2.1.

Biomedical Background

2.1.1.

Overall structure and composition of normal breast tissue

Optical mammography and diffuse optical spectroscopy of the human breast probe the absorption, scattering, and fluorescence properties of various components of healthy and diseased breast tissue. The human female breast consists mainly of glandular, adipose, and connective tissue, together with blood and lymphatic vessels, and it contains several simple mammary glands (lobes), each draining through a separate lactiferous duct. The lobes branch into several lobules consisting of intralobular ducts separated by rather loose (intralobular) connective tissue, containing microvasculature and small lymphatic channels. Each intralobular ductal tree terminates in a cluster of alveoli that will differentiate to produce milk on exposure to lactogenic hormones. The intralobular connective tissue and ductal network are surrounded by the interstitial connective tissue, being dense, less cellular, containing variable proportions of adipose tissue and extracellular matrix (ECM), and representing over 80% of the human breast volume.⁹

Besides blood and lymphatic vessels, the connective tissue (stroma) consists of various stromal cells and the ECM (see Fig. 1), providing a scaffold for stromal cells (e.g., fibroblasts, adipocytes, cells of the immune system, e.g., lymphocytes, macrophages).¹⁰ The ECM is composed of water, proteins, and polysaccharides.¹¹ Proteoglycans (PGs) and fibrous proteins (e.g., collagens, fibronectin, elastins) are the two main classes of macromolecules of the ECM, forming an intricate interlocking mesh.¹¹ The polysaccharides glycosaminoglycans (GAGs) are usually attached to ECM proteins to form PGs. PGs fill the majority of the extracellular interstitial space within the tissue in the form of a hydrated gel.¹¹ PGs have a net negative charge that attracts sodium ions and in consequence water, keeping the ECM and resident cells hydrated. Unlike other GAGs, the polysaccharide hyaluronic acid (or “hyaluronan”) contained in the ECM is not bound to matrix proteins. Hyaluronic acid in the extracellular space confers upon tissues the ability to resist compression providing a counteracting swelling force by absorbing significant amounts of water. Collagens constitute the main structural element of the ECM and, together with other fibrous proteins, provide mechanical strength and elasticity to the tissue.¹¹ Nonactivated tissue fibroblasts secrete various ECM proteins (collagens and elastins) and PGs including hyaluronic acid.¹¹ The ECM, containing various peptides, e.g., growth factors and enzymes, is being constantly remodeled for tissue homeostasis. Such remodeling is regulated by a careful balance between intracellular matrix synthesis, secretion, modification, e.g., crosslinking of collagens by lysyl oxidases (LOX) and enzymatic degradation, e.g., by matrix metalloproteases (MMPs).¹² The ECM not only represents scaffolding for the stromal cells but also conveys biochemical and biophysical signals to cells.

Fig. 1

Structure and function of ECM in (a) normal and (b) tumorous glandular epithelial tissue (e.g., breast). (Reprinted with permission from Frantz et al.¹¹)

2.1.3.

Abnormal extracellular matrix of solid tumors

It has long been known that tumor-derived ECM is biochemically distinct in its composition compared with normal ECM (Fig. 1).¹² Breast cancer progression is associated with changes in ECM composition, with inflammatory cell infiltration, and differentiation of fibroblasts.⁹ Tumor-derived ECM differs from that of the host tissue, owing to the disruption of the balance between ECM synthesis and secretion and owing to alterations in the normal levels of matrix-remodeling enzymes, such as MMP and LOX.¹² The ECM remodeling observed in tumors includes increased deposition of collagen, fibronectin, PGs, substantial MMP-dependent cleavage, and increased levels of LOX-dependent matrix crosslinking. There is evidence for an increased deposition of ECM in hypoxic tumor regions.³² Differentiated fibroblasts (e.g., myofibroblasts) deposit large amounts of ECM proteins.¹¹ The expression of MMPs is often highly upregulated in solid tumors with MMPs produced by myofibroblasts and tumor cells.¹¹ ECM deposition and leukocyte infiltration are often very pronounced at the tumor-stroma border.¹⁰ The majority of increased tumor and adjacent tissue stiffness occur as a result of increased ECM deposition.¹² Furthermore, increased LOX activity results in increased ECM stiffness.¹² As illustrated in Fig. 2, the expanding tumor mass exerts compressive stress (tissue solid stress) on the surrounding tissue, on the ECM, on intratumoral vasculature and lymphatics.^9,25 NIR imaging is directly affected by the modifications of the ECM occurring in tumors, resulting in changes of the tissue absorption spectrum. For example, additional deposition of collagen increases long-wavelength (1060 nm) absorption, while tissue solid stress on tumor vasculature may affect blood flow, resulting in changes of oxyhemoglobin concentration and corresponding modifications of the absorption spectrum.

Fig. 2

Extrinsic and intrinsic forces on (a) normal and (b) tumorous glandular epithelial tissue (e.g., breast). (Reprinted with permission from Frantz et al.¹¹)

2.2.

Historical Development of Optical Breast Imaging

First attempts to see suspicious lesions in the female breast using visible light were reported by Cutler in 1929.^33,34 The breast was transilluminated by holding a small size powerful lamp against the lower surface of the breast and by observing shadows of the light on the upper breast side by eye that arose from high-absorbing tissue structures. In the 1970s and 1980s, transillumination imaging of the breast was further developed by using improved light sources in combination with light detection by sensitive films or video cameras.^35–39 However, several clinical studies showed that sensitivity and specificity of the so-called lightscanning or “diaphanography” method were low in comparison to x-ray mammography.^40–43 The lightscanning approach was very simple to apply, but it had major intrinsic disadvantages, including the absence of discrimination between scattering and absorption of the tissue and the limited exploitation of spectral information.

New efforts in optical breast imaging started in the 1990s. At that time, mathematical models of light propagation in tissue became available that permitted separation of the scattering and the absorption properties of the tissue. In this way, absorption properties could be exploited by the methods of near-infrared spectroscopy to determine the composition of the tissue, i.e., to measure the concentrations of the main tissue absorbers including blood oxygen saturation. Basic concepts of optical breast imaging developed at this time are valid until today.

2.3.

Classification of Optical Breast Imagers

Normally, optical images of the breast are obtained by switching through a sufficient number of point-like light sources realized by illuminated optical fibers and, for each source, detecting the light at one or several suited detector positions. The instruments for optical breast imaging and spectroscopy can be classified with respect to the temporal profile of the laser radiation employed and according to their measurement geometry. Furthermore, a few instruments have been designed to detect fluorescence arising from an exogenous contrast agent. The three well-known groups of instruments are time-domain, frequency-domain, and continuous-wave (CW) systems. Briefly, time-domain systems measure the broadening of short (picosecond) laser pulses after propagation through the tissue. Absorption coefficients ${\mu }_{\mathrm{a}}$ and reduced scattering coefficients ${\mu }_{\mathrm{s}}^{\prime }$ of the tissue can be derived from the analysis of the detected pulse shape using an appropriate model of photon propagation. These coefficients characterize the tissue volume sampled by photons depending on the particular source–detector combination. For source–detector separations of a few centimeters, the width of the broadened pulses amounts typically to a few nanoseconds. A well-suited technique for the detection of these pulses is time-correlated single-photon counting. In the past, hardware for time-domain systems was expensive. However, it is now becoming possible to perform time-resolved data acquisition with more cost-competitive instrumentation.

Frequency-domain systems use intensity-modulated laser radiation and measure its demodulation and phase shift after passing through tissue. In principle, it is sufficient to perform such measurements at one modulation frequency to separate the absorption and the scattering coefficient of the sampled tissue volume. Typically, a modulation frequency of about 100 MHz is chosen. Frequency-domain measurements correspond to investigating the Fourier spectrum of time-domain measurements at one frequency. By employing several modulation frequencies up to at least 1 GHz, the information content of frequency domain measurements increases and becomes comparable to that of typical time-domain investigations.

CW systems use continuously emitting lasers or light-emitting diodes at several near-infrared wavelengths, or broadband light sources. The instruments detect the attenuation of the transmitted light. More specifically, the CW technique yields the attenuation coefficient $\kappa =\sqrt{{\mu }_{\mathrm{a}}/D}$, which is a combination of the absorption coefficient ${\mu }_{\mathrm{a}}$ and the diffusion coefficient $D=1/3{\mu }_{\mathrm{s}}^{\prime }$. To obtain absolute values of the absorption coefficients, prior knowledge on the reduced scattering coefficient is required, e.g., from additional investigations with the frequency-domain or time-domain technique or by using representative values from literature. However, because light scattering dominates over absorption within the optical window, the latter approach may lead to systematic errors. Alternatively, it is possible to assess both absorption and scattering properties by performing CW measurements at more than one source–detector separation, which has been used, e.g., to obtain properties of small (almost homogeneous) tissue regions.⁴⁴ Generally, it is also possible to exploit spectral features for the separation of absorption and scattering by using specifically selected optical wavelengths.⁴⁵ The main advantage of the CW approach is that light sources and detectors are comparably cheap. Therefore, such systems can be equipped with a large number of source and detector positions at low costs.

The geometry for optical breast imaging can be divided into three main groups (Fig. 3). In one approach, the breast is compressed between two parallel plates. This geometry is very close to the concept of x-ray mammography. It offers a simple way to compare an optical transillumination image with a corresponding x-ray mammogram. A second approach consists of investigating the freely pending breast with the woman being in prone position. This geometry has high similarity with MRI of the breast. The third basic approach is the use of a handheld probe that can be positioned at selected locations of the breast or moved over its surface similar to a breast ultrasound detector. It is obvious that these three geometries offer the possibility for combining optical imaging with x-ray, MR, or US breast imaging, which was pursued by several groups, mainly to gain prior information on spatial tissue composition, lesion size, and position when analyzing their optical and functional properties.

Fig. 3

Examples of measurement geometries: (a) parallel-plate geometry with moving source fiber and few detector fibers in transmission and reflection,⁴⁶ (b) parallel-plate geometry with fixed source and detector fibers (left) and CCD camera detection (right) (reprinted with permission from Choe et al.⁴⁷), (c) freely pending breast geometry with 255 source and 255 detector fibers (© 1999 IEEE. Reprinted with permission from Colak et al.⁴⁸), (d) freely pending breast with 32 sources and 32 detectors (reprinted with permission of Optical Society of America from Enfield et al.⁴⁹), (e) circular arrangement of 48 fibers in three rings with 16 translation stages to bring the fibers in contact with the tissue (reprinted with permission from Pogue et al.⁵⁰), and (f) handheld probe (reprinted with permission from O’Sullivan et al.⁵¹).

