Review
Women's Imaging
April 5, 2017

Optical Imaging of the Breast: Basic Principles and Clinical Applications

Abstract

OBJECTIVE. The objective of this article is to summarize the physical principles, technology features, and first clinical applications of optical imaging techniques to the breast.
CONCLUSION. Light–breast tissue interaction is expressed as absorption and scattering coefficients, allowing image reconstruction based on endogenous or exogenous contrast. Diffuse optical spectroscopy and imaging, fluorescence molecular tomography, photoacoustic imaging, and multiparametric infrared imaging show potential for clinical application, especially for lesion characterization, estimation of cancer probability, and monitoring the effect of neoadjuvant therapy.
Optical imaging (OI) refers to a variety of techniques using near-infrared (NIR) light (700–1000 nm) and visible light (400–700 nm) to provide molecular, morphologic, and functional information, probing absorption, scattering, and fluorescence properties of cells or tissues. One of the most promising applications is to the breast, a superficial organ where the remodeled vasculature and changes in cellular and extracellular tissue structure caused by malignant lesions creates a contrast suitable for OI [1].
In this article, we provide an overview of OI techniques for breast examination. The major OI techniques applied to the breast are summarized in Table 1. After an introduction of the basic principles and implementation of this technology, potential clinical applications are presented: lesion detection and characterization, monitoring of the effect of neoadjuvant therapy (NAT), and evaluation of the probability of breast cancer (BC). Discussion will be limited to the techniques that seem more likely to be translated to BC care. Thus, techniques such as bioluminescence (spontaneous light emission induced by biochemical reaction) will be not considered. Because of its intrinsic limitations in the depth of evaluable tissue (< 5 mm), optical coherence tomography was also considered beyond the scope of this article, although interesting results have been obtained for intraoperative evaluation of tumor margins [2, 3] and lymph nodes [4, 5].
TABLE 1: Main Characteristics of Major Optical Imaging Techniques
TechniqueImages ProvidedOutput and Measured ParametersEndogenous ContrastExogenous Contrast AgentResolutionDepthRemarks
Diffuse optical spectroscopyNoneAbsorption coefficient, scattering coefficientTotal hemoglobin, lipid, water, collagen, cell nuclei diameterNANANAMostly composed of handheld probes; does not produce images but estimates average functional properties such as concentration of oxy- and deoxyhemoglobin, lipid, and water; is useful to monitor treatments; very fast acquisition (tens of seconds)
Diffuse optical imaging2D and 3DAbsorption coefficient, scattering coefficientTotal hemoglobin, lipid, water, collagen, cell nuclei diameterNA10 mmUpto 10 cmSimilar to diffuse optical spectroscopy but much more complex; needs multiple source-detector combinations for data acquisition and image reconstruction
Photoacoustic imaging2D and 3DUltrasound intensityHemoglobin, melanin, waterCyanine dyes0.150 mm at 3 cm of depthUp to 3 cmSpatial and lateral resolution depending on the ultrasound transducer
Fluorescence optical tomography3DNIR fluorescence intensityAmino acids (phenylalanine, tyrosine, tryptophan); nicotinamide and flavins; porphyrins; structural proteins (collagen, elastin); fluorescent pigments (melanin, lipofuscin)Targeted or nontargeted fluorescent probes emitting in the NIR region such as cyanine dyes0.5 mmUpto 1 cmLow spatial resolution in deep tissue imaging; limited to molecular imaging and diagnosis of skin diseases
  NIR fluorescence lifetimeTissue environment such as pH and temperature    
Multiparametric infrared imagingNoneTime-resolved infrared 3D vascular mapAsymmetry of the vascular map between the two breastsNANANANeeds temperature stress for vasoconstriction

Note—NA = not available, NIR = near-infrared.

