E11. Advances in breast cancer imaging
A. Tardivon*, C. Malhaire, A. Athanasiou, F. Thibault, C. El Khoury
Department of Radiology, Institut Curie, 26 rue d’Ulm, 75248 Paris Cedex 05, France
The goal of cancer imaging is to detect the smallest
possible number of tumour cells and could be re-
duced to the signal-to-background ratio principle – the
signal generated by the cancer must be higher than
the background caused by a non specific signal or
nearby normal cells (contrast resolution) – and to the
spatial resolution of the imaging modality used. Breast
cancer detection using mammography and ultrasound
is based on natural endogenous tissue contrasts. With
mammography, cancer detection is easy in fatty breasts
and when the cancers exhibit distortion or associated
calcifications. These clinical situations are frequent in
the general population concerned (50−74 years) and
explain the success of screening programmes.
1
When the
breast density progressively increases, the sensitivity of
mammography rapidly decreases (<50%); this limitation
is particularly critical in the surveillance of high-risk
young women (gene-mutation carriers and women with
a significant familial history of breast/ovarian cancers).
2
Ultrasound (US) may improve this detection because
the normal dense fibro-glandular tissue appears “white’
whereas cancers are “black’.
3−4
However, when the
normal tissue is heterogeneous or the cancer atypical
(hyperechoic feature), false negative results may occur.
Another problem with US is its operator dependency with
a poor positive predictive value of biopsies (high percent-
age of false positive results). The introduction of full-field
digital mammography (FFDM) since 2000 has offered the
advantages of digital technologies (storage, transfer and
teleradiology) and produced new technological advances
that improve detection and characterisation: computed-
aided detection or diagnosis systems (important to
consider in the European context associating a decrease
in the number of radiologists and an increase in the
number of women concerned by screening in the next
decade), tomosynthesis (3D approach) that generates
thin contiguous slices of breast parenchyma limiting
superimposition problems and increasing background-
lesion contrast in heterogeneous breasts.
5−7
This 3D
anatomical approach is also available now with US using
a large field of view and automatic scanning coupled
to mammography data. Beyond these morphological
approaches, functional imaging is another tool to improve
detection and characterisation. The most relevant and
historical example is the detection of neoangiogenesis
that appears at early stages of the disease. This detection
can be obtained without contrast agents (US doppler
modes, optical imaging) or after injection of nontargeted
exogenous contrast agents (iodine agents for FFDM
and CT, microbubbles for US, gadolinium chelates for
MRI, 15O and 11C-labelled radiotracers for PET).
8
Most
routine analyses are still qualitative (present or not) and
semi-quantitative (e.g. with MRI: slopes of initial and
late enhancement, time-to-peak of enhancement, time-
intensity curves). Among all the imaging modalities
available for this approach, MRI has several advantages:
no use of X-Rays, 3D acquisition, analysis of both
breasts, no influence of breast density and possible
second opinions. Therefore, breast MRI has emerged as
the most sensitive modality for detecting breast cancer
and has been integrated in the surveillance protocols
of high-risk women.
9−10
However, because of the use
of low-molecular contrast agents, enhancement is not
specific and quantitative evaluation is still confidential
with MRI (no consensus about acquisition protocols
and analysis tools); research on macromolecular agents
has been ongoing for many years but without routine
implementation at this time. Recently, dedicated breast
computed tomography (CT) with low doses has been
tested with interesting preliminary results and will
probably re-emphasize perfusion CT, a robust method
to quantify precisely tumour flow parameters (response
evaluation under treatment).
8,11
Because cancers are
usually stiffer than benign lesions, elastography has
been studied since the 1980s and is now available
in real-time with standard US (semi-quantitative and
quantitative approaches) and MRI and demonstrated
a significant gain in specificity.
12−13
Other functional
imaging possibilities rely on the analysis of tumour
cell replication and metabolism (Sestamibi with SPECT,
choline components with MR-spectroscopy, diffusion
MRI, glucose metabolism with FDG PET).
13−15
If
there is a significant gain in specificity compared to
angiogenesis imaging, the limited spatial resolution of
all these technologies does not allow detection or
characterisation of small lesions (under 8 mm) but
are very useful, in cancers not candidates for breast-
conserving surgery, to evaluate response to systemic
treatments early after their initiation (one cycle of neo-
adjuvant chemotherapy) and before detection of mor-
phological changes. Technical improvements are ongoing
for MRI with the introduction of higher magnetic
1359-6349/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.