3.1.

Overview on Optical Breast Imagers

Figure 3 shows several examples of the measurement geometry and the source–detector arrangements that were realized in instrumentation for clinical studies on optical breast imaging.

A simple way to realize the compression geometry is the application of two transparent plates with variable distance [Fig. 3(a)]. Optical transillumination images of the breast can be obtained by moving a source fiber on one side and a detector fiber on the other side in tandem. Such systems have first been developed by the German companies Siemens and Carl Zeiss.^52,53 Both systems were equipped with frequency-domain instrumentation. Later, the Physikalisch-Technische Bundesanstalt⁵⁴ and the Politecnico di Milano⁵⁵ used this geometry to build-up the first time-domain scanning optical mammographs. At Tufts University, the compression geometry has been employed more recently in a hybrid instrument combining frequency-domain measurements at a few near-infrared wavelengths with broadband CW spectroscopy⁵⁶ and in an instrument using solely CW radiation.⁵⁷ With a scanning step size of a few millimeters or even less, typically, more than 1000 source–detector combinations (scan positions) are sampled.

As an example for the parallel plate geometry, the schematic in Fig. 3(a) shows the latest source–detector fiber arrangement of the PTB instrument with several detection fibers in transmission and also detection fibers in reflection.⁴⁶ The various fibers in transmission permit the detection of transillumination images at the implemented optical wavelengths under different projection angles, which can be exploited to reconstruct three-dimensional (3-D) images of the tissue.^46,58 The additional detection fibers in reflection can be used to improve the 3-D resolution close to the surface, as shown in phantom experiments.⁵⁹ However, with a transparent plate, only distances from the source fiber below, typically, 1 cm can be exploited due to multiple reflections occurring within the plate. Figure 3(a) shows a second source fiber (dashed line), that was added to the PTB optical mammograph for fluorescence measurements. The tandem scanner concept with several offset fibers in transmission is also employed in the latest version of the CW optical mammograph developed at Tufts University.⁵⁷

Figure 3(b) illustrates the parallel-plate instrument developed at University of Pennsylvania. This hybrid CW-frequency domain device uses the compression geometry with a source–detector arrangement optimized for tomographic reconstruction. The patient is in the prone position with both breasts hanging in a tank filled with a scattering fluid. The opaque compression plate on the left-hand side is equipped with 45 fixed source fibers and 9 fixed detector fibers. The frequency-domain approach is employed to measure the diffusely reflected light at four wavelengths. The CCD camera on the right-hand side in Fig. 3(b) is used to measure the diffuse transmittance of the breast by the CW approach at up to six wavelengths.⁴⁷ This device was extended with an option for fluorescence measurements using indocyanine green (ICG) as contrast agent.⁶⁰ There were also two time-domain devices developed that use the parallel-plate geometry with matching fluid chamber, the commercial system Softscan of Advanced Research Technologies (ART) Inc., Montreal,⁶¹ and a laboratory prototype at the Physikalisch-Technische Bundesanstalt Berlin.⁶² Both these devices employed the concept of scanning source and detector fibers along transparent plates. Depth resolution was achieved by either offset detection channels in transmission similar to Fig. 3(a) or by two CCD cameras similar to Fig. 3(b). The PTB device was also capable of performing fluorescence investigations. The company DOBI Medical International Inc. developed the parallel-plate device ComfortScan for imaging pressure-induced changes in the blood oxygen saturation of tumors. This instrument uses flat-field illumination at 640 nm by LEDs and CCD camera detection of the transmitted light.⁶³

Figures 3(c) and 3(d) show examples of devices with a cup-like chamber for investigations of the freely pending breast. The cup in Fig. 3(c) belongs to the CW instrument built by Philips, Eindhoven. In this device, a total of 255 source fibers and 255 interleaved detection fibers were used to reconstruct the tissue attenuation coefficient at three optical wavelengths.⁴⁸ The same geometry was used in the fluorescence instrument developed about 10 years later.⁶⁴ The time-domain instrument in Fig. 3(d) was built at the University College London. It is equipped with 32 source fibers and 32 detection fiber bundles and provides 3-D images of the absorption and scattering properties at two wavelengths.⁴⁹ Both these devices with cup geometry use a scattering fluid to get high quality optical coupling and to work with a well-defined geometry for reconstruction of the optical properties. In order to account for breasts of different sizes, cups with different diameters can be used.

At Dartmouth College, a frequency-domain instrument with six optical wavelengths was developed for investigations on the freely pending breast that does not need a scattering fluid [see Fig. 3(e)]. Here, 48 fibers, which are arranged on three rings, are brought in contact with the tissue under slight pressure.⁵⁰ In a newer version, the wavelength range of the device was extended up to 948 nm employing CW lasers.⁶⁵ The breast imager developed by NIRx Technologies uses the same principle of pressure-induced optode contact. From these four wavelengths, the CW instrument permits simultaneous investigations on both breasts of the patient. It has been designed for dynamic investigations to record the physiological response of the breast tissue to specific interventions, such as the Valsalva maneuver or to dynamically observe the effect of a contrast agent bolus.^66,67 The company, Imaging Diagnostics Systems, Inc., Fort Lauderdale, followed the principle of CT scanners and developed devices in which the freely pending breast of the patient in prone position is scanned by moving a laser beam and a detector array circularly around the tissue. By changing vertical positions, contiguous slices of the breast are acquired at one optical wavelength.^68,69 Hereby, the detectors are not in contact with tissue. An extended version of these devices was prepared for fluorescence investigations.

Detection limits of diffuse optical tomography systems were investigated theoretically and numerically.^70,71 The method consists of analyzing raw numerical phantom data by means of a chi-square test, obtained from forward simulations together with a realistic noise model, derived from the system hardware. Both parallel-plate and cup geometries were compared with respect to detection limits of heterogeneities at various positions within the tomographic volume investigated. In cup geometry, low detection sensitivity was obtained at the upper center of the cup [cf., Figs. 3(c) and 3(d)], i.e., close to the chest wall, where the tissue is exclusively sampled by source–detector combinations with large separations. In slab geometry, detection sensitivity shows only small variations between the outer and the inner tissue regions, since the breast is sampled with constant source–detector separation. Generally, for smaller breast sizes, lesions of 5-mm diameter could be detected in almost all parts of the compressed breast (parallel-plate geometry) and in the outer parts of the uncompressed breast (cup geometry), whereas for larger breasts, the detection limit moved toward 7.5-mm lesion size when a lesion-to-background absorption contrast of $2:1$ was assumed.

The third group of optical breast imagers comprises devices with handheld probes. As an example, Fig. 3(f) shows the probe of the diffuse optical spectroscopic imaging (DOSI) device developed at the University of California, Irvine.^51,72 This device combines the frequency-domain and the CW approach. It contains two source–detector pairs. The first one is used to perform frequency-domain measurements at six optical wavelengths ranging from 650 to 860 nm. In contrast to the frequency-domain instruments discussed above, the modulation frequency is varied here from 50 to 1000 MHz. Therefore, the amount of information obtained for the sampled tissue volume is comparable to that of time-domain approaches discussed above. The second source–detector pair is connected to a tungsten halogen white-light source and a spectrometer to record broadband reflectance spectra from 650 to 1000 nm. The scatter power law:

At the University of Pennsylvania, a handheld CW imager was used for breast imaging with a three wavelengths LED source and eight surrounding silicon diode detectors 4-cm apart from the source.⁷⁴ Another CW device was developed by ViOptix, Inc., Fremont.⁷⁵ Handheld probes have also been used for blood flow characterization in breast tissue and tumors employing diffuse correlation spectroscopy (DCS; cf. Sec. 3.5). A detailed overview about the various handheld NIR devices for breast imaging and other applications can be found in the review by Erickson and Godavarty.⁷⁶

3.2.

Multimodality Imaging

Several groups have combined optical breast imaging with other clinical breast imaging modalities. Hereby, the conventional clinical modality provides structural information about the breast tissue that is exploited in the reconstruction of the optical and physiological properties. In this way, problems of low spatial resolution and diffuse blurring of reconstructed optical data can be overcome, and optics can provide, e.g., metabolic information about lesions not accessible by the conventional modalities. At the University of Connecticut, a handheld probe device has been developed for combined investigations of breast tumors by NIR light and ultrasound. This probe contains a commercial US detector array together with 12 source and 8 detection fibers for diffuse reflectance frequency-domain measurements at several source–detector distances.⁷⁷ In the present version, three optical wavelengths between 660 and 830 nm are used.⁷⁸

The combination of optical and ultrasound measurements in a handheld probe is technically simple, whereas the combination of optical and MR measurements is more challenging. To avoid interference of the optoelectronic components with the high magnetic field, long fibers or fiber bundles have to be used to deliver and collect the light inside the MR bore. Furthermore, the restricted space inside the bore limits the number of source and detector fibers that can be installed. The first demonstration of concurrent optical and MR imaging of breast tissue was reported by Ntziachristos and Ma,⁷⁹ who placed a parallel-plate patient interface for time-resolved transmittance measurements inside the MR tomograph. This instrument allowed comparison of contrast-enhanced optical absorption imaging with ICG versus dynamic contrast enhanced MR imaging⁸⁰ and provided intrinsic hemoglobin and oxygen saturation contrast ${\mathrm{StO}}_2$ for malignant and benign breast lesions.⁸¹ At Dartmouth College, a 16 fiber ring holder interface was applied inside the MR scanner for simultaneous frequency-domain optical and MR interventions.⁸² During reconstruction of the optical images, the structural information from the MR investigations was used as prior information. This group also reported successfully incorporating water and fat information from MR imaging to improve the accuracy of the reconstructed hemoglobin concentration.⁸³

At the Massachusetts General Hospital, a combined optical and x-ray breast imager was developed that uses the parallel-plate geometry. Optical measurements are performed in transmittance by a hybrid frequency-domain and CW approach using a source–detector grid designed for 3-D reconstruction of the optical properties. The device can also be applied for functional monitoring of the breast tissue.⁸⁴

3.3.