Basics in Physical Principles

The first attempts to use visible light to diagnose BC were in the first half of the 20th century using a method called “transillumination” or “diaphonography” and resulted in a sensitivity as low as approximately 58% [6,7]. During the 1990s, the identification of the optical properties of tissues and their exploitation in determining the tissue composition opened a new era for OI, which presented a potential for identifying structural changes and functional abnormalities of tissues. An extensive technical review on this topic by Grosenick et al. [1] was recently published.
OI begins with the excitation of endogenous or exogenous chromophores within a volume of interest by an external source of light. The term “chromophore” refers to any molecule in the biologic tissue under evaluation that is able to interact with light. Light interaction includes absorption and scattering.
With absorption, photons hitting a chromophore disappear and release all their energy to molecular electrons. This absorption occurs only at unique molecular-specific values of frequency (energy). The most common chromophores are hemoglobin (both oxygenated and deoxygenated), water, and lipids. These chromophores may be thought of as acting like the mammary gland in the context of mammography, which absorbs x-rays allowing differentiation of the gland from lipid and calcifications, thus creating contrast. The absorbed energy may then be released as a delayed emission of light, the so-called autofluorescence [810], and as nonradiating heat to surroundings. In the latter case, a pressure wave in the form of detectable ultrasound (US) is produced via thermal expansion, a phenomenon called “photoacoustics” [11, 12]. Interestingly, although to a much smaller extent, this type of light-tissue interaction is somewhat similar to thunder, with the production of secondary light and sound after a very large electrical current has flown through the air. A schematic drawing of photoacoustic imaging is shown in Figure 1.
Fig. 1 —Drawing shows generation of acoustic wave. After absorption of energy in form of light by tissue, absorber (chromophore) warms up and starts to expand. This thermal expansion travels in form of pressure (acoustic) wave that may be detected by ultrasound probe.
With scattering, incident photons lose a minimal part of their energy and are deflected from the original direction. For a better understanding of scattering, a photon incident on a cell nucleus may be thought of like a bullet hitting a spherical target of the same size: After the collision, the emergent bullet may continue its way along any direction around the target, including behind the target (isotropic scattering). Thus, shooting a huge number of bullets would result in a uniform distribution of bullets around the target without any preferred direction. Thus, a spherical bowl centered in the target (working like a detector) would be punched uniformly in all its points.
In practice, the light used for OI excitation may be thought of as a huge flow of photons hitting the studied tissue that are scattered in the space around the tissue and are partially absorbed, as shown in Figure 2 [13].
Fig. 2 —Light propagation in tissues. (Reprinted with permission from [13]: Stuker F, Ripoll J, Rudin M. Fluorescence molecular tomography: principles and potential for pharmaceutical research. Pharmaceutics 2011; 3:229–274)
A, Images show what happens in highly absorbing tissue where mean free path (defined as distance between two scattering events or lsc) is large compared with 1/μs, where μs is scattering coefficient. In this case, light intensity decreases with distance.
B, Images show what happens in highly scattering tissue where lsc is comparable to 1/μs. In this situation, light propagation can be modeled with diffusion theory.

Operating Domains of Optical Imaging

Regardless of the actual physical phenomena on which OI is based (absorption, scattering, fluorescence, photoacoustics), three major implementation methods are used in practice according to the illumination scheme. Specifically, instruments for breast OI may be further classified according to the temporal profile of the light source used for excitation and according to the acquisition geometry. They are classified as time domain, frequency domain, and continuous-wave systems.

Time Domain

An exhaustive explanation of this topic is beyond the aim of this article. Here, we briefly mention that time domain systems use very short (on the order of picoseconds) laser pulses to excite tissues. After propagation through the breast, a tissue-specific absorption coefficient and scattering coefficient may be derived from analysis of the detected light. A well-established technique for the detection of these pulses is the time-correlated single-photon counting by which photons are counted at all detector positions along with their time delay (hence, the name “time domain”), something similar to the techniques used in nuclide-based imaging. Time domain systems are characterized by high spatial and depth resolution and are more accurate but are slower than frequency domain systems.

