Three-dimensional time-resolved optical mammography
of the uncompressed breast
Louise C. Enfield,1,* Adam P. Gibson,1Nicholas L. Everdell,1David T. Delpy,1Martin Schweiger,2
Simon R. Arridge,2Caroline Richardson,3Mohammad Keshtgar,3Michael Douek,3
and Jeremy C. Hebden1
1Department of Medical Physics and Bioengineering, University College London, Gower Street, London, WC1E 6BT, UK
2Department of Computer Science, University College London, Gower Street, London, WC1E 6BT, UK
3Department of Surgery, University College London, 4th Floor, The Medical School Building,
74 Huntley Street, London, WC1E 6AU, UK
*Corresponding author: email@example.com
Received 30 October 2006; revised 24 January 2007; accepted 31 January 2007;
posted 31 January 2007 (Doc. ID 76581); published 18 May 2007
Optical tomography is being developed as a means of detecting and specifying disease in the adult female
breast. We present a series of clinical three-dimensional optical images obtained with a 32-channel
time-resolved system and a liquid-coupled interface. Patients place their breasts in a hemispherical cup
to which sources and detectors are coupled, and the remaining space is filled with a highly scattering
fluid. A cohort of 38 patients has been scanned, with a variety of benign and malignant lesions. Images
show that hypervascularization associated with tumors provides very high contrast due to increased
absorption by hemoglobin. Only half of the fibroadenomas scanned could be observed, but of those that
could be detected, all but one revealed an apparent increase in blood volume and a decrease in scatter and
oxygen saturation. © 2007 Optical Society of America
Breast cancer is responsible for nearly one third of
cancer cases and one fifth of cancer deaths in women
in the UK, which represents ?40,000 cases and
?12,000 deaths each year . Early detection de-
creases mortality , so many countries in the devel-
oped world routinely screen for breast cancer. The
most commonly used breast-imaging tool is x-ray
mammography . However, x-ray mammography
has a number of disadvantages: specificity is poor for
some types of tumor, which can lead to unnecessary
biopsies; it uses potentially hazardous ionizing radia-
tion; and some women report that the required com-
pression of the breast is painful. Mammography is also
less suitable for younger women due to the increased
density of the breast tissue. In cases involving younger
women, ultrasound or magnetic resonance imaging
(MRI) is used [4,5], although neither of these tech-
niques is suitable for screening asymptomatic women.
Therefore a safe and effective imaging technique that
can provide a distinction between benign and malig-
nant lesions would be of very considerable benefit.
The use of near-infrared (NIR) light to examine the
breast has been actively researched since the 1980s,
when developments in source, detector, and comput-
ing technology made diffuse optical imaging feasible.
So-called optical mammography provides discrimina-
tion between tissues based on their optical properties
obtained from measurements of transmitted light
[6–8]. In principle, the different spectral absorption
characteristics of deoxyhemoglobin and oxyhemoglo-
bin allow images to be derived from blood volume and
blood oxygen saturation [9,10]. This provides valu-
able physiological information, which can also be
used, if necessary, combined with anatomical imag-
ing methods such as x-ray mammography or ultra-
sound . Many research groups have built and
begun to clinically evaluate optical imaging systems,
© 2007 Optical Society of America
3628 APPLIED OPTICS ? Vol. 46, No. 17 ? 10 June 2007
based on a variety of measurement types (e.g., con-
tinuous intensity, time and frequency domain) at var-
ious NIR wavelengths [12–19]. One of the major
challenges of optical mammography is to adequately
couple a large number of optical sources and detec-
tors to the breast, and various methods have been
explored. Some groups use a compressed breast ge-
ometry [14,16,17], where one or more sources and
detectors are scanned over both surfaces, and a single
projection image is produced. A disadvantage of this
method is that compression can lead to discomfort,
especially in women suffering from sore breasts. An
alternative method is to use a fixed array of source
and detectors that surround the breast, from which
either two-dimensional (2D) slices or a whole volume
image can be reconstructed [12,18,19].
Over the past two or three years, there has been a
large number of clinical studies published by various
research groups. A research group at Physikalische-
Technische Bundesanstalt, Berlin, has conducted a
major study of the in vivo blood concentration and
oxygen saturation in breast tumors, using a dual-
wavelength time-domain optical mammography sys-
tem [16,20,21]. Using a triple wavelength scanning
(compression geometry) time-resolved imaging sys-
tem, craniocaudal and mediolateral projections were
recorded from 154 patients suspected of having breast
cancer [20,21]. Images representing absorption, scat-
ter, total hemoglobin concentration, and blood oxygen
saturation were derived for each patient. Out of 102
histologically confirmed tumors, 72 were detected in
both views, 20 in just one projection, and 10 were not
detectable at all . On average, tumor absorption
coefficients were 2.5 times greater than surrounding
tissue at 670 nm, and reduced scattering coefficients
of tumors were 20% higher than surrounding tissue
at 670 nm. Hemoglobin concentrations were system-
atically larger in tumors than in surrounding tissue,
although blood oxygen saturation was found to be a
poor discriminator between healthy and tumor tissue
Meanwhile, researchers at the Politecnico di Mi-
lano have developed a similar scanning time-resolved
optical mammography system, which has been used
to scan 194 patients [17,22,23]. In the first part of
their study, 101 patients with 114 lesions were
scanned with a four-wavelength system. To improve
the spectral content of the images, the system was
expanded to seven wavelengths ?637–985 nm?. Im-
ages revealed sensitivity to tissue composition (oxy-
and deoxyhemoglobin, water, lipids) and physiology
(blood volume and oxygenation), as well as structural
changes. An 80% detection rate for cysts and malig-
nant lesion types was found when images were
acquired in both craniocaudal and oblique views.
Cancers exhibited an increased absorption coefficient
compared with surrounding tissue, while cysts had a
lower reduced scattering coefficient. The detection
rate for other benign lesions such as fibroadenomas
was much lower ??40%?.