Data Analysis and Reconstruction

One aim of data analysis is the generation of optical mammograms, i.e., the generation of 2-D or 3-D images that display lesions and structures inside the breast with high contrast. Another aim is the determination of the optical and physiological parameters of lesions and of healthy breast tissue. Hereby, the generation of images is not really required. Generally, optical mammograms can provide such values. However, the accuracy of values characterizing lesions is often limited by the diffuse blurring and by partial volume effects. These limitations can be overcome by using prior information about the size and location of the lesions together with adequate heterogeneous models.

The generation of optical mammograms for the circular tomographic geometry requires a reconstruction of the optical or physiological properties of the breast. For the parallel-plate devices, reconstruction can also be applied, but it is not mandatory. Since the distance between source fiber and the detection fiber in scanning devices with parallel-plate geometry is the same at all scan positions, data at each scan position can be analyzed independently.

Figure 4 illustrates some general features of optical mammograms obtained by the various data analysis methods. Transillumination images from parallel-plate instruments [Figs. 4(a)–4(c)] often exhibit not only the carcinoma but also superficial blood vessels and other localized regions of high vascularization. Reconstructed slices from parallel-plate instruments show more blurred structures [Figs. 4(d) and 4(e)]. Similarly, reconstructed images from circular tomographic measurements [Figs. 4(f) and 4(g)] display the carcinoma often within a nonuniform background with additional structures showing high correlation to the source and detector fiber positions.

Fig. 4

Examples of absorption and HbT optical mammograms of patients with carcinomas (labeled by arrows or lines): (a) late-gate transillumination image at 670 nm from a parallel-plate instrument displaying a 3.5-cm carcinoma (from Rinneberg et al.⁸⁵); (b) transillumination image at 785 nm from a parallel-plate instrument displaying difference in absorption coefficient $\mathrm{\Delta }{\mu }_{\mathrm{a}}$ from average background analyzed within an inhomogeneous model (from Quarto et al.⁸⁶); (c) HbT transillumination image with 3.9-cm carcinoma from a CW parallel-plate instrument (reprinted with permission from Anderson et al.⁸⁷), (d) reconstructed slice of relative total hemoglobin concentration (designated rTHC) with a 2.2-cm carcinoma from a parallel-plate instrument with matching fluid (reprinted with permission from Choe et al.⁴⁷), (e) reconstructed slice of HbT with a 2.5-cm carcinoma obtained with prior knowledge from x-ray mammography (reprinted with permission of the Radiological Society of North America from Fang et al.⁸⁸), (f) reconstructed slice (HbT map) with a carcinoma about 4 cm in size (reprinted with permission from Wang et al.⁶⁵), and (g) reconstructed absorption image with carcinoma (indicated by yellow circle) from a CW tomographic instrument with matching fluid (image reproduced from van de Ven et al.⁸⁹).

3.4.

Broadband Spectroscopy

Most imagers display mammograms at a limited number of discrete wavelengths, with a minimum of three wavelengths. Operating at seven selected wavelengths (635 to 1060 nm) and applying a perturbative approach for data analysis, it was possible to generate at each wavelength absorption difference $\mathrm{\Delta }{\mu }_{\mathrm{a}}$ maps, with respect to the average absorption of the same breast. From these data, corresponding concentration difference $\mathrm{\Delta }C$ maps for each tissue constituent (HbR, ${\mathrm{HbO}}_2$, water, lipid, and collagen) with respect to the average concentration in the same breast were deduced.⁸⁶ As an example, in Fig. 5, the adipose nature of the breast is apparent from the high lipid content in the entire breast (in agreement with what was seen in the corresponding x-ray image). As expected based on physiology, the residual of the mammary gland (upper quadrants), which causes marked x-ray attenuation, is characterized by high content of oxygenated blood, water, and also collagen. A blood vessel is clearly detected in the maps of HbT and ${\mathrm{HbO}}_2$ content, similar to the x-ray image. A carcinoma is characterized by strong hemoglobin, water, and collagen absorption, as confirmed by quantitative estimates in the reference area (see Fig. 5).

Fig. 5

Concentration difference maps $\mathrm{\Delta }C$ of the main breast constituents (HbR, ${\mathrm{HbO}}_2$, HbT, lipid, water and collagen) of the craniocaudal view of the left breast of a patient with a 25-mm invasive ductal carcinoma in the retroareolar area (compressed breast thickness 46 mm); x-ray image (top left) and reference image (bottom left) show carcinoma (red arrow), and lesion area (black square) selected to estimate lesion properties, respectively. The color bar range for $\mathrm{\Delta }C$ maps depends on the particular constituent, for details, see Quarto et al.⁸⁶ (Reprinted with permission of Optical Society of America from Quarto et al.⁸⁶)

In contrast to imagers employing discrete wavelengths, the handheld device developed at the University of California, Irvine, exploits CW broadband operation with halogen lamp illumination covering the range of 650 to 1000 nm together with frequency-domain diode lasers at discrete wavelengths (see Sec. 3.1).^72,98 Tufts University has also recently developed a scanning system that combines CW broadband Xenon arc lamp illumination and detection with a CCD camera coupled to a spectrograph. Spectral images of the compressed breast are acquired every 8 nm between 650 and 950 nm.⁸⁷ Although only wavelengths shorter than 850 nm are exploited for data analysis, thus excluding strong absorption peaks of major tissue constituents, the spectral differences among tissue absorbers proved significant enough to allow the estimate of oxy- and deoxyhemoglobin, water, and lipid content.

Some breast studies (e.g., assessment of breast density) aim at quantifying the average properties of breast tissue. Thus, they may not require imaging capabilities. The set-up developed by Lilge and coworkers belongs to the latter category.¹¹⁶ A halogen lamp was used as a broadband light source, and light transmitted through the compressed breast was collected coaxially and detected by a spectrophotometer coupled to a 2-D CCD camera to achieve continuous operation between 625 and 1060 nm with 3-nm resolution. Measurements were typically performed at four positions on the compressed breast, thus providing some information on breast heterogeneity. Principal component analysis was widely applied to interpret the transmittance spectra, showing that four principal components account for most of tissue variability.¹¹⁶ Recently, a spectrally constrained global fitting procedure based on photon diffusion has been introduced to estimate the concentrations of ${\mathrm{HbO}}_2$, HbR, the percentage content of lipid and water.¹¹⁷

Another broadband system was also applied to investigate breast tissue composition. It operated in the time domain, scanning continuously between 610 and 1050 nm. A synchronously pumped dye laser and an actively mode-locked Ti:Sapphire laser provided picosecond pulses over the range of 610 to 700 nm and 700 to 1050 nm, respectively, and time-correlated single-photon counting was used to detect the time distribution of either the transmitted or the reflected signal.¹¹⁸ This was a complex table–top laboratory set-up, not suitable for use in a clinical environment. However, it allowed investigation of breast heterogeneity^119,120 and changes in breast tissue with the menstrual cycle,¹²¹ assessing HbT and ${\mathrm{StO}}_2$, water and lipid content, and scattering properties. The scattering properties were also further analyzed to estimate the equivalent size of the scattering centers and the anisotropy factor $g$ in the different experimental conditions.

3.5.

Diffuse Correlation Spectroscopy

DCS is a method to characterize blood flow in tissue, most likely blood flow in the microvasculature.¹²² It measures fluctuations in light intensity caused by the movement of scatterers like RBCs. The method employs a near-infrared CW laser with long coherence length to detect the intensity autocorrelation function. The corresponding electric field autocorrelation function is then analyzed to derive a blood flow index. Although a detailed microscopic model relating the drop in the field autocorrelation function with correlation time is missing up to now,¹²³ changes in this index are taken as a quantitative measure for changes in blood flow.¹²² The drop in the field autocorrelation is best fitted to a Brownian motion model, rather than a random flow model. Choe et al.¹²⁴ used a two-fiber handheld probe with a source–detector separation of 2.5 cm to perform DCS measurements on breast tissue in reflection. This probe was moved along a line crossing the lesion of interest. In a study by Busch et al.,¹²⁵ DCS was applied for measuring the blood flow index in transmission through the breast, which was strongly compressed similar to the conditions in x-ray mammography.

4.1.

Detection and Characterization of Breast Cancer

One of the fundamental questions of optical breast imaging is how this technique can contribute to diagnosis and therapy control of breast cancer, either as a stand-alone method or as an adjunct modality to conventional imaging modalities. The potential of the technique depends, in particular, on differences in the optical and functional properties of healthy breast tissue and of the various types of malignant and benign lesions. Since optical and functional properties of breast tissue show large intersubject variability, relative changes are of particular interest.

First, exploratory studies with the various imaging devices discussed in Sec. 3 had been conducted as case studies with a small number of patients. Hereby, breast carcinomas were found to show a large increase in HbT, whereas benign lesion exhibited a smaller enhancement.^81,126–129 When tissue blood oxygen saturation ${\mathrm{StO}}_2$ was determined, a reduced value was reported for carcinomas compared to normal breast tissue as result of the initial studies.^81,126–128 The first observation corresponds to the increased vascularization of carcinomas due to neoangiogenesis and circumferential growth. The reduced oxygen saturation was interpreted as an indication of a high metabolic rate in carcinomas. However, as will be discussed below in some detail, larger studies revealed that a considerable number of breast tumors are characterized by an increase in tissue blood oxygen saturation, i.e., vascular oxygen supply may overcompensate transvascular oxygen loss and tissue metabolic oxygen consumption.

Over the years, clinical results on a large number of subjects have become available.^{47,72,86–88,130–133} Table 1 summarizes cohort average values of functional properties for lesions and healthy tissue that have been published by several research groups. Table 2 lists results on tumor-to-normal tissue contrast (given as relative values or differences), which represent averages of individual contrast values taken over the patient cohort investigated. It has to be taken into account that instrumental characteristics, methods of data analysis, and clinical protocols differed significantly among studies. Thus, results obtained from different studies are often difficult to compare quantitatively. Despite this caveat, more recent and extended studies confirmed the expectation concluded from tumor neoangiogenesis that breast carcinomas have an increased content of hemoglobin compared to host tissue. Average values of this increase range from about 15%⁴⁷ to 200%.^134,135 In tendency, smaller values are typically obtained by the application of a homogeneous model^72,131 or by reconstruction,^47,88,136 whereas heterogeneous models yield larger ratios.^86,134,135 The ratio ${\mathrm{HbT}}_{(T)}/{\mathrm{HbT}}_{(N)}$ reflects large intersubject variability. Figure 6(a) gives an example of the data spread from a study on 87 carcinomas analyzed by a heterogeneous model.¹³⁴ High intersubject variability was also confirmed by Fang et al.,⁸⁸ who detected statistically significant difference in HbT when malignant lesions are compared with healthy tissue in a paired test but not in an unpaired test.