Frequency Domain

Frequency domain systems use intensity-modulated lasers to excite chromophores. Different from time domain systems, frequency domain systems do not count photons but measure the light demodulation and phase shift to calculate absorption and scattering coefficients. In these systems, the excitation light should not be thought of as a collection of pointlike photons but as a flow of electromagnetic waves. On impact on a chromophore, these waves emerge modified in intensity, frequency, and phase. The term “demodulation” refers to the capability to detect these changes. Several modulation frequencies ranging from 100 MHz to 1 GHz are typically used. In practice, a laser pulse with a given fixed frequency is used to excite the studied tissue, and both the intensity and phase of the emergent light are measured by detectors placed around the tissue. Then, another excitation pulse is used with a slightly higher frequency, and the measurements are repeated once more and so on for many exciting frequencies. Roughly, this technique is similar to MR spectroscopy, with the only difference that only one broadband radiofrequency pulse is used in the latter technique instead of multiple single-frequency pulses. From a mathematic point of view, frequency domain measurements correspond to investigate the Fourier transform of time domain measurements. In the parallel to MR spectroscopy, we may say that time domain systems analyze the free induction decay (the temporal profile of the analog signal detected by the receiving coil) and that frequency domain systems analyze the associated spectrum (frequencies included in the free induction decay). Frequency domain systems have high spatial and depth resolution; they are relatively less expensive and are faster than time domain systems but are less accurate.

Continuous Wave

Continuous-wave systems use continuously emitting light sources (i.e., lasers or light-emitting diodes) at several NIR wavelengths or broadband lights. The instrument measures the average attenuation coefficient of the transmitted light, which is a combination (integral) of the absorption coefficient and scattering coefficient. Systems using continuous wave are cheaper and faster than time domain and frequency domain systems. Thus, a large number of source-detector combinations may be used in continuous-wave systems at low cost.
It is important to highlight that the mentioned classification is based on the technical implementation of OI. Theoretically, all combinations of the physical phenomena (absorption, scattering, fluorescence, and photoacoustics) and technical implementations (time domain, frequency domain, and continuous wave) are possible.