A breast-imaging system developed at Dartmouth
College is based on 16 sources and 16 detectors
mounted on linear translation stages arranged in
three circular arrays . The patient lies prone with
her breast suspended within the arrays, and the
sources and?or detectors are moved in direct contact.
Variations in hemoglobin, water content, and tissue-
scattering properties between subjects have been ob-
served, and changes over the menstrual cycle have
been explored . The parameters vary greatly be-
tween subjects, indicating that tumor hemoglobin
concentration relative to background rather than ab-
solute values may be a better diagnostic indicator.
Studies have shown that tumors have a higher he-
moglobin concentration and higher oxygen satura-
tion compared with background  and that images
representing the scattering properties often highlight
lesions relative to surrounding healthy tissue .
A research group at the University of Pennsylvania
has developed and tested a combined frequency do-
main and continuous-wave clinical system that em-
ploys six wavelengths . The prone subject’s breast
is lightly compressed in an Intralipid-filled tank, and
sources and detectors are placed on either side. In-
tralipid is a fat emulsion that is widely used for optical
tissue equivalent phantoms. The system has also been
combined with positron emission tomography (PET)
using labeled fluorodeoxyglucose (FDG) to examine
metabolic activity  and with MRI to provide simul-
taneous anatomical and physiological data . A case
study of a patient undergoing neoadjuvant chemother-
apy with MRI and optical imaging showed a decrease
in tumor size, in agreement with MRI . A decrease
in the total hemoglobin concentration contrast be-
tween the tumor and healthy tissue was also observed.
The group has also been looking at the use of extrinsic
contrast agents such as indocyanine green (ICG) to
examine the uptake and outflow of the dye in tumors
and healthy breast tissue .
A breast-imaging system has been developed by
a team at the State University of New York that al-
lows both breasts to be imaged simultaneously .
The system employs four wavelengths (between 760
and 830 nm) and 32 source–detector fibers for each
breast. The subject lies with her breast pendant in
the imaging array, and the optical fibers are trans-
lated radially until they are in direct contact with the
breast. Data have been collected from a healthy vol-
unteer and from a patient with a 6 cm ductal carci-
noma in her left breast. Both were recorded at rest
and when performing a controlled Valsalva maneu-
ver. Data showed a marked difference between the
data in the dynamic response of the cancerous versus
the healthy breast .
The combination of optical imaging with conven-
tional diagnostic modalities is also being investigated
by researchers at Massachusetts General Hospital.
Optical imaging probes have been integrated into an
x-ray mammography system by mounting detectors
and sources on removable compression plates .
Optical and x-ray data are recorded successively with
the breast held in the same position. Scans on 18
patients with benign and malignant lesions revealed
an increase in optical absorption near the benign and
10 June 2007 ? Vol. 46, No. 17 ? APPLIED OPTICS 3629
malignant lesions, as well as in the dense fibroglan-
dular tissue, relative to the surrounding tissues.
The in vivo optical properties of breast lesions and
healthy tissues have also been extensively investi-
who have developed a hand-held optical probe, which
is placed against the breast. Diffuse reflection mea-
surements are made at points on the surface of the
breast . Studies on 12 premenopausal patients
with cancer show significant contrast between normal
and tumor regions in the concentrations of deoxyhe-
moglobin, oxyhemoglobin, water, and lipids .
The contribution of our group at University Col-
lege London (UCL) toward the development of op-
tical mammography is based upon a 32-channel
time-resolved optical imaging system . Our ini-
tial measurements on patients were performed us-
ing a simple interface consisting of two rings of
different sizes, to which the sources and detectors
were attached, mounted on a frame . The breast
was placed into whichever ring provided the bet-
ter fit for that patient. Twenty-one patients were
scanned, with 17 out of 19 lesions being detected,
including one subject scanned repeatedly over one
year [18,33]. However, certain limitations were iden-
tified, such as inconsistent coupling between the fi-
bers and the breast. This led to unusable intensity
data and subsequently to an inability to distinguish
between the scatter and absorption properties of the
tissue. The fixed rings also proved unsuitable for
older women as the breast tissue was too pendu-
lous. Furthermore, a lesion may be missed if it was
not located within or near the plane of the ring. To
overcome these problems, an alternative patient in-
terface has been developed , based on an ap-
proach originally adopted by Philips Research
Laboratories . The patient lies with her breast in a
hemispherical cup filled with tissuelike intralipid-
based scattering fluid. This method had three major
benefits. First, cups can be made of a sufficient size to
accommodate a large range of breast sizes and
shapes, enabling the entire three-dimensional (3D)
volume of the breast to be sampled. Second, the cou-
pling of the source–detector optics at the surface is
constant and independent of the subject, enabling
intensity to be used as a data type in the image
reconstruction. And third, the external geometry of
the reconstructed volume is known exactly, so an
accurate model can be generated. Initial experiments
using isolated absorbing and scattering targets
within the cup were conducted to evaluate the per-
formance of the system in terms of the contrast, spa-
tial resolution, and localization accuracy . Results
dent on the location of the targets within the imaged
volume. Initial studies on several healthy volunteers
revealed subtle heterogeneity, particularity in the
distribution of scatter. In this paper, we summarize
the results of subsequent scans performed on a cohort
of 38 patients with a range of benign and malignant
The UCL 32-channel time-resolved optical imaging
system  employs a portable fiber laser (IMRA Inc.,
USA) that produces pulses ?2 ps in duration at 780
and 815 nm, interlaced with a repetition rate of
80 MHz. The pulses are transmitted through a 32-
way optical switch that illuminates each source fiber
in sequence. Each source fiber is integrated along the
axis of a detector fiber bundle. The bundles transmit
light to four eight-anode microchannel-plate photo-
multiplier tubes (MCP-PMTs). The MCP-PMTs pro-
duce an electronic pulse for each photon detected
and are protected from overexposure by variable
optical attenuators. Timing electronics record the
times taken for photons to travel through the illu-
minated medium and produce a time-of-flight his-
togram known as a temporal point-spread function
(TPSF) for each source–detector combination. Im-
ages are reconstructed (Subsection 2.D) using data
types extracted from the TPSFs, such as integrated
intensity, mean flight time, and variance. For the
studies reported here, intensity and mean flight
time have been used, as they have been proven to be
most robust and the most effective at separating
scatter and absorption .