Table 1

Functional properties of malignant and benign breast lesions (N=number of patients, FAD=fibroadenoma) and of healthy breast tissue from in vivo tissue optical spectroscopy.

Table 2

Differences or ratios of functional properties between malignant or benign breast lesions (N=number of patients) and healthy breast tissue. Data represent average values of the individual contrast data. Values in parentheses were calculated from the cited papers by the authors of this review.

Average data for benign lesions, such as fibroadenomas and fibrocystic changes, show a smaller increase in HbT than carcinomas^86,133 or no significant differences at all.⁴⁷ However, particular benign mastopathic alterations in the breast can result in a large contrast in HbT images.¹³⁸ In a study on about 150 patients with suspicious lesions, only 40% of the histologically confirmed carcinomas dominated the contrast in the optical mammogram.⁸⁵ In the other cases, either benign lesions or highly vascularized tissue regions showed up with higher contrast. Accordingly, the specificity of optical breast imaging as a stand-alone imaging modality was found to be poor.^52,69,85

The initial result that carcinomas generally exhibit lower tissue blood oxygen saturation was not confirmed later by several larger studies. The clinical studies performed during the European OPTIMAMM project on optical mammography have shown that, on average, tissue blood oxygen saturation ${\mathrm{StO}}_2$ in breast carcinomas is not noticeably reduced.^134,135 This observation has then been confirmed by studies of other groups.^47,72,88 As illustrated by the data in Fig. 6(b), there are carcinomas with ${\mathrm{StO}}_2$ above (${\mathrm{StO}}_{2(T)}>{\mathrm{StO}}_{2(N)}$) and below (${\mathrm{StO}}_{2(T)}<{\mathrm{StO}}_{2(N)}$) normal. As discussed above, tissue blood oxygen saturation reflects the balance between oxygen influx, i.e., oxygen supply, and oxygen transvascular flux or tissue metabolic rate of oxygen consumption, i.e., oxygen demand. This balance depends on the tumor vascular network obtained by remodeling of the host vascular network following vessel cooption, angiogenesis, circumferential growth, and regression of tumor vessels. Breast tumors with tissue blood oxygenation above normal [${\mathrm{StO}}_{2(T)}>{\mathrm{StO}}_{2(N)}$, see Fig. 6(b)] may be associated with high tumor blood flow through large-diameter tumor blood vessels, whereas breast tumors with ${\mathrm{StO}}_2$ below normal (${\mathrm{StO}}_{2(T)}<{\mathrm{StO}}_{2(N)}$) may indicate restricted perfusion at least in parts of the tumor vasculature or high oxygen consumption in tumor tissue. A plot of tumor blood oxygenation ${\mathrm{StO}}_{2(T)}$ versus total hemoglobin concentration ${\mathrm{HbT}}_{(T)}$ on a patient by patient basis [see Fig. 5(a) of Ref. 134], corresponding to the data illustrated in Fig. 6, shows lower tumor blood oxygenation at smaller total hemoglobin concentrations and ${\mathrm{StO}}_{2(T)}$ to approach an upper limit at high total hemoglobin concentrations (${\mathrm{HbT}}_{(T)}$). This observation indicates that perfusion (rBF) and total hemoglobin concentration (HbT), or vascular volume density (rBV), are positively correlated. Similarly, a scatter plot of tumor to normal oxygen saturation $r{\mathrm{StO}}_2={\mathrm{StO}}_{2(T)}/{\mathrm{StO}}_{(N)}$ versus tumor to normal total hemoglobin concentration $r\mathrm{HbT}={\mathrm{HbT}}_{(T)}/{\mathrm{HbT}}_{(N)}$ on a case by case basis reveals that tumors with tissue blood oxygen saturation falling below normal are associated with only small increases in total hemoglobin concentration, whereas tumors with ${\mathrm{StO}}_2$ above normal exhibit larger increases in $\mathrm{HbT}$.^134,139 This spread in $r{\mathrm{StO}}_2$ shows tumor-to-normal blood oxygen saturation obtained from diffuse optical imaging to be unsuited for tumor versus normal discrimination. Furthermore, conventional diffuse optical imaging probes intravascular oxygen concentration, it does not directly assess tissue partial oxygen pressure and hence cannot detect tumor hypoxia. Saturation data for benign breast lesions were found to be in the range of the values for healthy breast tissue.^47,86,88 Recently, a study on 26 breast cancer patients employing CW spectroscopic imaging (650 to 850 nm) yielded lower tissue blood oxygenation by $\mathrm{\Delta }{\mathrm{StO}}_2/{\mathrm{StO}}_{2(N)}=-5\%$.⁸⁷ However, scattering was not measured for each patient separately but was accounted for by using literature values both for the reduced scattering coefficient at a reference wavelength and for the scattering power to estimate the scattering spectrum, raising doubts on the accuracy of the results obtained.

Fig. 6

(a) Total hemoglobin concentration and (b) oxygen saturation of 87 carcinomas plotted versus the corresponding reference value of healthy tissue of the same breast (from Grosenick et al.¹³⁴).

Statistically relevant data on water and lipid concentrations in carcinomas were first published by Cerussi et al., who employed a combined frequency domain and broadband spectroscopy system, using, however, a smaller source–detector distance than other groups. In this way, sufficient signal-to-noise-ratio could be reached at wavelengths above 900 nm. Water concentration was found to be larger in carcinomas, whereas lipid concentration was smaller in carcinomas than in normal tissue of the same breast.⁷² Recently, Quarto et al. reported similar results obtained from a parallel-plate breast scanner. Additionally, they reported corresponding data for benign lesions, showing the same qualitative trend but less marked differences than in the case of malignant lesions with respect to normal tissue.⁸⁶

In order to account for the increased Hb and water concentration and for the decreased lipid concentration in a single parameter, the following tissue optical index was proposed to achieve high contrast for carcinomas:⁷²

In recent years, it is becoming more common to try not only to detect optical contrast between breast lesions and healthy tissue but also to correlate optically assessed macroscopic parameters with microscopic information on tumors (e.g., grade, proliferation), even though up to now this has often been attempted only on small patient cohorts. Broadband CW acquisition was performed by Anderson et al.⁸⁷ on 26 breast cancer patients. They observed significant correlation between the optical contrast due to the different tumor-to-normal tissue composition and the tumor grade as quantified by Nottingham histologic score, which is a measure of tumor aggressiveness and is used to guide therapy and predict its outcome. Water content showed the strongest correlation and HbT the weakest one. Interestingly, no significant correlation was observed between optically derived parameters and tumor size, which is expected to increase optical contrast, as a homogeneous model was used for data analysis.⁸⁷ At the University of Pennsylvania, Chung et al.¹⁴⁰ investigated the correlation of optically estimated blood parameters with the expression of Ki67 (proliferation marker, $N=18$) and CD34 (angiogenesis marker, $N=19$). Positive correlation was observed between ${\mathrm{StO}}_2$ and Ki67 expression (which is used to predict patient prognosis and response to chemotherapy), and between HbT and mean vessel area. At Dartmouth College, exploring optical tomography for monitoring of NAC (see Sec. 4.4) in 11 subjects, significant correlation was observed between pretreatment HbT values and expression of CD105 (marker of tumor-induced blood vessels) only in subjects that turned out to be responders to therapy.¹⁴¹ These outcomes suggest that complex information can be obtained noninvasively by optical means, and further work is needed to understand it in depth and to fully exploit its potential.

Several clinical studies have shown that, on average, the reduced scattering coefficient of carcinomas is larger than that of healthy tissue. Fang et al.⁸⁸ and Grosenick et al.¹³⁴ observed an increase by about 20%, Choe et al.⁴⁷ even by about 50%. Higher scattering in malignant lesions than in benign lesions, or healthy tissue, also agrees, at least qualitatively, with the results of a small pilot study performed to compare optically assessed scattering properties with the pathologic analysis of microscopy images of cellular morphology.¹⁴² As shown in Table 1, several groups have measured, in particular, higher scatter power coefficients $b$ in carcinomas than in healthy tissue. Different results on the scattering properties of carcinomas have been reported by Spinelli et al.,¹³⁵ who observed a reduction of scatter power and also a slight reduction of ${\mu }_{\mathrm{s}}^{\prime }$, possibly indicating that the limits of perturbation theory used for data analysis had been reached. For solid benign lesions, either no significant change^47,135 or a small increase of ${\mu }_{\mathrm{s}}^{\prime }$ was reported.⁸⁸

Recently, the concentration of collagen was considered as a further parameter for the characterization of carcinomas and benign lesions. Based on data from 62 patients with carcinomas or benign lesions, Quarto et al.⁸⁶ described a large increase of the collagen concentration in carcinomas compared to a smaller increase in benign lesions and supposed that collagen might have high diagnostic relevance. Taking tissue absorption by collagen into account reduces the total hemoglobin concentration and increases ${\mathrm{StO}}_2$ since collagen contributes to tissue absorption not only above 900 nm but also below 700 nm.¹⁴³

Choe et al.¹²⁴ employed DCS to investigate the microvascular blood flow of healthy tissue and malignant breast lesions. Results on 32 subjects revealed on average a 2.25-fold larger blood flow index for malignant lesions compared to normal tissue of the same breast. Similar results were obtained when the contralateral breast was used as reference, which implies that malignant tumors have significantly higher blood flow than normal tissue.

Several papers discussed optical properties of (liquid filled) cysts. These lesions could be distinguished from carcinomas and solid benign lesions by their smaller reduced scattering coefficient.^85,135,144 By contrast, Fang et al.⁸⁸ did not see such a reduction. Instead, these authors reported a significantly reduced ${\mathrm{StO}}_2$ value in cysts as compared to healthy tissue and to other lesions.