Endogenous and Exogenous Contrast

Both absorption and scattering as well as autofluorescence and photoacoustics secondary to absorption may be exploited to obtain tissue-specific biochemical information or may be handled to provide images. Obviously, the instrumentation, data analysis, and image reconstruction depend on the specific physical phenomena used by each technique.
The diagnostic performance relies on existing differences in the optical properties of healthy tissues and various types of malignant and benign lesions. From analysis of the signal arising from a volume of interest (voxel) measured by detectors, two average coefficients are calculated: the absorption coefficient and the scattering coefficient of that specific voxel. These two coefficients are closely related to changes of concentration (in the case of absorption) or morphology (in the case of scattering) of endogenous or exogenous chromophores. Endogenous chromophores refer to molecules naturally present in the biologic tissue under evaluation such as hemoglobin, water, and lipids. Exogenous chromophores refer to molecules with dedicated optical properties that are developed to enhance the tissue contrast. The latter are typically IV injected and generally improve cancer detection and differentiation.
The in vivo measurement of the absorption coefficient of hemoglobin is one of the main approaches in diffuse optical spectroscopy, which we discuss in the next section. Similar to MRI, where radiofrequency pulses are not absorbed by the tissue unless their frequency equals the Larmor frequency, photons reaching hemoglobin are mainly scattered around (not absorbed) unless the photon frequency equals unique values, when they are absorbed instead. Indeed, the absorption spectra of oxy- and deoxyhemoglobin differ from each other, with a slight difference in the absorption frequency. This difference is measurable, creates an optical contrast, and allows distinguishing the two states of hemoglobin. In normal conditions, the ratio between oxyand deoxyhemoglobin is highly conserved. Thus, changes in this ratio may indicate an underlying abnormality. Melanin, myoglobin, and water are other endogenous absorbers in the range of 600–1000 nm.
Biologic tissues are extremely heterogeneous and, according to the Mie theory [8, 9], incident photons are scattered mainly by particles of the same or larger size than the photon wavelength. A detector placed around the tissue in a near-spherical geometry would measure a nonuniform distribution of the scattered light, with preferred directions depending on the tissue biochemical composition. The measured 3D distribution of the scattered photons can represent a tissue-specific signature. For example, changes in the size and number of mitochondria and cell nuclei can be used to differentiate neoplastic tissues from normal tissues. In fact, cell nuclei in normal epithelial tissues have a diameter of 4–7 μm, much smaller than those in cancers (≈ 20 μm), which also contain multiple nuclei [8, 9]. Thus, normal and cancerous tissues have different scattering coefficients (signatures), a difference (contrast) that may be measured and used for diagnosis.
The absorption coefficient is much easier to calculate than the scattering coefficient. In fact, the latter needs the implementation in the calculating algorithm of a model of light transport in biologic tissues, is still approximate, and relies on simplifying assumptions. Moreover, it is time-consuming and demands high computing power. Indeed, most OI techniques mainly use absorption properties of chromophores. Only more advanced techniques include also the study of the scattering properties of compounds such as collagen and lipid. The calculation of these coefficients proceeds automatically and does not require specific skills from the operator.
Exogenous contrast agents are used for a wide array of applications, from the study of cell regulation mechanisms to cancer diagnosis. Optical probes that absorb and emit in the visible light are mainly used for superficial tissue imaging (e.g., skin cancer). However, the background tissue autofluorescence may bias the measurements in the visible range, while it is negligible in the NIR range. Thus, in vivo deep tissue imaging (especially breast imaging) primarily uses molecular probes in the NIR region such as cyanine dyes to minimize autofluorescence.
Two exogenous contrast agents are typically used to increase the endogenous tissue contrast and help differentiate malignant from benign lesions. One is indocyanine green, which has already been approved for other applications (e.g., hepatic function and ocular fundus fluorescence angiography). Indocyanine green strongly binds to macromolecules in the blood and extravasates in carcinomas because of their vessel leakiness, acting there for an extended time as a fluorescent label. After less than 10 minutes, the dye is cleared from the blood by the liver. The other exogenous contrast agent is named “omocianine” but still does not have approval for medical applications [14].

Major Optical Imaging Techniques

Diffuse Optical Spectroscopy

Diffuse optical spectroscopy is the most established OI technique for clinical applications, although it does not provide images [1521]. As we mentioned earlier, it relies mainly on the in vivo measurement of hemoglobin and blood oxygen saturation (SO2) in tissues within a volume of interest. After calibration, diffuse optical spectroscopy may be used for obtaining absolute measurements.

Diffuse Optical Imaging

Diffuse OI is based on diffuse optical spectroscopy, but it is much more complex. In fact, diffuse OI needs multiple source-detector pairs for data acquisition and image reconstruction, providing 2D or 3D images. Basically, the process of diffuse optical spectroscopy is repeated for multiple voxels, and a colorimetric map is presented (Fig. 3).
Fig. 3 —53-year-old woman with 2.2-cm invasive ductal carcinoma in right breast. (Reprinted with permission from [61]: Choe R, Konecky SD, Corlu A. Differentiation of benign and malignant breast tumors by in-vivo three-dimensional parallel-plate diffuse optical tomography. J Biomed Opt 2009; 14:024020)
A and B, Sagittal (A) and axial (B) dynamic contrast-enhanced MR images. Red dashed line in A shows level of axial plane depicted in (B).
C, Image shows tumor region (red) based on optical data with guidance of MRI.
D–I, Diffuse optical tomography images show relative total hemoglobin concentration (D), relative blood oxygen saturation (E), relative oxygenated hemoglobin (F), relative deoxygenated hemoglobin (G), relative tissue scattering at 786 nm (H), and optical index (I). Black solid lines indicate tumor region.
Many prototypes have been designed for application to the breast, with three different geometries. Siemens Healthcare (Erlangen, Germany) and Carl Zeiss Meditec (Oberkochen, Germany) have developed systems in which the breast is compressed by two transparent parallel plates; both systems are equipped with frequency domain instrumentation [22].

