Invasive breast carcinoma - IARC

Fig. 1.73 A Supervised classification on prognosis signatures, using a set of prognostic reporter genes
to identify optimally two types of disease outcome from 78 sporadic breast tumours into a poor prognosis
and good prognosis group. B Each row represents a tumour and each column a gene. Genes are
ordered according to their correlation coefficient with the two prognostic groups. Tumours are ordered
by the correlation to the average profile of the good prognosis group. Solid line, prognostic classifier with
optimal accuracy; dashed line, with optimized sensitivity. Above the dashed line patients have a good
prognosis signature, below the dashed line the prognosis signature is poor. The metastasis status for
each patient is shown in the right panel: white indicates patients who developed distant metastases
within 5 years after the primary diagnosis; black indicates patients who continued to be disease-free for
at least 5 years. From L.J. van’t Veer et al. {2986}.
association between specific mutations
and short survival and poor response to
t reatment was seen, emphasizing the
i m p o rtance of DNA sequence analysis
of the entire coding region of T P 5 3
when evaluating its prognostic and predictive
value.
Loss of heterozygosity (LOH)
LOH at the T P 5 3 gene has been shown
to be a marker for prognosis and predictor
of response to certain therapies
(see above). Other regions with LOH
that appear to correlate to short survival
include 11q23 and several regions on
3p {1216,1552}. Deletion of 9q13 is also
associated with shorter survival. The
target gene(s) in these areas have still
to be identified {1326}.
D N A a m p l i f i c a t i o n
Conventional as well as array-based
CGH have identified a number of amplified
regions containing putative oncogenes
with prognostic potential. Amplification
of the FGFR1 gene on 8p12 has
been correlated with reduced diseasefree
survival, especially if the gene is
amplified together with the cyclin D1
gene {596}. The MYC gene on 8q24 is
amplified in approximately 20% of breast
c a rcinomas, which is associated with
estrogen receptor negativity {596}, locally
advanced disease and poor prognosis
{250}. On 11q13, cyclin D1 (CCND1) is
amplified in 15-20% of breast tumours. In
ER-positive tumours, CCND1 amplification
is associated with a relatively poor
prognosis {596,2582}, and is more frequent
in lobular carcinomas compared to
ductal carcinomas.
Expression profiling
Much recent work has been focused
on the potential of gene expression
p rofiles to predict the clinical outcome
of breast cancer { 2 5 7 , 1 2 5 7 , 2 3 2 8 , 2 7 5 7 ,
2986,2990}. These studies, although
h e t e rogeneous in patient selection and
numbers of tumours analysed, have indicated
that gene expression pattern s
can be identified that associate with
lymph node or distant metastasis, and
that are capable of predicting disease
course in individual patients with high
accuracies (circa 90%). In the largest
study to date {2990}, analysing 295
tumours, the expression profile was a
s t rong independent factor and outcompeted
lymph node status as a pre d i c t o r
of outcome. These findings suggest
that some primary tumours express a
"metastasis signature", which is diff i c u l t
to reconcile with the classic tumour prog
ression model in which a rare subpopulation
of tumour cells have accumulated
the numerous alterations re q u i re d
for metastasis to occur. Intere s t i n g l y,
some of the genes in the signature
seem to be derived from non-epithelial
components of the tumour {2328}, suggesting
that stromal elements re p re s e n t
an important contributing factor to the
metastatic phenotype. Survival diff e r-
ences were also noted between the diff
e rent subtypes of breast tumours as
defined by expression patterns {2756,
2757}. The patients with basal-like and
ERBB2+ subtypes were associated with
the shortest survival, while a diff e re n c e
in the outcome for tumours classified as
luminal A versus luminal B was also evident.
The luminal subtype B may re p resent
a class of ER-positive tumours with
poor outcome, possibly not re s p o n d i n g
to tamoxifen. This strongly supports the
idea that many of these breast tumour
subtypes re p resent biologically distinct
disease entities with diff e rent clinical
o u t c o m e .
A remarkable feature of the expre s s i o n
s i g n a t u res identified in these studies is
that they usually involve fewer than 100
genes {257,2986}, in one instance even
only 17 genes {2328}. However, somewhat
confusing is that the overlap
between the diff e rent sets of genes thus
defined is incomplete {1257,2757}.
F u rther comparative studies are
re q u i red to elucidate the critical components
of the poor prognosis signature ,
while the clinical utility of this new diagnostic
tool must now be demonstrated
in a prospective trial setting. At a more
fundamental level, it will be intere s t i n g
to establish whether the observed association
between expression signature s
and survival reflects an intrinsic biological
behaviour of breast tumour cells or
a diff e rential response to therapy.
Invasive breast carcinoma
59