Statistical tests performed on the discussed composition and functional properties of carcinomas, healthy tissue and benign lesions revealed significant differences leading to the conclusion that these properties are suited to diagnose carcinomas and to differentiate between malignant and benign lesions, different from what was suggested by earlier studies.^52,69,85 Accordingly, high sensitivity and specificity values for optical imaging were obtained: e.g., when HbT is used as criterion for diagnosis, Choe et al.⁴⁷ reported a sensitivity of 98% and a specificity of 90% and Zhu et al.¹³⁰ obtained a sensitivity of 92% and a specificity of 93%. Chance et al.⁷⁴ detected 44 carcinomas from 116 subjects with a sensitivity of 96% and a specificity of 93% by using HbT and ${\mathrm{StO}}_2$. However, the latter results should be taken with caution since only CW radiation at three different wavelengths was used and separation of absorption from scattering thus was difficult. Also, Deng et al.¹¹³, exploiting data fusion on x-ray imaging and optical tomography to build HbT maps, were able to differentiate 30 malignant from 24 solid benign lesions ($p=0.006$ in a two-sample $t$-test). For the discrimination between 16 malignant and 9 benign lesions, Zhao et al.¹⁴⁵ tested both HbT and the tissue optical index:

When interpreting the high values of sensitivity and specificity reported, one should have in mind that they were in part derived by exploiting prior knowledge about the existence and even the location of the lesion. The task of optical imaging focuses on determining functional parameters for tissue regions defined by other modalities. This approach differs from the original attempts of optical mammography that aimed at the detection of lesions from the optical image itself and that revealed a low specificity of the optical method.^{52,55,131,148} With lesion position known from a conventional breast imaging modality, tissue regions with high contrast in optical mammograms that do not coincide with the position of the known lesion in the conventional modality are ignored. This approach avoids, in particular, the problem of low specificity and yields values above 90%.

Generally, the high values for sensitivity and specificity are encouraging to use optical imaging as an adjunct modality to conventional breast imaging. The various clinical studies have shown that HbT, water, lipid and collagen concentration, the scattering properties, and related parameter combinations can potentially be exploited to improve the workflow in breast cancer diagnosis.

4.2.

Dynamic Response to External Impact

The work on the detection and differentiation of breast tumors discussed so far was related to properties of breast tissue under static conditions. There have been several attempts to induce dynamic changes of the intrinsic functional properties in order to discriminate malignant from benign lesions and from healthy breast tissue. With respect to the origin of the optical contrast, one can distinguish between (1) alterations that primarily vary blood volume and blood flow and (2) alterations that primarily affect blood oxygen saturation. Due to their disorganized vasculature and the particular metabolism of cancer cells, carcinomas should show a different transient response with respect to healthy tissue and possibly to benign alterations.

Schmitz et al.⁶⁶ proposed a Valsalva maneuver, which was expected to temporarily increase the Hb concentration in the breast due to vein congestion. Using their dynamic dual-breast imager, a single case study revealed a significantly slower recovery of the reconstructed HbR concentration after the end of the breath hold for a carcinoma compared to the contralateral healthy breast. More recently, measurements on two patients with carcinomas using a novel instrument confirmed the sluggish recovery that reflects the increased resistance of the tumor vasculature to blood flow.¹⁴⁹

Blood volume and blood flow can also be varied by the application of external pressure to the breast. The parallel-plate instrument ComfortScan of DOBI Medical International Inc. temporarily applies a pressure of 10 mm Hg to the breast tissue in order to affect cessation of blood flow in the abnormal capillaries of malignant tumors. The local trapping of RBCs is expected to cause a fast reduction of ${\mathrm{StO}}_2$ compared to healthy tissue and benign lesions, where blood flow is still going on. Such a reduction is observable by a reduction of light transmission around 640 nm. A multicenter clinical trial revealed 74% sensitivity and 92% specificity.¹⁵⁰ In a recent study, a sensitivity and specificity of 80% and 87% were obtained, respectively.¹⁵¹ Furthermore, a study on 617 young women showed a sensitivity of 98% and a specificity of 87%.¹⁵² The applied pressure of only 10 mmHg should be compared to the elevated interstitial pressure in breast carcinomas known to rise up to 30 mmHg or even higher.¹³

Xu et al. used a handheld probe with CW light sources and an integrated load sensor to temporarily increase the pressure on top of a lesion known from ultrasound for 10 s. However, measurements on 36 cases did not reveal any significant decrease in ${\mathrm{StO}}_2$ originating from the pressure change.¹⁵³

A temporal change in blood oxygen saturation can be induced by changing either the oxygen or the carbon dioxide (${\mathrm{CO}}_2$) concentration in the inhaled gas. Enhanced oxygen inhalation results in a global increase of ${\mathrm{StO}}_2$, which then causes vasoconstriction. Enhanced ${\mathrm{CO}}_2$ leads to a right shift of the oxy-Hb dissociation curve corresponding to a decrease in ${\mathrm{StO}}_2$, which finally results in vasodilation. Carpenter et al. used these effects in a pilot study on two patients with carcinomas. The patients received inhalation gas during four cycles, each cycle corresponding to 2 min 100% ${\mathrm{O}}_2$ supply followed by 2 min of carbogen gas (95% ${\mathrm{O}}_2$, 5% ${\mathrm{CO}}_2$). Using MR-guided optical imaging, normal tissue showed a stronger modulation between decreased HbT due to vasoconstriction and enhanced HbT due to vasodilation than the carcinomas. This difference was probably an indicator for the lack of adequate response of smooth muscle cells in the carcinoma vessels required for adapting the vessel diameters. Furthermore, strong temporal delays between the vessel diameter modulation in the healthy and the tumor tissue were observed.¹⁵⁴ Using a CW transillumination method, Dixit et al.¹⁵⁵ investigated patients with malignant and benign breast lesions, who were inhaling carbogen. However, the results on temporal changes in ${\mathrm{HbO}}_2$ and HbR for the various lesions derived by an eigenvector approach did not clearly show common features of the lesions. A general increase in oxygenation visible in healthy as well as diseased subjects was explained by the increased oxygen content in carbogen compared to air. The difficulties of Dixit et al. to observe a robust hemodynamic response by using carbogen are in line with a study on healthy subjects by Carpenter et al.,¹⁵⁶ who reported that air/carbogen as well as air/oxygen breathing yield less robust data than the oxygen/carbogen modulation method.

The effects of pressure-induced changes in breast tissue were also investigated in a number of studies, on healthy subjects, mainly, to understand changes in the tissue functional parameters by compression plates or pressure-induced optode coupling to tissue. In a first study on five healthy subjects, Jiang et al.¹⁵⁷ reported that the pressure of a circular optode arrangement on the tissue results in an increase of HbT and ${\mathrm{StO}}_2$ and in a decrease of the tissue water content. By contrast, a later study of this group on 17 healthy subjects revealed a decrease of HbT and no significant alterations in ${\mathrm{StO}}_2$ and water concentration.¹⁵⁸ Similar to the latter result, Carp et al. observed a decrease in HbT when using high compression in their parallel-plate combined optical and x-ray approach. Moreover, a decrease in ${\mathrm{StO}}_2$ and ${\mu }_{\mathrm{s}}^{\prime }$ was observed.^159,160 The reduced HbT indicates that due to higher pressure blood is partly pressed out of the breast vasculature. Furthermore, for high compression similar to x-ray mammography, a strong reduction in blood flow of about 88% was observed using DCS measurements.¹²⁵ Altogether, these results indicate that tissue compression could reduce optical imaging contrast hampering the detection of tumors. Busch et al.¹²⁵ report on a noticeable effect of soft tissue compression. On the other hand, results on tumor contrast obtained with soft tissue compression did not noticeably differ from data obtained with pressure-free techniques. This is also the case for dynamic measurements using a bolus of ICG (cf. next section).

4.3.

Exogeneous Contrast Agents and Fluorescence Mammography

The results on endogeneous tissue contrast discussed in Secs. 4.1 and 4.2 show several promising features that could be helpful in breast cancer diagnosis. However, they did not yet reveal sufficient potential to substantially change the clinical workflow in diagnostic mammography. Exogeneous contrast agents offer the possibility to exploit contrast mechanisms that are superior to endogeneous tissue contrast and, therefore, could help to improve cancer detection and differentiation. In particular, molecular imaging has been discussed and investigated for the past decade as a new way for early detection of tumors with high specificity.

However, the development and approval of a new dye suited for applications on human subjects is a long and expensive process. So far, two contrast agents have been applied in optical imaging studies on patients with breast cancer. The first one is the well-known NIR dye ICG, which is routinely used in clinical practice for other applications (e.g., hepatic function and ocular fundus fluorescence angiography) since many years. ICG has been approved by regulatory agencies for investigation of the microcirculation yet has not been approved for fluorescence imaging of the breast. The second dye is a newly developed substance with the name omocianine, which does not have any approval for medical applications so far.

ICG can be detected in the breast by its absorption around 760 nm as well as by its fluorescence around 830 nm. The pharmacokinetic behavior is well known. After intravenous injection, ICG quickly binds to high molecular weight plasma molecules like lipoproteins. Within typically less than 10 min, the dye is cleared from the blood by the liver.

A first application of this dye in optical mammography was reported by Ntziachristos et al.,⁸⁰ who reconstructed the change in absorption observable about 3 min after intravenous injection of a bolus of ICG. Optical imaging was restricted to a tissue slice selected from concurrent Gd-enhanced MRI taken in the same parallel-plate geometry as the optical imaging. Investigations on two subjects with a malignant (1-cm invasive ductal carcinoma) and a benign (1.5 cm fibroadenoma) lesion showed good spatial correlation of high ICG absorption with the Gd enhancement visible for both lesions. Outside the lesions, areas with smaller ICG enhancement were visible, similar to moderate enhancement obtained for the healthy case. The possible benefit of the ICG enhanced absorption with respect to the measured intrinsic lesion absorption was not considered by the authors.