University College London Optical Tomography
To accommodate breasts of different sizes, three rigid
hemispherical cups were constructed, with diameters
of 130, 160, and 180 mm, and depths of 70, 85, and
100 mm. Each was made by coating a latex rubber
hemisphere with a 5 mm thickness of fiberglass im-
pregnated with polyester resin (KS55, Alec Tiranti
Ltd., London, UK) mixed with black polyester paint
(Alec Tiranti Ltd., London, UK). Thirty-one regularly
spaced 1 cm diameter holes were made in the hemi-
sphere surface and Perspex windows were fixed into
each hole, providing a watertight volume . The
optical fiber bundles are attached to the exterior sur-
face of the cup by inserting them within short plastic
tubes fixed over each window. Despite having made
three different cups, the 160 mm diameter cup has
proven to be sufficient for the vast majority of pa-
tients, and all the results presented here were ob-
tained using this cup. The cup is attached to a plastic
ring that fits within an aperture in a wooden table,
as shown in Fig. 1. The coupling fluid that fills the
remaining space between the breast and the hemi-
sphere is made from a solution of Intralipid and near-
infrared dye in de-ionized water, which provides an
absorption coefficient of ?a ? 0.004 mm?1and a
transport scatter coefficient of ?s? ? 0.8 mm?1at
800 nm. These values were chosen to ensure that the
signal acquired for the largest source–detector sepa-
ration produces a measured photon count rate of
?50–100 photons?s, sufficient to obtain useful data
types for image reconstruction. The cup is filled with
fluid from below using a peristaltic pump, which then
3630APPLIED OPTICS ? Vol. 46, No. 17 ? 10 June 2007
circulates fluid continuously and slowly to maintain a
constant level of fluid. A channel is cut into the ring,
which allows coupling fluid that overflows from the
cup to return via plastic tubing to the fluid reservoir,
where it is maintained at a constant temperature of
37 °C using an electrically isolated water heater. The
table is covered with a layer of foam, and a pillow and
towels are provided. Patients lie prone with either
breast pendant in the cup, and most have reported
no discomfort within scan times of between 5 and
20 min. The fluid-filled cup mounted in the table is
shown in Fig. 2.
As reported previously , imaging using the
hemispherical cup employs a so-called difference ap-
proach, which involves reconstructing the changes in
optical properties between the cup containing both
the coupling fluid and the breast and those of the cup
containing the fluid alone. This eliminates the effect
of uncertainty and variability in the surface coupling,
allowing use of intensity measurements. The disad-
vantage of difference imaging is that errors will occur
if the properties of the reference medium are not
known exactly. To provide an adequate reference for
patient studies it is necessary to extend the height
of the reference medium beyond the top surface of the
fluid-filled cup, since light traveling between sources
and detectors located near the top of the cup can
migrate into and out of the chest wall. This extension
was provided using a thin, hollow latex cylinder, of
height 100 mm and diameter 200 mm, mechanically
supported by a plastic ring. To acquire a reference
measurement the cylinder is placed on top of the
fluid-filled cup and filled with fluid with the same
optical properties as that in the cup to a height of
50 mm above the cup. The latex base of the cylinder
is assumed to be sufficiently thin to have a negligible
influence on the data.
Suitable patients were recruited from among women
attending appointments with breast surgeons. In
most cases, a discrete lump had already been con-
firmed by conventional imaging (mammography or
ultrasound). The patients received an optical scan
either directly after their appointment at the clinic or
arrangements were made for them to be scanned a
few days later. Each scan took 11 min (using an illu-
mination of 10 s per source). Scans were made of the
diseased breast and, whenever possible, the healthy
breast for comparison. Data were recorded simulta-
neously at both source wavelengths.
Thirty-eight patients, aged between 22 and 80
(mean 41 ? 15 yr old) with a variety of breast dis-
eases were scanned: 19 women had one or more ma-
lignant lesions, 3 had cysts or galactoceals; 1 suffered
from mastalgia, 1 had a leaking breast implant, 1 had
a benign phyllodes, 1 had a malignant lesion in one
breast and a fibroadenoma in the other, 1 had fibro-
cystic changes, and 11 had fibroadenomas.
Three-dimensional images were generated from the
data using the time-resolved optical absorption and
scatter tomography (TOAST) reconstruction package
developed at UCL . TOAST employs the finite-
element method (FEM) to model the propagation of
light through tissue using the diffusion approximation
to the radiative transfer equation. TOAST produces
3D images of scatter and absorption coefficients. Im-
age reconstruction involves iteratively adjusting the
optical properties assigned to the FEM mesh to min-
imize the difference between the data types predicted
by the model and the data. The model starting values
are the measured absorption and scatter coefficients
of the coupling fluid. A hemispherical finite element
mesh with the 50 mm cylindrical extension (to sim-
ulate the cup with the chest wall extension) with
46,415 tetrahedral elements and 67,537 nodes was
generated using NETGEN software . The mesh is
assigned a greater density at the positions of the
source–detector bundles, where the rate of change of
light intensity is greater. The image reconstruction
was performed using a conjugate gradient solver and
Robin boundary conditions. In all cases, the 10th it-
eration was analyzed as this was found to be the
number beyond which no further improvement to the
images was observed. Each iteration required 48 min
on a 2.2 GHz Xeon processor and used approximately
800 MB of RAM.
We assume that the optical absorption represented in
the absorption images is due to three components:
oxyhemoglobin ?HbO2?, deoxyhemoglobin (Hb), and a
Derivation of Blood Volume and Oxygen Saturation
imaging. The patient lies with her breast pedant in a hemispher-
ical cup filled with scattering fluid.
