopausal
group, but the difference was not statistically signifi-
cant (MAI, P ? .1849; SMI, P ? .4098). On the other
hand, the difference in mitotic counts between the
lymph node–positive and -negative tumors was sig-
nificant (MAI, P ? .0038; SMI, P ? .0008). Lymph
node–positive tumors had higher values.
The proliferative indices in the whole material,
and in different subgroups defined by the meno-
pausal status, lymph node status, tumor size, clin-
ical stage, histologic grade, and type are shown in
TABLE 2. Average Mitotic Activity as Expressed by MAI and SMI in Nigerian Breast Cancer Cases
P SMI (SD)b
Lymph node (LN)
Tumor size (cm)
300 30.5 (25.1)42.6 (27.5)
?0.0001 30.3 (24.6)
MAI, mitotic activity index; SMI, standardized mitotic index.
aMitotic figures per 10 high-power fields.
bMitotic figures per square millimeter of neoplastic epithelium.
Biology and Prognostication in Nigerian Breast Cancer (O.F.R. Ikpatt et al.)785
Table 2. Higher values are seen in large tumors and
those of higher histologic grade. The difference in
the proliferative indices between invasive ductal
carcinoma and lobular carcinoma was statistically
significant (P ? .0001 for both MAI and SMI).
A positive correlation between the SMI and MAI
was observed (Pearson’s r ? 0.80, P ? .001). Other
statistically significant correlation coefficients be-
tween SMI and other variables were as follows:
apoptotic index, r ? 0.28; tumor size, r ? 0.31;
clinical stage, r ? 0.17; histologic grade, r ? 0.68;
fraction of tubular differentiation, r ? ?0.56; nu-
clear area, r ? 0.41; nuclear perimeter, r ? 0.42;
nuclear diameter, r ? 0.45; and standard deviation
of nuclear area, r ? 0.41.
Table 3 shows the results of the univariate sur-
vival analysis performed in the whole material and
subgroups of patients. Tumor size and nodal status
clearly predicted breast cancer death in the whole
SMI and MAI were significant predictors of sur-
vival in the overall material. These proliferative in-
dices did not come out as significant prognostica-
tors in tumors of ?2 cm in diameter. The reason for
this was purely technical, as the algorithm did not
calculate the prognostic influence if there were no
deaths in the studied group.
Lymph node status as a predictive variable was
significant in the whole data set (P ? .031) and in
premenopausal patients (P ? .052).
Determination of decision cut points in the Ni-
gerian material resulted in an obvious cut point at
92. This figure turned out to be the one with the
greatest relevance for both SMI and MAI. The anal-
ysis detected only one cut point surrounded by less
significant cut points. For morphometric grading,
we propose SMI ? 32 as the lower cut point for
morphometric grading in the Nigerian material.
Also at this threshold, the difference on both sides
of the cut point is statistically significant.
Multivariate analyses were performed using tu-
mor size (5-cm cut point), MAI, and/or SMI as
prognosticators (Table 4). Because of limited num-
ber of deaths among lymph node– or ?2-cm
groups, the multivariate significance of lymph node
status and the ?2-cm cut point could not be ana-
lyzed. In multivariate analysis with the grading-
associated cut points (17 mitotic figures per square
millimeter for SMI, 10 mitotic figures per 10 high-
power fields for MAI; 14) of SMI and MAI, tumor
size at the 5-cm cut point was the most significant
independent prognosticator (more significant than
the mitotic counts). However, when the cut point of
92 was used, SMI (RR 6.9) and MAI (RR 7.5) had
about the same significant risk ratio as the tumor
size at 5-cm cut point (RR 5.5). Among the pre-
menopausal patients, the importance of mitotic in-
dices increased (SMI, RR 14.9). Mitotic indices were
also significant prognosticators when treated as
continuous variables among all patients and in the
Figure 1 demonstrates the SMI-associated sur-
vival curves as defined by the cut points at 32 mi-
totic figures per square millimeter of neoplastic
TABLE 3. Univariate Analysis on the Significance of the Most Important Prognosticators in the Nigerian Material
Patient GroupPrognostic FeaturePRisk Ratio
All (n ? 129)SMI17
Tumor size of ?2 cm
Tumor size ?5 cm
Tumor size of ?2 cm
Tumor size ?5 cm
Tumor size of ?5 cm
Tumor size of ?5 cm
Premenopausal (n ? 97)
Postmenopausal (n ? 32)
Node positive (n ? 76)
Tumor size of 5 cm (n ? 66)
Thresholds are 17 and 92 mitotic figures per square millimeter and 10 and 92 mitotic figures per 10 high-power fields for SMI and MAI, respectively.
tissue in the materials from the 129 patients avail-
able for survival studies.