Intes et al.¹⁶¹ measured the full-time course of ICG uptake and washout on three patients using a single-wavelength CW imager. The circular arrangement of sources and detectors on the freely pending breast was positioned according to the expected lesion position determined from palpation. Figure 7(a) displays the time courses obtained from reconstruction of absorption changes using a perturbation approach. The two malignant cases show a significantly slower inflow and wash-out in comparison to the benign lesion and to the surrounding normal tissue. This slower inflow and wash-out likely indicated the increased resistance to blood flow arising from the chaotic vessel structure in the carcinomas, and possibly indicated extravasation of ICG.¹⁶¹

Fig. 7

(a) ICG inflow and washout characteristics after bolus injection for three patients with breast tumors measured by Intes et al.;¹⁶¹ case I: fibroadenoma (diameter 1 to 2 cm), case II: adenocarcinoma (diameter 2 to 3 cm), case III: invasive ductal carcinoma (size $4\mathrm{cm}\times 3\mathrm{cm}$) (reprinted with permission of American Association of Physicists in Medicine from Intes et al.¹⁶¹); (b) time courses of ICG absorption after bolus injection normalized to HbT for a 2-cm invasive ductal carcinoma and two healthy reference locations measured by Rinneberg et al.⁸⁵

Rinneberg et al.⁸⁵ used the first generation PTB parallel-plate optical mammograph to investigate the time course of ICG absorption after bolus injection for 11 patients with suspicious lesions. These measurements showed distinct patient-to-patient variations of the wash-out curves. However, the temporal characteristic for the carcinoma and healthy tissue of the same breast was almost identical in all cases. As an example, Fig. 7(b) shows the time courses for a 2-cm invasive ductal carcinoma and for two reference positions within the surrounding normal tissue, chosen on different sides of the tumor. In each case, ICG absorption has been normalized to the HbT value measured at the same location of the breast. Since the three curves fall on top of each other after normalization, ICG absorption is proportional to absorption by hemoglobin. This observation reflects the known strong binding of ICG to blood plasma proteins, which hinders extravasation of the dye molecules. The authors concluded that the investigation with ICG yields essentially the same tumor absorption contrast as accessible from the intrinsic total hemoglobin concentration.

In 2007, Corlu et al.⁶⁰ demonstrated that ICG can be detected in breast cancer by fluorescence measurements. The investigations were performed with the frequency-domain/CW parallel-plate instrument of the University of Pennsylvania equipped with spectral filters for fluorescence detection by the installed CCD camera. Fluorescence data were acquired in a time window from 6 to about 16 min after injection of the ICG bolus. Images of the ICG concentration at 10.2 min after ICG injection and of other functional parameters, such as HbT, ${\mathrm{StO}}_2$, and ${\mu }_{\mathrm{s}}^{\prime }$ were derived by 3-D reconstruction. Data on three patients with invasive ductal carcinomas revealed a lesion-to-background contrast in the reconstructed ICG concentration of $4:1$ to $6:1$ compared to about $1.3:1$ contrast in HbT and typically, $1.5:1$ contrast in ${\mu }_{\mathrm{s}}^{\prime }$. The strong difference in the ICG and HbT concentration contrast was discussed as an indicator for extravasation of ICG due to leaky tumor vasculature.⁶⁰ This result disagrees with the data of Rinneberg et al. displayed in Fig. 7(b), which exhibits similar contrast for ICG and HbT about 10 min after bolus injection. As discussed above, extravasation was negligible within the considered 10-min time window after bolus injection for the 11 cases investigated in this study.⁸⁵

In 2009, first results of a clinical study with the PTB fluorescence mammograph were published in which extravasation of ICG in carcinomas was explicitly exploited as contrast mechanism.¹⁶² This study differed in two main aspects from all previous investigations discussed above. First, the ICG concentration was not only measured during the washout period but also, at a later time at which no more ICG could be seen circulating through the blood vessels. Second, the ICG bolus was followed by a typically 20 min lasting period of constant ICG infusion. The aim of this infusion was twofold. It should allow for a measurement at constant ICG concentration in the vascular compartment during infusion to avoid difficulties arising from the continuous drop of the ICG concentration following a bolus injection during image acquisitions times of several minutes. Furthermore, ICG infusion over an extended period of time should increase ICG extravasation compared to the application of a bolus followed by rapid ICG clearance from blood by the liver. Figure 8 illustrates the results obtained at different times of the investigation for a patient with an invasive carcinoma. The mammogram in Fig. 8(a) shows the late-gate intensity absorption contrast at 660 nm before injection of the dye, Fig. 8(b) shows the ICG fluorescence signal that was measured during the infusion period, and Fig. 8(c) shows the ICG fluorescence signal recorded about 25 min after the infusion had been stopped.

Fig. 8

Optical mammograms of a patient with an invasive ductal carcinoma (arrow) from the study of Hagen et al.:¹⁶² (a) absorption (normalized late-gate intensity) at 660 nm without contrast agent, (b) ICG fluorescence during dye infusion, and (c) ICG fluorescence after washout of the dye from the vessels.

The first two images exhibit essentially the same structures in the breast. In Fig. 8(a), the carcinoma as well as several blood vessels are visible due to the corresponding hemoglobin absorption (at 660 nm mainly deoxy-Hb), and in Fig. 8(b), the same structures are visible from fluorescence since the ICG is contained almost entirely in the vasculature inside and outside the carcinoma. The situation changes dramatically in Fig. 8(c). Now, the ICG from the vessels has been washed out completely, and the small amount of ICG that extravasated through the leaky vessel walls of the carcinoma becomes visible. A study on 20 patients showed that ICG extravasation occurs only for malignant and not for benign tumors.¹⁶³ This result is in accordance with the EPR effect of macromolecules in solid carcinomas. ICG, which strongly binds to the lipoproteins in the blood, acts as a fluorescent label for these macromolecules. Such macromolecules can extravasate in carcinomas due to the leakiness of the vessels and remain there for an extended period of time. Interstitial flow is reduced due to the impaired lymphatic system. By contrast, the vasculature in the benign lesion was not permeable for the ICG-labeled macromolecules. Hence, fluorescence mammography with ICG offers a way to differentiate malignant from benign lesions in the breast, which is a clear advantage compared to measurements exploiting the intrinsic tissue contrast only.

The dye omocianine mentioned above, also known under the name tetrasulfonated carbocyanine dye, does not show strong binding to blood plasma proteins.¹⁶⁴ This contrast agent circulates more than 24 h through blood vessels.⁶⁴ So far, it was used in two dose finding studies. Investigations on 52 patients were performed in a multicenter study with a prototype CW fluorescence mammograph from Imaging Diagnostic Systems, Inc. (Ft. Lauderdale, United States).¹⁶⁵ This study revealed considerable variations in sensitivity at different omocianine concentrations. At the dose optimum, a detection rate of 100% was found for malignant lesions, whereby a large number of additionally visible lesions indicated low specificity. Van de Ven et al.⁶⁴ reported on a study with omocianine performed with the Philips fluorescence mammograph. Dye enhancement was observed in lesions as well as in glandular tissue and at the nipple, which indicates problems with differentiation of lesions from healthy tissue, too.

Compared to these results, fluorescence mammography with ICG seems to offer a higher potential for the differentiation of malignant and benign lesions. ICG fluorescence mammograms recorded during the extravascular phase are superior to the intrinsic tissue contrast because (1) carcinomas are visible without background contrast from blood vessels and (2) benign lesions do not show contrast despite enhanced hemoglobin concentration. So far, these results are based on a small number of cases only, and further studies are required to obtain a sufficient data base for evaluating tumor detection and differentiation based on the EPR effect.

4.4.

Monitoring Patient Response to Neoadjuvant Chemotherapy

The capability of near-infrared spectroscopy to measure functional parameters of carcinomas at known locations within the breast has led to its application for monitoring the response to NAC. This therapy is increasingly used for patients with locally advanced breast cancer to shrink tumor size prior to surgery. NAC has been shown to increase the rate of breast-conserving surgery.¹⁶⁶ Moreover, patients with a pathological complete response to chemotherapy are associated with longer disease-free and overall survival.¹⁶⁷ It is desirable to predict the likely success of NAC as early as possible, preferentially before or right after its start. For this purpose, functional imaging methods seem to be better suited than morphological imaging techniques.¹⁶⁸ Compared to techniques like dynamic contrast-enhanced MRI or PET mammography, NIR spectroscopy is technically less demanding and does not need a contrast agent.

The first application of optical imaging for NAC treatment monitoring was demonstrated at the University of California, Irvine.¹⁶⁹ Meanwhile, several of the instruments discussed in Sec. 3 have been employed in corresponding clinical studies including multimodality imagers. Using the DOSI instrument, Cerussi et al.¹⁷⁰ found that for a 3-month NAC treatment with adriamycin/cytoxan, pathologic complete responders and nonresponders could be distinguished already after the first week of the therapy. At this time point, deoxy-Hb concentration and water concentration had dropped by $27\%\pm 15\%$ and $11\%\pm 15\%$, respectively, for responders, whereas nonresponders did not show any significant changes in these parameters. The differentiation between responders and nonresponders was also obtained when the reduction of HbT was considered instead. Furthermore, the responder group showed a significantly larger tumor-to-normal ratio of oxy-Hb than the nonresponders before NAC. In other words, the carcinomas in the nonresponder group should originally have had a lower ${\mathrm{StO}}_2$ value compared to the responders. This assumption was then confirmed in a later paper, where a pretreatment ${\mathrm{StO}}_2$ threshold value of 76.7% was found between the two groups.¹⁷¹ Another study with the DOSI instrument revealed that pathologic complete responders and partial responders could already be distinguished from nonresponders 1 day after start of the NAC by a characteristic flare in the tumor oxy-Hb concentration. On average, this increase in the responder group amounted to more than 40% of the baseline value.¹⁶⁸

These investigations caught the interest of a number of groups involved in breast tissue optical spectroscopy. As discussed above, the objective is to discriminate responders from nonresponders as early as possible, ideally before or shortly after the start of chemotherapy. Soliman et al. proposed to consider the amount of deoxy-Hb and the water volume instead of the deoxy-Hb concentration and the water volume fraction when investigating patients that obtained different NAC treatments.^172,173 By taking these quantities for the entire breast, not only for the lesion volume, a differentiation between responders and nonresponders was obtained after the first week of treatment. At this time point, an increase of both quantities was observed for responders, whereas nonresponders showed a decrease. Four weeks after the start of the treatment, separation worked best for the amount of deoxy-Hb in the lesion and not in the whole breast. The investigations were performed with a commercial instrument (SoftScan from ART, Canada) that provides hemoglobin concentration, fractional water volume, lipid fraction as well as scatter power and scatter amplitude values from measurements at four wavelengths between 690 and 830 nm.¹⁷³