Fluid-coupled patient interface for the 3D optical breast
Fig. 2. Fluid-filled cup viewed from above.
10 June 2007 ? Vol. 46, No. 17 ? APPLIED OPTICS3631
“background” component with the same absorption
coefficient at both wavelengths (780 and 815 nm),
assumed to be due to lipid and water. A background
value of 70% of the absorption averaged over the
image was selected, based on estimates of tissue op-
tical properties published elsewhere . However,
we also tested the influence of the assumed size of the
background by selecting a broad range of values be-
tween 55% and 85% of average absorption. The val-
ues of oxygen saturation varied by ?10%, while the
blood volume varied by ?2%. The coefficient of vari-
ance for an individual image was greater than this,
indicating that the absolute properties are not sensi-
tive to changes in background.
The combination of data at two wavelengths and a
knowledge of the appropriate extinction coefficients
can be used to derive the regional blood volume (rBV)
and the regional tissue oxygen saturation ?rStO2? as
rBV?volume of whole blood in tissue
total tissue volume
where ?HbO2? and [Hb] are the concentrations of oxy-
hemoglobin and deoxyhemoglobin, respectively. The
normal range of values for hemoglobin concentra-
tion in blood ?Hbb? in the healthy adult female is
12.2–15.0 g?dl ?1.9–2.3 mM?, and a typical value of
14 g?dl ?2 mM? is assumed here . The regional
blood volume (%) can be converted to an absolute unit
of tissue hemoglobin (?mol of blood in 1000 cm3) as-
suming a hemoglobin concentration of 2 mM as given
The 38 patient studies produced 42 scans of pathol-
ogies (21 malignant and 21 benign) and 27 scans of
healthy breasts. Some women had abnormalities in
both breasts, but due to patient time constraints not
all healthy breasts were scanned. Of all breast scans,
a total of 58 images were obtained: 35 from patholog-
ical conditions and 23 of healthy breasts. A small
number of scans failed to produce data sufficient to
reconstruct an image, due to temporary instability in
the system electronics or hardware malfunction.
Images typical of common types of lesion are
shown, and the results obtained on the full cohort are
discussed. Images representing scatter coefficient,
blood volume, and oxygen saturation are shown. For
each reconstructed 3D image, a single transverse
slice across the breast-containing cup, at the level of
the lesion, is displayed. Each scatter image obtained
at 780 nm was qualitatively indistinguishable from
the corresponding scatter image obtained at 815 nm
and, in each case, only the scatter image at 780 nm is
shown. The blood volume and blood oxygen satura-
tion images are derived from the 10th iteration of the
absorption images. The orientation of either breast in
the cup is illustrated in Fig. 3. However, it is impor-
entire cup. The breast often does not fill the cup, and
the center of the cup does not necessarily correspond
images are shown together, a common color scale is
Nineteen women with some form of malignancy in one
26 to 56 yr (mean 48 ? 13 yr). The malignant condi-
tions included carcinomas, a malignant phyllodes, and
an intracystic carcinoma. The first three patients were
scanned using only one wavelength, due to technical
difficulties; therefore only absorption and scatter im-
ages at 780 nm were obtained, and it was not possible
images. These first optical images are presented in
Figs. 4(a)–4(c). The images shown are coronal slices of
the 3D image of the breast through the region of the
lesion with maximum contrast.
Figure 4(a) is of a 53 yr old woman with a large
carcinoma behind the nipple of her left breast. Breast
MRI demonstrated a 4.3 cm ? 5.1 cm lobulated mass.
The left breast only was scanned due to discomfort of
the patient as a result of a bad back. The absorption
image shows a region of increased absorption in the
center of the image consistent with the known posi-
tion of the lesion. The scatter image shows a corre-
sponding region of increased scatter. This increase in
scatter may be due to physiological changes in the
cancerous cells or due to parameter cross talk, i.e.,
insufficient separation between the absorption and
scatter parameters (as previously observed in optical
images of the breast) [18,33].
Figure 4(b) shows the optical images from a 52 yr
old woman with bilateral breast cancer. MRI images
had revealed a distortion in the right breast, in-
creased stroma in the right-upper and lower-outer
quadrants of the breast, and a malignant mass in the
left-upper-outer quadrant of the left breast. An in-
Quadrants of the breast used to identify lesion loca-
3632 APPLIED OPTICS ? Vol. 46, No. 17 ? 10 June 2007
creased region of absorption and decreased scatter
can be seen in the images of the left breast in the
upper-outer quadrant. No obvious lesions are seen in
the right breast.
Figure 4(c) shows the optical images from a 36 yr
old woman diagnosed with breast cancer in the upper-
outer quadrant of her right breast. A MRI image indi-
cated that the lesion measured 5.3 cm ? 3.3 cm
? 5.3 cm. Only the right breast was scanned, due to
All subsequent patients were scanned using both
wavelengths, allowing us to derive oxygen saturation
and blood volume maps. The images in Fig. 5(a) are
from a 55 yr old woman with a 13 mm diameter in-
tracystic carcinoma in the upper-outer quadrant of
her right breast, in line with and 3 cm posterior to the
nipple. The MRI showed poor enhancement in the
center, indicating either calcifications or bleeding fol-
lowing aspiration or biopsy. Her left breast showed
areas of enhancement in the lower-inner quadrant of
the breast. Both breasts were scanned optically. The
optical images of the diseased breast show a region of
increased scatter and decreased blood oxygen satu-
contralateral breast. The images are displayed on the
with a carcinoma in the lower-inner side of her left
breast. In the x-ray mammogram, there was an ill-
defined 15 mm diameter mass in the left breast and
the right breast was normal. Ultrasound demon-
strated a 22 mm ? 17 mm ? 17 mm mass with vas-
cularization. There were also two prominent lymph
nodes in the upper-outer quadrant of the left breast.