Figure 2 shows survival curves at the cut point of
92 mitotic figures per square millimeter in the
whole data set (129 patients). There is a dramatic
survival difference at this cut point.
Differences in the age at presentation and histo-
logical dissimilarities between breast cancer in Ni-
geria and Finland had been previously described (4,
Although demographic differences between the
Nigerian and Finnish population may contribute to
the lower age at presentation, dietary, genetic, and
environmental factors may independently or simul-
taneously influence the biological and clinical pat-
terns observed (16–19). The reproductive factors
seem to influence the occurrence of breast cancer
in a similar fashion in the two countries.
The proliferative activity is an established prog-
nostic feature in breast carcinomas (5–9). Studies
have suggested that prognostic variables may have
different prognostic association at different time
periods during the follow-up (10). The significant
factors observed in the Nigerian material will,
therefore, reflect the prognostic variables in the
early stages of follow-up in Nigerian breast carci-
nomas. The limited follow-up time is a potential
weakness in this study, and future studies with
longer follow-up times will be beneficial.
The mean values of SMI and MAI were 42.64
mitotic figures per square millimeter and 30.53 mi-
totic figures per 10 high-power fields, respectively
in the Nigerian material. The corresponding values
in the Finnish material were 13.8 mitotic figures per
square millimeter and 10.7 mitotic figures per 10
high-power fields (14). Proliferative differences be-
tween the Nigerian and Finnish tumors were signif-
icant in the overall material and in the different
subgroups studied (P ? .001). Because there were
no Stage 4 cases in the Finnish material, a compar-
ison was made using the Nigerian material of Stages
1 to 3. Although the proliferative indices of the
Finnish material were within ranges observed in
other studies (5–8), the values in the Nigerian ma-
terial are much higher.
It is important to note that material used in the
other studies was from Western countries. Our
study is in concordance with observations in the
United States, where after adjusting for age, stage,
TABLE 4. Multivariate (Cox’s Regression) Analyses with Different Prognosticators and Cut Points
VariableP Risk Ratio95% CI Data Group
Tumor size, 5 cm
MAI (cut point of 10)
Tumor size, 5 cm
SMI (cut point of 17)
Tumor size, 5 cm
MAI (cut point of 92)
SMI (cut point of 92)
Tumor size, 5 cm
MAI (cut point of 92)
SMI (cut point of 92)
MAI expressed as mitotic figures per 10 high-power fields. SMI expressed as mitotic figures per square millimeter of neoplastic epithelium. CI,
confidence interval; MAI, mitotic activity index; SMI, standardized mitotic index.
FIGURE 1. Survival of patients divided according to SMI values
among 102 patients with Stage 1–3 tumors. The used cut point is 32,
proposed as a relevant cut point for grading Nigerian breast cancer at
the present time. The upper curve represents survival of patients with
SMI of ?32 (n ? 37; 1 dead) whereas the lowest curve incorporates
patients with SMI of ?32 (n ? 65, 15 dead). There is a clear and
significant survival difference (log-rank P ? .006).
FIGURE 2. Survival of 102 patients with infiltrating breast carcinoma,
with Stage 1–3 tumors (Stage 4 excluded from the follow-up material of
129 patients), divided by standardized mitotic index (SMI, also called
volume fraction–corrected mitotic index, M/Vv-index). Patients with
mitotic counts of ?92 (n ? 9, 8 dead) have a dramatically worse
survival than patients with SMI of ? 92 (n ? 93, 8 dead). The survival
difference is highly significant (log-rank P ? .0001).