A clinical study with the combined optical and ultrasound handheld probe of the University of Connecticut showed for the first time that responders and nonresponders to NAC are characterized by differences in the pretreatment tumor hemoglobin level. Pathologic complete and near-complete responders exhibited significantly higher tumor concentrations in HbT ($75.7\pm 18.8\mu \mathrm{M}$) as well as in oxy- and deoxy-Hb than subjects with modest or no response (tumor average HbT of $51\pm 14\mu \mathrm{M}$).¹⁷⁴ Recently, Zhu et al.¹⁷⁵ proposed a prediction model for NAC response combining HbT concentrations with conventional pathologic tumor parameters. Unlike Zhu et al., who utilized absolute values of HbT concentrations, Jiang et al.¹⁷⁶ considered relative baseline values of pretreatment tumor HbT with respect to the contralateral breast. In a clinical study on 19 patients with the combined MR-optical instrument developed at Dartmouth College, these relative values were found to be significantly different between responders and nonresponders. Perfect differentiation was obtained when the percent change in tumor HbT within the first cycle of treatment was taken as criterion, which was similar to a previous observation on seven patients.¹⁷⁷ In an exploratory single case study, NIR spectroscopy was performed together with DCS using a handheld probe. Hereby, the tumor-to-normal blood flow contrast was significantly increased on day 3 and significantly decreased on day 5 after start of treatment.¹⁷⁸ Recently, investigations have been started with the dynamic dual breast imager developed by NIRx Technologies. Two weeks after start of the NAC, pathological complete responders showed a faster recovery of deoxy-Hb after a short breath hold than partial and nonresponders.¹⁷⁹

Several studies discussed the time-dependent change of the tumor functional properties over different treatment cycles, in part until the end of NAC.^{141,169,172,173,180–182} Pakalniskis et al.¹⁴¹ reported that tumor HbT decreases significantly during NAC for pathologic complete responders and not for partial responders. A similar trend is visible in the long-term tumor data of Falou et al.^172,173 Schaafsma et al.¹⁸² focused on the investigation of patients with HER2-negative breast cancer, which is associated with decreased angiogenesis resulting in less pronounced differences in HbT between tumor and healthy tissue. The 18 responders in this study showed a statistically significant decrease in oxy-Hb concentration already after the first treatment cycle compared to the four nonresponders. The predictive value for differentiating responders and nonresponders was comparable to DCE-MRI based measurements of tumor volume carried out after the third of six treatment cycles and appeared to outperform DCE-MRI evaluation according to the response evaluation criteria of solid tumors guidelines.¹⁸²

The various investigations have shown that optical imaging is a well suited tool for monitoring the progress during NAC. Functional parameters are highly correlated to the patient’s response to treatment. In this scenario, the location of the tumor is known and the original size of the tumor is large, which enables one to determine the functional parameters with high signal-to-noise ratio. On the other hand, there are still many differences in the investigated types of NAC, in patient groups, in measurement protocols, data registration, and analysis that all together hamper the comparison of the results obtained. In particular, different approaches were used to determine the numbers of responders and nonresponders from pathological assessment. Generally, subjects can be divided into pathologic complete responders, partial responders, and nonresponders. Hereby, complete response is typically defined as microscopic evidence that invasive cancer cells had entirely disappeared. In some studies, subjects with a reduction in tumor size of at least 50% were considered as partial responders, otherwise as nonresponders.¹⁶⁸ In other studies, classification of partial responders and nonresponders was based on grading of tumor cellularity.^174,182 In part, the size criteria was combined with the additional demand of a significant decrease in tumor cellularity to classify a subject as (partial) responder.^172,173 The different ways of pathological classification can have considerable influence on the evaluation of the optical monitoring method, e.g., Cerussi et al.¹⁷⁰ mentioned explicitly that their number of responders differed when using either a binary or a tertiary pathologic classification scheme. Furthermore, since a pathological complete response is accepted as important prognostic factor, some investigators evaluated NIR monitoring of NAC by considering only pathologic complete responders as true responders, whereas partial responders and nonresponders were combined in the overall nonresponder group. Other investigators took complete and partial responders into the same group and the remaining nonresponders as the alternative group.

This variety of approaches emphasizes the need for improved consensus, and that care has to be taken when comparing results from various studies available. Furthermore, to achieve clinical translation, joint efforts should be undertaken to impressively demonstrate the advantages of the optical monitoring technique as a contribution to personalized treatment of patients. A first step into this direction is the multicenter clinical trial of DOSI instruments, recently started at five clinical sites.¹⁸³ It remains to be seen how early and how reliable breast tissue optical spectroscopy can predict the outcome of NAC, preferentially even for pretreated subjects, with clear potential benefits for patients and cost reduction for the healthcare systems.

4.5.

Assessment of Breast Density and Estimate of Cancer Risk

Breast density is a measure of the extent of radiodense fibroglandular tissue in the breast.¹⁸⁴ It reflects composition in terms of fat, stromal, and epithelial tissues. As implied by its definition, it is assessed either from the qualitative or quantitative analysis of x-ray mammograms. In the former case, which is still most common in clinical practice, x-ray mammograms are visually inspected by an expert radiologist, and breast density is most often classified through BI-RADS categories:¹⁸⁴ (1) almost entirely fat; (2) scattered fibroglandular densities; (3) heterogeneously dense; and (4) extremely dense. However, the quantitative estimate of breast density is becoming more and more widespread. It is performed analyzing x-ray mammograms with dedicated percentage of dense (fibroglandular) tissue present in the breast.

The importance of the assessment of breast density comes from its identification as a risk factor. Over the years, a large number of studies have investigated breast density, and it is now widely recognized that it is a strong independent risk factor for developing breast cancer. Actually, it is the strongest risk factor, apart from gender and age: it has been observed that risk is two to six times higher in women with dense breasts (BI-RADS category 4) than in women with fatty breasts (BI-RADS category 1).¹⁸⁵ High density also increases significantly the risk of cancer recurrence.¹⁸⁶

The information on breast density can be useful to identify women that are at high risk and deserve personalized screening/diagnostic paths. Moreover, different from other risk factors, breast density can be modified. Thus, its estimate can help monitoring the effects of interventions that aim at reducing breast density and the related risk. Actually, not only the preventive administration of chemotherapy drugs can reduce breast density but also nonpharmaceutical interventions, like changes in the diet or lifestyle.^187,188 Breast density can also be exploited as an intermediate endpoint in studies on breast cancer etiology. Finally, information on breast density has recently proved useful to predict the pathologic response to neoadjuvant chemo-¹⁸⁹ or endocrine therapy.¹⁹⁰

As mentioned above, x-ray mammography is the gold standard for the assessment of breast density and the method that is generally applied in clinical practice. This implies that the information on breast density becomes known only at the time of the first mammography, typically at the age of 50.¹⁹¹ On the contrary, both for monitoring women at high risk and for devising actions to reduce breast density, it would be beneficial to know it earlier in life. MRI has effectively been applied to estimate the amounts of fibroglandular and fat tissue.^192,193 However, high costs, long measurement times, and, in some cases, limited patient acceptance make MRI unsuitable for screening wide populations. The sensitivity of optical techniques to tissue composition and their general features (relatively inexpensive and simple to operate, well accepted, generating absolute, operator-independent results) makes them potentially interesting candidates as alternative means for the estimate of breast density.

Starting a decade ago, this possibility has been extensively investigated by Lilge and coworkers in Toronto.^116,194 In particular, they have been performing CW transmittance spectroscopy over a broad spectral range (625 to 1060 nm) at four positions (central, medial, distal, and lateral) on both breasts and interpreting transmittance data with principal component analysis. The principal components reflect variations in tissue attenuation (due to both absorption and scattering properties) and, thus, contain information on tissue composition and density. Each spectrum is a weighted average of component spectra, each weighted by a score. Four principal components proved to account for almost the entire tissue variability. Collecting data on more than 200 volunteers, they have shown good correlation between the four scores and percentage mammographic density (assessed using Cumulus,¹⁹⁵ a computer-assisted thresholding software for the analysis of x-ray mammograms).¹⁹⁴

At Politecnico di Milano, we have applied time-resolved diffuse optical imaging performed at seven wavelengths (635 to 1060 nm) to estimate breast tissue composition and structure of 2018 subjects. In particular, time-domain operation in a wavelength range extending beyond 900 nm allowed not only the assessment of blood parameters (total hemoglobin content and oxygenation level), as typically performed by optical instruments, but also of tissue composition in terms of water, lipid, and collagen. The scattering parameters (amplitude $a$, i.e., scattering at 600 nm, and power $b$) were also estimated. The results confirmed strong positive correlation with optically derived water content (and corresponding negative correlation with lipid content), as expected, considering that water strongly absorbs x-rays, while adipose tissue is essentially translucent to x-rays. An optical index based on tissue composition and structure proved suitable to classify noninvasively breast density in agreement with BI-RADS classification performed by an expert radiologist.¹⁹⁶

Even more interesting, collagen content and scattering slope $b$ increased progressively and significantly with breast density.¹⁹⁶ Some other limited studies also confirmed the positive correlation between scattering parameters and mammographic density.^197–199 The scattering properties, especially the scattering power, are expected to depend on collagen structure in tissue. So, the combination of collagen content and scattering parameters may be used to obtain more comprehensive information on collagen in breast tissue. Collagen is a protein family and a major constituent of connective tissue. Fibrillar collagens are essential in determining stromal architecture [see Figs. 1(b) and 2(b)]. Morphological alterations that can be attributed to changes in collagen content and structure occur in breast stroma not just during invasive tumor growth but also in the preneoplastic phase: alterations of stromal architecture and composition are a well-known aspect of both benign and malignant pathologies, and may play an initial role in breast carcinogenesis.^200–202 The optical assessment of collagen content (and potentially also of collagen structure through scattering parameters) may prove of great interest for the assessment of cancer risk, which is associated to mammographic density. Based on what has been discussed above on the link between collagen and cancer development, it might provide noninvasively a more direct estimate of risk than x-ray mammography, which is more sensitive to water content than to collagen. Recently, very preliminary data have also suggested that collagen might provide independent information on risk with respect to mammographic density.²⁰³ If confirmed, this hypothesis would further increase the interest in optical means for the estimate of cancer risk.

4.6.