The optical images show an area of increased blood
volume in the lower-inner quadrant of her left breast
that is not seen in the right breast.
Of the 18 scans of malignancies that resulted in
images, one case of ductal carcinoma in situ showed
no changes in optical properties. Of the remaining 17,
all show an increase in absorption, and all but one
show an increase in blood volume. This correlates
with what we know about angiogenesis occurring in
the tumor. Eight of the tumors show a decrease in
oxygen saturation in the region of the lesion, while
and scatter images of a 52 yr old woman with breast cancer. MRI showed a distortion in the right breast and increased stroma in the
right-upper and lower-outer quadrants of the breast. In the left breast there is a malignant mass in the left-upper-outer quadrant. (c)
Absorption and scatter images from a 36 yr old woman with a cancerous lesion in the upper-outer quadrant of her right breast.
(a) Absorption and scatter images from a 53 yr old woman with a large cancer in her left breast behind the nipple. (b) Absorption
10 June 2007 ? Vol. 46, No. 17 ? APPLIED OPTICS3633
the others show no change. Two of the tumors show
no change in scatter, ten show an increase, and five
show a decrease.
Fibroadenomas are benign breast lumps with low
vascularization and a distinctive histological appear-
ance. They are most commonly seen in younger
women and are usually assessed using mammogra-
phy or ultrasound. They are sometimes identified as
suspicious on mammograms, leading to unnecessary
biopsies. Our previous study using the fixed ring of
sources and detectors observed fibroadenomas in six
patients with varied levels of contrast , largely
consistent with another optical imaging study sug-
gesting that fibroadenomas display no unique optical
characteristics . Twelve women diagnosed with
fibroadenomas, with ages between 22 and 48 yr (mean
26 ? 6 yr), have so far been scanned using the liquid-
coupled interface at UCL.
Figure 6(a) shows optical images from a 29 yr old,
with a fibroadenoma in the lower-inner quadrant of
her right breast. The ultrasound demonstrated a
8 mm ? 11 mm ? 5 mm discrete mass. Both breasts
were optically scanned. A region of increased blood
volume and decreased oxygen saturation is seen in
the images of the right breast.
The optical images in Fig. 6(b) are from a 48 yr old
woman with a fibroadenoma reported to be in the
upper-outer quadrant of her right breast. The fibroad-
enoma measured 11 mm ? 5 mm ? 7 mm in size on
the ultrasound and was positioned 3 cm from the nip-
ple. The optical images show a region of decreased
scatter at the expected location of the lesion and
slightly increased blood oxygen saturation and blood
Twelve women with fibroadenomas were scanned,
resulting in 11 usable images. Of these images, four
showed no obvious change in optical properties in the
region of the lesion. Of the remaining seven, all
showed an increase in absorption and blood volume.
saturation, while six showed a decrease.
Cysts are fluid-filled sacs that develop within the
breast tissue and are generally easily diagnosed upon
ultrasound examination. The fluid in most cysts is
optically clear, though sometimes there may be blood
products present. Consequently cysts are commonly
revealed in optical images as regions of very low scat-
ter [20,33]. Three women aged between 27 to 57 yr
(mean 40 ? 15 yr) were optically scanned. One of
these women had cysts in both breasts, while another
patient had three partially aspirated galactoceals in
the same breast, which are milk-filled cysts that de-
velop in women who are breast-feeding. The third
patient had three cysts located very close to each
other in her left breast.
The breast images of the patient with the three
galactoceals exhibited regions of low scatter and high
absorption. While the low scatter was expected, the
high absorption may be due to minor bleeding follow-
ing the partial aspiration. Meanwhile optical images
of the women with three close cysts in her left breast
[Fig. 6(c)] show regions of very low oxygen saturation
and blood volume in the region of the group of cysts
but no associated decrease in scatter. No similar fea-
tures are seen in the contralateral breast. Unfortu-
nately, data recorded on the third woman with cysts
did not yield usable images.
carcinoma in the upper-outer quadrant of her right breast, in line with and 3 cm posterior to the nipple. (b) Scatter, blood volume, and blood
oxygen saturation images from a 45 yr old woman with a carcinoma in the lower-inner side of her left breast.
(a) Scatter, blood volume, and blood oxygen saturation images from a 55 yr old woman with a 13 mm diameter intracystic
3634 APPLIED OPTICS ? Vol. 46, No. 17 ? 10 June 2007
Four women with other benign conditions, aged be-
tween 30 and 44 yr (mean age 35 ? 6 yr), also took
part in the study. One woman had a diagnosis of
fibrocystic change (a condition that includes a range
of changes within the breast involving both the glan-
dular and stromal tissues). One of the women
scanned had a leaking breast implant, one suffered
from bilateral mastalgia, and one had a benign phyl-
lodes tumor. Scans of the patients with the leaking
implant and the benign phyllodes tumors did not
yield data of sufficient quality from which to recon-
struct images. No obvious lesions or changes in opti-
cal parameters were seen in the optical images from
the patient with mastalgia, which is consistent with
previously reported images . The optical images
from the patient with fibrocystic changes showed an
increase in scatter in the expected region of the
Other Benign Conditions
our initial results with the liquid-coupled system de-
scribed above demonstrate good sensitivity to a broad
range of lesions; at least one of the optical images
(scatter, blood volume, or blood oxygenation) usually
shows some evidence of a prediagnosed lesion, unless
it is very close to the chest wall, where the optical
sampling is still relatively sparse. The good sensitiv-
ity is underlined by the results of the image classifi-
cation shown in Table 1. Each set of breast images
has been assessed using a quantitative scoring
method proposed by Grosenick et al.  and Taroni
et al. . The scoring system is given in the footnote
of the table. Scores vary between 0—which is as-
signed when the lesion produces no observable fea-
ture in any of the images, and 5—for an image where
the dominant feature is due to the known lesion. An
X means that a scan was performed but an image was
not obtained due to an error occurring in the data
collection or in the calibration measurement, which
renders the data unusable. N?A means not applica-
ble, when a discrete change in optical properties is
not expected, such as in cases of mastalgia.