Biology and Prognostication in Nigerian Breast Cancer (O.F.R. Ikpatt et al.)787
socioeconomic status, reproductive experience,
and health care access, African-American patients
had significantly higher mitotic activity and grade
nuclear atypia than did their Caucasian counter-
The Nigerian material was processed after vari-
able fixation time. Despite this (20, 21), the higher
proliferative activity could still be determined with-
out practical problems. Considering the little infor-
mation available on the proliferative activities of
epithelial tissues in the normal breast and benign
mammary conditions in the Nigerian female, the
interpretation of these findings is problematic.
However, several explanations can be presented.
The prevalence of obesity in the Nigerian female
(22, 23) is 22.3% (versus 18% among Finnish wom-
en; 24). Obesity is usually associated with higher
plasma estrogen levels (25). This, together with
other exogenous sources of dietary estrogens, may
contribute to the observed increase in proliferative
rates. Estrogen as a carcinogen may have permis-
sive, promotional, and/or tumor growth–inducing
influences in the multistage development process
of breast carcinoma (25). However, differences in
the prevalence of obesity alone are not sufficient to
account for the significant differences in prolifera-
tive activity of the tumors in the two studied pop-
regulating genes may contribute to the high mitotic
activity seen (26). However, there is no evidence of
corresponding influences in the Nigerian material.
Finnish premenopausal patients (14) had higher
values of the proliferative indices than postmeno-
pausal patients. In the Nigerian material, there was
no statistically significant difference between pre-
menopausal and postmenopausal patients. This
finding may reflect the more advanced nature of
the disease in the postmenopausal patients.
Aaltomaa et al. (7) suggested that all breast tu-
mors could be graded using the same principles
when the mitotic indices are determined. This was
based on their observing insignificant proliferative
activity differences between ductal carcinomas and
the special forms of breast cancers (10).
The finding of a significant correlation observed
between the SMI and other prognostic factors, in-
cluding nuclear measurements, in the Nigerian ma-
terial also exists in the Finnish (7, 9, 14) and other
studies (5, 6, 8). This is understandable because the
nuclear size (8) is also dependent on the activity of
the cell cycle.
Our results show that the earlier proposed cut
points (14) for MAI and SMI are also applicable in
the Nigerian material in separating patients with
favorable and unfavorable outcome of the disease
when used independently or in combination with
other prognosticators as observed in earlier Finnish
studies (7, 9, 14).
However, more significant cut points (SMI as the
example) are 32 and 92 mitotic figures per square
millimeter. At 92 mitotic figures per square milli-
meter, there is a true cut point. At 32 mitotic figures
per square millimeter, the significance of the cut
point is still increasing towards 92 mitotic figures
per square millimeter, which is the only true ditch
in a curve of P values over the range of Nigerian SMI
values. The cut points could be used for morpho-
metric grading in Nigeria, though the system out-
lined by Kronqvist et al. (14) is still applicable in the
Nigerian material. It is therefore necessary to vali-
date the determined cut points in separate Nigerian
materials after an adequate follow-up period, as the
conventional Western grading parameters may be
suboptimal in African tumors.
Counting of mitotic figures can be performed
with good reproducibility, and it is a relatively
cheap process (8, 11). The use of this technique in
the prognostic evaluation of African breast cancers,
especially in lymph node–positive premenopausal
patients with large tumors, should be encouraged.
This is because the availability of other methodol-
ogies, like flow cytometry or immunohistochemis-
try, is still limited in Africa.
1. Ijaduola TG, Smith EB. Pattern of breast cancer among
white-American, African-America, and non-immigrant west-
African women. J Natl Med Assoc 1998;90:547–51.
2. Ihekwaba FN. Breast cancer in Nigerian women. Br J Surg
3. Oluwole SF, Fadiran OA, Odesanmi WO. Diseases of breast
in Nigeria. Br J Surg 1987;74:582–5.
4. Ikpatt OF, Collan Y. Histological pattern of breast cancer in
Nigeria. Pathologica 2001 (in press).
5. Ladekarl M, Jensen V. Quantitative histopathology in lymph
node-negative breast cancer. Prognostic significance of mi-
totic counts. Virchows Arch 1995;427:265–70.