Dependence of Breast Tissue Composition and Structure on Demographic Parameters

Two recent works on the estimate of breast density (see Sec. 4.5) have applied broadband¹¹⁷ or multiple-wavelength²⁰⁴ acquisition over an extended spectral range (650 to 1000 nm or wider) on more than 200 subjects. Blackmore et al.¹¹⁷ studied how blood parameters (HbT and ${\mathrm{StO}}_2$), water and lipid content, and scattering parameters (amplitude and slope) vary with age and body mass index (BMI) in pre- and postmenopausal women. Taroni et al.²⁰⁴ estimated tissue composition (water, lipids, and collagen), functional information (blood parameters HbT and ${\mathrm{StO}}_2$), and information on the microscopic structure of tissue (scattering amplitude $a$, i.e., scattering coefficient at 600 nm, and power $b$), investigating the correlation with age, BMI, menopausal status, and use of oral contraceptives. Both studies highlighted how age, BMI, and menopause affect tissue composition, increasing the weight of the adipose fraction at the expense of the fibroglandular fraction. This corresponds to significant decrease in water and collagen content and corresponding increase in lipid content. The total hemoglobin content also decreases, while the trend for oxygen saturation is less clear.

The use of oral contraceptives seems to have opposite effect, corresponding to an increase in the fibroglandular tissue fraction. After age adjustment, the trend is the same, but the difference between users of oral contraceptives and nonusers are no longer statistically significant or only close to significance. However, it should be taken into account that complete information (e.g., duration of use) was missing for several subjects, thus further studies are needed to draw definite conclusions. In particular, data referring to the effects of the menopausal status are reported in Table 3. In one case, tissue composition is estimated as relative percentage values and in the other one as absolute concentrations. However, the trends with menopause are very similar for all measured parameters, except for the scattering power, which showed a significant decrease in the study by Taroni et al., while it did not vary significantly in the study of Blackmore et al.

Table 3

Dependence of tissue composition and structure (mean±SD) on menopausal status (N=number of subjects, N.A. = not available).

Over the years, the DOSI system developed at the University of California, Irvine,⁵¹ has also provided pieces of information on breast tissue composition affected by menopausal status. A recent example is represented by the study performed to monitor changes in tissue composition in 28 breast cancer patients undergoing neoadjuvant chemotherapy (see Sec. 4.4). In that case, baseline (pretherapy) water and lipid content in breast tissue was shown to depend markedly on the menopausal status, in agreement with what described above, and also to correlate with fibroglandular density assessed by MRI.

It is important to stress that optically derived data reported in the literature on tissue composition and structure and their changes are fully consistent with the present knowledge on breast physiology, which has often been obtained through ex vivo analyses or studies performed with invasive or more complex and expensive approaches than optical ones (e.g., x-ray mammography and MRI, respectively). The good correspondence between optically derived information and medical knowledge, combined with well-known and already mentioned advantages of optical means (including noninvasiveness, limited costs, good patient acceptance), indicates that optical spectroscopy could take on an effective role to investigate breast physiology in vivo and monitor changes in breast tissue through longitudinal studies (e.g., during chemotherapy or with age).

5.1.

Photoacoustic Mammography

PI (or optoacoustic imaging) of the breast combines the high contrast arising from light absorption in tissue with the high spatial resolution associated with ultrasound detection. The local absorption of light by tissue chromophores or contrast agents leads to thermoelastic expansion of the tissue and to the formation of pressure waves. From the detected ultrasound signals, the location of the absorbers can be estimated with superior spatial resolution compared to diffuse optical imaging. At present, photoacoustic mammography relies essentially on blood contrast and, hence, carcinomas and benign lesions can be detected due to the increase in local deoxy-, oxy-, and total hemoglobin concentration associated with angiogenesis of these lesions. Although a laser optoacoustic imaging system for the breast was proposed by Oraevsky et al.²⁰⁵ as early as in 1996, only more recently clinical systems have been applied in exploratory studies. Here, we will shortly report on the main features of breast PI prototypes. An extensive survey has recently been performed by Menke.²⁰⁶

The most recent clinical studies have been performed using the Twente Photoacoustic Mammograph. In the latest version of this prototype,^207,208 the patient is lying in prone position with the breast pending through an aperture and mildly compressed between a glass window and an ultrasound detector array containing 588 elements. The breast is illuminated through the glass plate by a pulsed laser at 1064 nm with 10-ns pulses at a repetition rate of 10 Hz using a beam diameter of $\sim 70\mathrm{mm}$. Information from x-ray mammograms and ultrasound images are used to define a region of interest on the breast, which is then scanned by the laser beam. Depending on penetration depth, the lateral resolution, and the axial resolution range from 2.3 to 3.9 mm and from 2.5 to 3.3 mm, respectively.^207–209

The case studies performed over a period of 8 years on a total of 23 patients with breast tumors have shown tumors with high imaging contrast. In the most recent study on 11 patients,²⁰⁷ with MRI data available, excellent correspondence in lesion localization was found between PI and MRI images, together with a good to reasonable correlation in lesion shape and appearance. Moreover, it is important to notice that PI images were not affected by the breast density. In patients with cysts (six cases), the lesion was visible with high contrast only at the front and/or at the back of the cyst where impedance changes occur, leading to possible misdiagnosis since malignancies can also appear as one or multiple confined contrast areas.^208,210

The Canon prototype presented a configuration similar to that of the Twente mammograph with respect to patient position and breast compression geometry.²¹¹ The illumination system was different and relied on two sources to provide forward and backward illumination with respect to the 345 element array transducer. The forward and backward laser pulses could be applied simultaneously to maximize optical fluence in the region of interest or alternately to evaluate forward and backward images separately. The laser source was a tunable Ti:Sapphire laser (15-ns pulses at 10-Hz repetition rate). The lateral resolution was 2 to 3 mm. Clinical data were obtained from 26 patients with breast cancer. Images were acquired at both 756- and 797-nm wavelengths to evaluate total hemoglobin content and oxygen saturation. The tumors were correctly detected in 20 cases. The PI images showed high contrast related to the higher hemoglobin content, with tumor margins less defined as compared to MRI images.

The original laser optoacoustic imaging system (LOIS) proposed by Oraevsky et al.²¹² was further developed to perform optoacoustic tomography (LOIS-64).²¹³ With the patient in prone position, the uncompressed breast was suspended through a circular hole in a probe cup. An arc-shaped 64 elements ultrasound transducer array was rotated around the breast to acquire a set of 2-D images each with a depth resolution of 0.5 mm. The expanded laser beam was provided by a Q-switched Alexandrite laser emitting at 757 nm, with pulse duration of 75 ns at a repetition rate of 10 Hz. In the clinical study performed on 27 patients, the LOIS-64 prototype revealed 18 out of 20 histologically proven malignant lesions and 4 out of 8 benign lesions. In all cases detected, the tumor area presented high contrast attributed to high tumor vascularization.

Although not yet tested on breast cancer patients, the OptoSonics 512 prototype presents interesting features.²¹⁴ With the patient in prone position, the breast is placed in a spherical cup, designed to be optically transparent and with an acoustic transmittance larger than 70% at 2 MHz. This cup is positioned in a hemispherical bowl equipped with 512 ultrasound transducers ($3\mathrm{mm}\times 3\mathrm{mm}$ size). The pendant breast being pushed toward the chest wall by the cup is illuminated from below by an Alexandrite laser at 756 nm with a pulse repetition rate of 10 Hz. The system was tested on four healthy volunteers with breast thickness between 2.3 and 4 cm. The reconstructed images clearly showed the breast vasculature with submillimeter resolution. The estimated spatial resolution of the system is 0.42 mm and, based on the phantom study, the authors expect to be able to image tissues as thick as 5.3 cm.

The features of the presently available clinical prototypes illustrate the potential of the photoacoustic approach. Several problems, however, are still open ranging from technical aspects (e.g., detectors sensitivity, field of view), the methodology of the reconstruction algorithms, and the biological basis for discriminating malignant from benign lesions. Clearly, larger clinical studies on both malignant and benign lesions are needed. As for diffuse optical imaging and tomography, improvements might come from multispectral operation and/or integration with other diagnostic modalities, as, e.g., diffuse optical imaging/tomography and ultrasound imaging, to combine information on tissue properties from different approaches. Examples of these integrated modalities are already present in the literature with tests on phantoms or ex vivo specimens.^215–217

A further proposed approach is related to the development of handheld photoacoustic devices, which have the advantage to be simple, portable, and can be easily integrated in an ultrasound probe.^218–220 As for other diagnostic modalities, improvements might come from molecular imaging. In fact, suitable chromophores might enhance the sensitivity and specificity of the diagnostic images, which presently rely essentially on blood contrast.²²¹ Similarly, in hybrid ultrasound and photoacoustic systems, ultrasonic contrast agents can be useful for both modalities.

5.2.

Long Wavelength Broadband Spectroscopy

Up to now, multiple-wavelength and broadband imaging and spectroscopy of the breast have been performed at most up to 1000 to 1100 nm. However, it is interesting to observe that the spectral range available for tissue spectroscopy is extending toward longer wavelengths, up to $2\mu \mathrm{m}$ and beyond. The total attenuation coefficient was measured ex vivo on thin ($200\mu \mathrm{m}$) slices of breast tissue (healthy and cancerous) between 400 and 2500 nm, showing that in the second (1100 to 1350 nm) and even more in the third (1600 to 1850 nm) NIR optical window, deeper light penetration can be achieved compared to the conventional first NIR window (650 to 950 nm). Moreover, in the second and third windows, the scattering becomes progressively less than in the first window, and light attenuation is dominated by tissue chromophore absorption, suggesting the possibility to perform imaging with reduced blurring with respect to what typically observed in diffuse optical images.²²² Moreover, it has been reported that fitting tissue composition and scattering parameters over an extended spectral range (500 to 1600 nm instead of 500 to 1000 nm), which includes more spectral features, can significantly improve the results (i.e., the confidence level on the estimate of water and lipid fractions can improve by a factor of 4).²²³ Recent advances in laser sources, especially supercontinuum fiber sources, and detectors have made it possible to develop portable systems and apply them to perform diffuse optical spectroscopy in clinical settings at least up to 1350 nm,²²⁴ suggesting that results on in vivo breast studies beyond 1000 to 1100 nm will soon become available.