The optical images of breast cancer revealed a sig-
nificant increase in absorption and regional blood vol-
ume in most cases, presumably as a result of tumor
hypervascularization. Low blood oxygen saturation is
(b) Images of a 29 yr old with a fibroadenoma in the lower-inner quadrant of her right breast. The ultrasound revealed a lump of
8 mm ? 11 mm ? 5 mm. (c) Images of a 27 yr old woman with three cysts in her left breast. The cyst diameters obtained from ultrasound
were 29, 15, and 9 mm.
(a) Images of a 48 yr old woman with an 11 mm ? 5 mm ? 7 mm fibroadenoma in the upper-outer quadrant of her right breast.
10 June 2007 ? Vol. 46, No. 17 ? APPLIED OPTICS3635
also often observed at the location of the lesion, as is
a decrease in scatter. However, the latter may be due
to parameter cross talk (i.e., a change in one coeffi-
cient causing a small apparent change in the other).
Cross talk was observed with previous phantom stud-
ies, described by Yates et al. . In the case of the
patient with an intracystic carcinoma, there is a de-
crease in blood oxygen saturation, increased scatter,
and no evidence of increased blood volume [Fig. 5(a)].
Intracystic carcinomas are a rare form of breast can-
cer and do not behave like normal tumors. They are
a form of papillary carcinoma, are slow growing, and
are of low density . Their slow growth may sug-
gest that they do not develop the vasculature com-
monly associated with tumors.
change in blood volume is expected. Only half of the
fibroadenomas we scanned revealed any optical con-
trast in the region of the lesion relative to the sur-
rounding tissue and?or the contralateral breast,
which is consistent with results of other researchers
. All fibroadenomas that exhibited contrast indi-
cated an increased blood volume, and all but one
suggested a decrease in oxygen saturation. A study of
ex vivo tissue properties has predicted a possible in-
crease in absorption and low scatter for fibroadeno-
Cysts are typically low-scattering regions, so a de-
crease in blood volume and oxygen saturation in the
region of cysts, as well as a decrease in scatter, is
expected and is commonly observed. However, the
galactoceals were exhibited as regions of low scat-
ter and high absorption, probably due to aspiration-
induced bleeding, while the cysts in another patient
were revealed only as regions of low blood volume and
oxygen saturation, with no evidence of low scatter.
However, the scatter images for this 27 yr old subject
exhibited a very high degree of heterogeneity in both
breasts, which is commonly observed for younger
The observed maximum percent blood volume was
2.4% for tumors and 1.3% for benign lesions. These
values are equivalent to total hemoglobin ?Hbb? of 48
and 27 ?M, respectively (assuming 2 mM hemoglo-
bin concentration). Background values of ?0.7%–
1.5% are seen in both healthy and diseased breasts
?13–30 ?M?. Cysts show a decrease in blood volume
compared with the background tissue.
Because calculations include areas occupied by the
coupling fluid, the blood oxygen saturation values
displayed in the 3D images range from ?0.7% to over
100%. However, when examining the regions occu-
pied by the breast alone, the oxygen saturation val-
ues are within 56% ? 20%. Whereas blood volume
depends on total absorption, determination of oxy-
gen saturation requires more subtle spectroscopic
changes. Other researchers have found bulk healthy
tissue oxygen saturation values of 68% ? 8% ,
74% ? 9% , and 67%–70% .
A successful medical diagnostic technique usually
requires more than just good sensitivity; it is neces-
sary to distinguish between one type of pathology and
another and between benign and malignant lesions,
in particular. So far, our study has been exclusively
retrospective and has intended to provide a prelimi-
nary indication of the typical appearance of a broad
range of pathologies. The next important step is to
establish the specificity of the imaging technique and
determine if any optical signatures can be identified
that are indicative of the type of lesion. Ultimately
this requires a prospective study involving a suffi-
cient number of patients so that a meaningful statis-
tical analysis can be performed.
Better discrimination is being sought by improving
both the spatial resolution and the quantitative ac-
curacy of the images. Our efforts to achieve this are
largely focused on introducing prior structural infor-
mation into the image reconstruction and informa-
tion on the boundary between the breast and the
surrounding fluid in particular. One approach we are
currently investigating involves deliberately using a
coupling liquid with lower scatter and absorption in
order to clearly identify the boundary in the images.
This, combined with a knowledge of the breast vol-
ume derived from the displacement of fluid, enables
Table 1. Scoring Systema
Number of Cases and
Number of Images
Carcinomas21 (21)185, 5, 0, 3, 5, 5, X, X, 4, 4, 4,
X, 3, 5, 4, 5, 4, 5, 5, 5, 4
5, 4, X, X
5, 1, 4, 5, X, 4, 0, 0, 4, 3, 3, 5
4, 3, X
a5, contrast of lesion dominates; 4,
but still clearly identifiable; 2,
X,image not obtained.
contrast comparable to other inhomogeneities; 3,
weak contrast; 1, hardly perceivable inhomogeneity, only detected if position is known; 0,
contrast inferior to other inhomogeneities
3636APPLIED OPTICS ? Vol. 46, No. 17 ? 10 June 2007
the reconstructed volume to be confined to that occu-
pied by the breast.
Although improved quantitative accuracy is ex-
pected to help distinguish between lesions, a recent
optical imaging study reported by Pogue et al. 
has already identified a marked variation in the nor-
mal physiological parameters between patients and
suggested that it may be necessary to examine the
ing tissue rather than absolute values of optical pa-
Finally, optical imaging of the breast offers a po-
tentially powerful tool for monitoring long-term
changes in the breast, such as the study of tissue
recovering following surgery  or alterations dur-
ing and after treatment with chemotherapy or radio-
therapy. Since optical scanning is safe, repeated scans
over a long period can be performed without the expo-
sure risks associated with x-ray imaging.
