Breast Cancer Prevention
Risk prediction and prevention is a key research
programme within the Manchester Breast Centre which along with the
Christie NHS Foundation Trust, encompasses The University of
Manchester, Faculty of Medical and Human Sciences; South and
Central Manchester NHS Trusts; Greater Manchester & Cheshire
Cancer Research Network.
The Principle Investigators in this programme are Anthony
Howell, Professor of Medical Oncology (Christie NHS Foundation
Trust) and Gareth Evans, Professor of Clinical
Genetics (Central Manchester and Manchester Children's
University Hospital trust).
Research Overview
One in nine women develop breast cancer during their lifetime in
the UK. Breast cancer remains the most common tumour in women and
the incidence continues to increase with over 40,000 new cases per
year. Because of the increase and the availability of agents such
as tamoxifen there is more emphasis on prevention of the disease.
However a limitation of the preventive use of tamoxifen, for
example, is the large numbers of at risk women needed to prevent
one cancer, which in turn, is an indication of the deficiencies of
our methods to predict risk of the disease. Thus the aims of our
studies in this programme is to attempt to improve risk prediction
and to devise more effective methods for prevention by enhancing
current approaches and attempting to devise novel ones.
Risk Prediction - Background
Current methods of prediction of risk depend upon integrating
risks associated with a family history of breast cancer and other
well-known 'endocrine' risk factors such as age of first pregnancy
and menopausal status. The available risk programmes include one or
both groups of factors. We evaluated four risk prediction packages
in 1,933 women at risk in our Family History Clinic (FHC). Fifty
two cancers developed and we computed the expected over observed
number of cancers for each package. The Gail (E/O 0.48, 95% CI
0.54-0.90), Claus (E/O 0.56, 95% CI 0.59-0.99) and Ford (E/O 0.49,
95% CI 0.52-0.86) performed less well than a new model devised by
Tyrer and Cuzick (E/O 0.81, 95% CI 0.85-1.41). The greater
predictive power of the new model in our at risk population
probably relates to inclusion of a full family history (Gail does
not) and several endocrine risk factors (Claus and Ford do not) and
warrants general use in the clinic and efforts to improve its
predictive power.
With our national and international collaborators we have
devised methods for improving prediction of risk associated with
BRCA1/2 mutations and attempted to discover new risk associated
genes, with some success. Our studies suggest that the FHC setting
is appropriate for such investigations and there is good patient
collaboration and satisfaction. Our FHC provides an important
research resource for future studies. We advise over 290 families
with BRCA1/2 and TP53 mutations and over 600 high-risk mutation
negative families.
Although mammographic screening is not a predictor of risk its
evaluation in the context of risk in young women is important. Our
group is particularly involved in assessing new methods of
mammogram visualisation and computer aided detection, especially in
view of the denser mammograms often found in younger women. We are
involved in trials to assess the value of MRI and mammography in
women at increased risk. Assessment of screening in our own centre
suggests it may be of value and may improve survival compared with
symptomatic women. The above considerations indicate that, although
some improvements have occurred in risk prediction there are still
outstanding questions which we are attempting to address through
new research projects.
Can we improve current methods of risk prediction?
One approach is to attempt to add additional risk factors such
as mammographic density and weight to the current algorithms.
Mammographic density is a major risk factor for breast cancer.
Having performed preliminary studies in order to relate density to
other risk factors in the IBIS tamoxifen prevention trial, we now
plan to perform a case control study where we will relate breast
density in 240 women in our programme who have developed breast
cancer with 4 matched controls without cancer from our clinics.
Although of lower magnitude we and others have shown that weight
is an important risk factor for breast cancer. We have demonstrated
that postmenopausal breast cancer is associated with general rather
than central obesity, current BMI is equivalent to weight gain as a
risk factor, and 5% weight loss reduced risk by 25-40% compared
with continues weight gain. Thus current weight and/or BMI may be
used to improve the predictive power of risk algorithms and we will
work with Cuzick to attempt to improve the Tyrer-Cuzick algorithm
and an international consortium also interested in combining risk
factors.
Can we develop new methods to determine risk?
Most women who develop breast cancer do not have any appreciable
risk factors for the disease. There must be other factors, which
confer risk. We hypothesise that there will be a gene signature
associated with risk in tissue taken from the normal breast in
women at risk. Since we also hypothesise that the risk-associated
profiles will be the net effect of interaction between stromal and
epithelial cells and thus whole rather than micro-dissected tissue
should be used. We have preliminary evidence that gene expression
profiles can distinguish between population and high-risk women. We
now wish to search data bases and perform additional clinical
studies to find profiles, which indicate particular physiological
states of the breast, which may lead to determination of further
important risk genes.
Prevention - Background
Epidemiological studies indicate reduction in lifetime exposure
to ovarian steroid hormones reduces breast cancer risk and
clinically oophorectomy and anti-oestrogens reduce risk is high in
risk populations. In addition, weight loss, exercise and the use of
non-steroidal anti-inflammatory drugs are also associated with
reduced risk albeit the data are only from epidemiological studies.
Studies are required to improve on current endocrine approaches and
to detect novel ones.
Can we improve on current approaches to breast cancer
prevention?
We are currently analysing the 7 year results of the IBIS I
Tamoxifen Prevention trial which will give an assessment of this
approach 2 years after completion of treatment. The ATAC trial
showed that anastrozole reduced contralateral breast cancer by
about 50% compared with tamoxifen (Programme 5, ref 51). This has
led to the initiation of the IBIS II trial comparing 5 years of
anastrozole with placebo (AH Co-PI). Importantly we have
incorporated subprotocols to assess the effect of anastrozole on
bone, cognitive function and body composition. We have performed
pilot studies to assess the feasibility of 2 years ovarian
suppression with goserelin with add back raloxifene and also
another study in collaboration with RMH of 3 years phytoestrogen in
the form of red clover tablets. These studies are currently being
analysed and we will then make decisions whether to study these
approaches in large international trials.
Can we develop new approaches to breast cancer prevention?
The current approach to prevention of larger international
trials has given important insights into this approach but the
studies are expensive and lengthy. Others have and are attempting
to use some form of breast biopsy technique before and after a
potential agent pioneered by Harper-Wynne et al at RMH. We believe
this is an important avenue and have initiated proof of principle
studies to assess the effect of known preventive strategies
(oophorectomy and anastrozole) and weight loss on breast gene
expression. The results of these studies will inform our future
approaches to new agents.
Key Research Achievements Since 2001
- Risk assessment packages compared
- Computer aided mammographic detection developed
- New method for determination of gland volume developed
- Further delineation of the Li-Fraumeni syndrome.
- Tamoxifen prevents breast cancer in the short term (IBISI)
- Low uptake in prevention but not screening trials reported
- Scoring system for implementation of gene testing
developed
- Favourable psychosocial impact of tamoxifen prevention
reported
- Large study showing 5% wt loss reduced breast cancer risk
- Effectiveness of mammographic screening in high risk women
- Effectiveness of prophylactic surgery delineated