6. Van Diest PJ, Baak JP, Matze-Cok P, Wisse-Brekelmans ECM,
van Galen CM, Kurver PHJ, et al. Reproducibility of mitosis
counting in 2469 breast cancer specimens: results from mul-
ticenter morphometric mammary carcinoma project. Hum
7. Aaltomaa S, Lipponen P, Eskelinen M, Kosma V-M, Marin S,
Alhava E, et al. Mitotic indexes as prognostic predictors in
female breast cancer. J Cancer Res Clin Oncol 1992;118:75–
8. Baak JPA, van Dop H, Kurver PHJ, Hermans J. The value of
morphometry to classic prognosticators in breast cancer.
9. Haapasalo H, Pesonen E, Collan Y. Volume corrected mitotic
index (M/V-INDEX). The standard of mitotic activities in
neoplasms. Pathol Res Pract 1989;185:551–4.
10. World Health Organization. Histological typing of breast
tumours. In: International histological classification of tu-
mours. 2nd ed. Geneva: World Health Organization; 1981.
11. Bloom HJG, Richardson WW. Histopathological grading and
prognosis in breast cancer. Br J Cancer 1957;11:359–77.
12. Baak JPA, Oort J. A manual of morphometry in diagnostic
pathology. New York: Springer; 1983.
13. Kuopio T, Collan Y. Mitotic counts in breast cancer: the
influence of training performance. Pathol Res Pract 1996;
14. Kronqvist P, Kuopio T, Collan Y. Morphometric grading in
breast cancer: thresholds for mitotic counts. Hum Pathol
15. Cutler SJ, Ederer F. Maximum utilisation of the life tables
method in analysing survival. J Chron Dis 1958;8:699–712.
16. Pathak DR, Osuch JR, He J. Breast carcinoma aetiology:
current knowledge and new insights into the effects of re-
productive and hormonal risk factors in black and white
populations. Cancer 2000;88:1230–8.
17. Wu AH. Diet and breast carcinoma in multiethnic popula-
tions. Cancer 2000;88:1239–44.
18. Bergman-Jungestrom M, Gentile M, Lundin A, Wingren S.
Association between CYP17 gene polymorphism and risk of
breast cancer in young women. Int J Cancer 1999;84:350–3.
19. Johnson-Thompson MC, Guthrie J. Ongoing researches to
identify environmental risk factors in breast carcinoma.
20. Cross SS, Start RD, Smith JHF. Does delay in fixation affect
the number of mitotic figures in processed tissue? J Clin
21. Bergers E, Jannink I, van Diest PI, Cuesta MA, Meyer S, van
Mourik JC, et al. The influence of fixation delay on mitotic
activity and flow cytometric cell cycle variables. Hum Pathol
22. Oguntona CR, Kuku O. Anthropometric survey of the elderly
in south-western Nigeria. Ann Hum Biol 2000;27:257–62.
23. Okoronkwo MO. Body weight and uterine leiomyomas
among women in Nigeria. W Afr J Med 1999;18:52–4.
24. Pietinen P, Vartiainen E, Mannisto S. Trends in body mass
index and obesity among adults in Finland from 1972 to
1992. Int J Obes Relat Metab Disord 1996;20:114–20.
25. Bland KI, Copeland EM. Growth regulation of normal and
malignant breasts epithelium. In: The breast: comprehen-
sive management of benign and malignant diseases. Phila-
delphia: Saunders; 1991.
26. Taioli E, Bradlow HL, Garbers SV, Sepkovic DW, Osborne
MP, Trachman J, et al. Role of estradiol metabolism and
CYP1A1 polymorphism in breast cancer risk. Cancer Detect
27. Ikpatt F, Kuopio T, Collan Y. Nuclear morphometry in Nige-
rian breast cancer: the prognostic value. Image Anal Stereol
2002;21 (in press).
Histological characteristics of breast carcinoma in blacks and
whites. Cancer Epidemiol Biomarkers Prev 1994;3:127–35.
29. Simon MS, Severson RK. Racial differences in breast cancer
survival: the interaction of socio-economic status and tu-
mour biology. Am J Obstet Gynecol 1997;176:233–9.
30. Hall IJ, Newman B, Millikan RC, Moorman PG. Body size and
breast cancer risks in black women and white women. The
Carolina breast cancer study. Am J Epidemiol 2000;151:754–
Biology and Prognostication in Nigerian Breast Cancer (O.F.R. Ikpatt et al.) 789