The authors acknowledge the financial support of
Cancer Research UK and the Engineering and Phys-
ical Sciences Research Council.
1. Office for National Statistics, Mortality Statistics: Cause En-
gland and Wales 2004 (The Stationary Office, 2004).
2. L. Tabar, M.-F. Yen, B. Vitak, H.-H. Chen, R. A. Smith, and
S. W. Duffy, “Mammography service screening and mortality
in breast cancer patients: 20-year follow-up before and after
introduction of screening,” Lancet 361, 1405–1410 (2003).
3. R. W. Blamey, A. R. M. Wilson, and J. Patnick, “Screening for
breast cancer,” BMJ 689, 693 (2003).
4. M. D. Schnall, “Application of magnetic resonance imaging to
early detection of breast cancer,” Breast Cancer Res. 3, 17–21
5. D. D. Pavic, M. D. Koomen, C. D. Kuzmiak, and E. D. Pisano,
“Ultrasound in the management of breast disease,” Curr.
Womens Health Rep. 3, 156–164 (2003).
6. A. E. Cerussi, A. I. Berger, F. Bevilacqua, N. Shah, D.
Jakubowski, J. Butler, R. F. Holcombe, and B. J. Tromberg,
“Sources of absorption and scattering contrast for near infra-
red optical mammography,” Acad. Radiol. 8, 211–218 (2001).
7. S. B. Colak, M. B. van der Mark, G. W. Hooft, J. H. Hoogen-
raad, E. S. van der Linden, and F. A. Kuijpers, “Clinical optical
tomography and NIR spectroscopy for breast cancer detection,”
IEEE Quantum Electron. 5, 1143–1158 (1999).
8. V. G. Peters, D. R. Wyman, M. S. Patterson, and G. L. Frank,
“Optical properties of normal and diseased human breast tis-
sue in the visible and near infrared,” Phys. Med. Biol. 35,
9. X. Cheng, J. Mao, R. Bush, D. B. Kopans, R. H. Moore, and M.
Chorlton, “Breast cancer detection by mapping hemoglobin
concentration and oxygen saturation,” Appl. Opt. 42, 6412–
10. S. Jiang, B. W. Pogue, T. O. McBride, and K. D. Paulsen,
“Quantative analysis of near-infrared tomography: sensitivity
to the tissue-simulating precalibration phantom,” J. Biomed.
Opt. 8, 308–315 (2003).
11. A. Li, E. L. Miller, M. Kilmer, T. J. Brukilacchio, T. Chaves,
J. J. Stott, Q. Zhang, T. Wu, M. Chorlton, R. H. Moore, D. B.
Kopans, and D. A. Boas, “Tomographic optical breast imaging
guided by three-dimensional mammography,” Appl. Opt. 42,
12. C. H. Schmitz, D. P. Klemer, R. Hardin, M. S. Katz, Y. Pei,
H. L. Graber, M. B. Levin, R. D. Levina, N. A. Franco, W. B.
Solomon, and R. L. Barbour, “Design and implementation of
dynamic near-infrared optical tomographic imaging instru-
mentation for simultaneous dual-breast measurements,” Appl.
Opt. 44, 2140–2152 (2005).
13. Y. Young, M. Shnall, S. Zhao, S. Orel, C. Xie, S. Nioka, B.
Chance, and A. Solin, “Optical imaging of breast tumour by
means of continuous waves,” Adv. Exp. Med. Biol. 441, 227–
14. J. P. Culver, R. Choe, M. J. Holboke, L. Zubkov, T. Durduran,
A. Slemp, V. Ntziachristos, B. Chance, and A. G. Yodh, “Three-
dimensional diffuse optical tomography in the parallel plane
transmission geometry: evaluation of a hybrid frequency do-
main?continuous wave clinical system for breast imaging,”
Med. Phys. 30, 235–247 (2003).
15. B. W. Pogue, M. Testorf, T. McBride, U. Osterberg, and K.
Paulsen, “Instrumentation and design of a frequency-domain
diffuse optical tomography imager for breast cancer detection,”
Opt. Express 1, 391–403 (1997).
16. D. Grosenick, H. Wabnitz, K. T. Moesta, J. Mucke, M. Moller,
C. Stroszczynski, J. Stobel, B. Wassermann, P. M. Schlag, and
H. Rinneberg, “Concentration and oxygen saturation of hemo-
globin of 50 breast tumours determined by time-domain opti-
cal mammography,” Phys. Med. Biol. 49, 1165–1181 (2004).
17. P. Taroni, G. Danesini, A. Torricelli, A. Pifferi, L. Spinelli, and
R. Cubeddu, “Clinical trial of time-resolved scanning optical
mammography at 4 wavelengths between 683 and 975 nm,”
J. Biomed. Opt. 9, 464–473 (2004).
18. T. D. Yates, J. C. Hebden, A. P. Gibson, N. L. Everdell, S. R.
Arridge, and M. Douek, “Optical tomography of the breast
using a multi-channel time-resolved imager,” Phys. Med. Biol.
50, 2503–2517 (2005).
19. B. W. Pogue, S. Jiang, H. Dehghani, C. Kogel, S. Soho, S.
Srinivasan, X. Song, T. D. Tosteson, S. P. Poplack, and K. D.
Paulsen, “Characterization of hemoglobin, water and NIR
scattering in breast tissue: analysis of intersubject variability
and menstrual cycle changes,” J. Biomed. Opt. 9, 541–552
20. D. Grosenick, K. T. Moesta, M. Möller, J. Mucke, H. Wabnitz,
B. Gabauer, C. Stroszczynski, B. Wassermann, P. M. Schlag,
and H. Rinneberg, “Time-domain scanning optical mammog-
raphy: I. Recording and assessment of mammograms of 154
patients,” Phys. Med. Biol. 50, 2429–2449 (2005).
21. D. Grosenick, H. Wabnitz, K. T. Moesta, J. Mucke, P. M.
Schlag, and H. Rinneberg, “Time-domain scanning optical
mammography: II. Optical properties and tissue parameters of
87 carcinomas,” Phys. Med. Biol. 50, 2451–2468 (2005).
22. P. Taroni, A. Torricelli, L. Spinelli, A. Pifferi, F. Arpaia, G.
Danesini, and R. Cubeddu, “Time-resolved optical mammog-
raphy between 637 and 985 nm: clinical study on the detection
23. L. Spinelli, A. Torricelli, A. Pifferi, P. Taroni, G. Danesini, and
R. Cubeddu, “Characterization of female breast lesions from
Med. Biol. 50, 2489–2502 (2005).
24. T. O. McBride, B. W. Pogue, E. D. Gerety, S. B. Poplack, U. L.
Osterberg, and K. D. Paulsen, “Spectroscopic diffuse optical
tomography for the quantitative assessment of hemoglobin
concentration and oxygen saturation in breast tissue,” Appl.
Opt. 38, 5480–5490 (1999).
25. H. Dehghani, B. W. Pogue, S. P. Poplack, and K. D. Paulsen,
“Multiwavelength three-dimensional near-infrared tomogra-
phy of the breast: initial simulation, phantom and clinical
results,” Appl. Opt. 42, 135–145 (2003).
26. B. W. Pogue, S. Jiang, X. Song, S. Srinivasan, H. Dehghani,
K. D. Paulsen, T. D. Tosteson, C. Kogel, S. Soho, and S. P.
Poplack, “Near-infrared scattering spectrum differences be-
tween benign and malignant breast tumours measured in vivo
10 June 2007 ? Vol. 46, No. 17 ? APPLIED OPTICS3637
with diffuse tomography,” in OSA Biomedical Topical Meeting Download full-text
27. R. Wiener, R. Choe, A. Corlu, K. Lee, S. M. Crinivas, J. R.
Saffer, R. Freifelder, J. S. Karp, A. G. Yodh, and S. D. Konecky,
“Diffuse optical tomography and positron emission tomogra-
phy of human breast,” in Biomedical Optics Topical Meeting
28. R. Choe, A. Corlu, K. Lee, T. Durduran, S. D. Konecky, M.
Grosicka-Koptyra, S. R. Arridge, B. J. Czernieki, D. L. Fraker,
A. DeMichele, B. Chance, M. A. Rosen, and A. G. Yodh, “Dif-
fuse optical tomography of breast cancer during neoadjuvant
chemotherapy: a case study with comparison to MRI,” Med.
Phys. 32, 1128–1139 (2005).
29. X. Intes, J. Ripoll, Y. Chen, S. Nioka, A. G. Yodh, and B.
Chance, “In vivo continuous-wave optical breast imaging en-
hanced with Indocyanine Green,” Med. Phys. 30, 1039–1047
30. O. Zhang, T. I. Brukilacchio, A. Li, J. J. Scott, T. Chaves, E.
Hillman, T. Wu, M. Chorlton, E. Rafferty, R. H. Moore, D. B.
Kopans, and D. A. Boas, “Coregistered tomographic x-ray and
optical breast imaging: initial results,” J. Biomed. Opt. 10,
31. B. J. Tromberg, A. Cerussi, N. Shah, M. Compton, A. Durkin,
D. Hsiang, J. Butler, and R. Mehta, “Diffuse optics in breast
cancer: detecting tumours in pre-menopausal women and
monitoring neoadjuvant chemotherapy,” Breast Cancer Res. 7,
32. F. E. W. Schmidt, M. E. Fry, E. M. C. Hillman, J. C. Hebden,
and D. T. Delpy, “A 32-channel time-resolved instrument for
medical optical tomography,” Rev. Sci. Instrum. 71, 256–265
D. W. Chicken, M. Douek, and M. R. S. Keshtgar, “Monitoring
recovery after laser surgery of the breast with optical tomog-
raphy: a case study,” Appl. Opt. 44, 1898–1904 (2005).
34. T. D. Yates, J. C. Hebden, A. P. Gibson, L. Enfield, N. L.
Everdell, S. R. Arridge, and D. T. Delpy, “Time-resolved optical
mammography using a liquid coupled interface,” J. Biomed.
Opt. 10, 054011 (2005).
35. S. R. Arridge, J. C. Hebden, M. Schweiger, F. E. W. Schmidt,
M. E. Fry, E. M. C. Hillman, H. Dehghani, and D. T. Delpy, “A
method for 3D time-resolved optical tomography,” Int. J. Im-
aging Syst. Technol. 11, 2–11 (2000).
36. J. Schoberl, “NETGEN—an advancing front 2d?3D-mesh gen-
erator based on abstract rules,” Comput. Vision Sci. 1, 41–52
37. S. J. Matcher, M. Cope, and D. T. Delpy, “Use of the water
absorption spectrum to quantify tissue chromophore concen-
tration changes in near-infrared spectroscopy,” Phys. Med.
Biol. 38, 177–196 (1993).
38. E. M. C. Hillman, “Experimental and theoretical investiga-
tions of near-infrared tomographic imaging methods and clin-
ical applications,” Ph.D. thesis (University of London, 2002).
39. D. Grosenick, K. T. Moesta, H. Wabnitz, J. Mucke, C. Stroszc-
zynski, R. Macdonlad, P. M. Schlag, and H. Rinneberg,
“Time-domain optical mammography: initial clinical results
on detection and characterization of breast tumours,” Appl.
Opt. 42, 3170–3186 (2003).
40. M. A. Peters, D. R. Voegeli, and K. A. Scanlan, Breast Imaging
41. S. Fantini, S. A. Walker, M. A. Franeschini, M. Kaschke, P.
Schlag, and K. T. Moesta, “Assessment of the size, position and
optical properties of breast tumors in vivo by noninvasive op-
tical properties,” Appl. Opt. 37, 1982–1989 (1998).
3638APPLIED OPTICS ? Vol. 46, No. 17 ? 10 June 2007