Diffuse Optical Tomography

The Physikalisch-Technische Bundesanstalt (Braunschweig, Germany) [23] and Politecnico di Milano (Milan, Italy) [24] later used the parallel-plate geometry to build the first time domain optical mammography system, which, with further refinements, has been developed for providing 3D images (diffuse optical tomography). However, with this geometry, only a depth of up to 1 cm may be typically studied.
Philips Healthcare (Eindhoven, The Netherlands) has developed a device with a cup-like geometry in which the subject lies in the prone position with a freely pendant breast [25]. In this continuous-wave instrument, a total of 255 source fibers and 255 detection fibers were used to reconstruct the tissue attenuation coefficient. A scattering fluid fills the cup to get high-quality optical coupling.
Imaging Diagnostic Systems (Ft. Lauderdale, FL) followed the principle of CT and developed a device in which the freely pendant breast of the patient in a 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 [26].
Generally, with diffuse optical tomography of small breasts, lesions of 5 mm in size can be detected in all parts of the compressed breasts (parallel-plate geometry) and in the outer part of the uncompressed breasts (cup geometry); for large breasts, the detection limit is 7.5 mm [1].
Some optical breast imagers uses hand-held probes. The device developed at the University of California (Irvine, CA) uses two source-detector pairs and combines the frequency domain and the continuous-wave approaches [27]. A detailed overview about handheld NIR breast imaging devices can be found in the review by Erickson and Godavarty [28].

Fluorescence Optical Tomography

Fluorescence optical tomography exploits the fluorescence emission intensity profile of both endogenous fluorophores (autofluorescence) and exogenous contrast agents. Here, the term “fluorophore” refers to special chromophores that emit light with properties similar to those emitted by fluorine. The excitation light is also acquired and used for normalization. A charge-coupled device camera, similar to those used in photography, is mounted as a detector. For studies of fluorescence biodistribution, continuous-wave technology may be advantageous thanks to the higher signal-to-noise ratio and operational simplicity and robustness [29]. Crucial are the noncontact measurements—that is, the measurements where the sources and detectors never touch the tissue [30, 31]. For 3D reconstruction, it is important to illuminate the breast using a large number of projections, similar to other tomographic techniques such as digital breast tomosynthesis.
The use of targeted probes as exogenous contrast agents allowed optical tomography to enter the molecular imaging world. This hybrid technique was originally termed “fluorescence-mediated molecular tomography,” but it is currently known as fluorescence optical tomography [32].
A variant of fluorescence optical tomography is performed by measuring fluorescence lifetime in addition to intensity [10, 3335]. Fluorescence lifetime is an intrinsic property of endogenous and exogenous fluorophores (and fluorine in general), defined as the average time a fluorophore remains in the excited state before it emits light. Importantly, fluorescence lifetime is less dependent on the fluorophore concentration, but it is sensitive to changes in the local tissue environment such as pH and temperature.

Photoacoustic Imaging

As many other imaging techniques, OI consists of perturbing somehow the studied tissue and detecting its response or reaction to such perturbation. In this regard, the tissue may be considered to be like a black box of which we wish to understand the biochemical composition. The trick is to exploit potential differential behavior of compounds that are sensitive to the perturbation.
In the techniques presented, both the perturbation and detection are made by means of visible or NIR light. In this section, we briefly present two additional OI techniques using different methods to perturb the tissue or to detect the signal. Of note, a major drawback of fluorescence imaging is the low spatial resolution in deep tissues due to the scattering of the emission light. Photoacoustic imaging is a new technique promising to overcome this limitation [9, 3638]. As we mentioned earlier, the energy absorbed from the exciting photons by chromophores such as hemoglobin and water may be released in a nonradiative way to the surroundings, increasing its vibrational status and, therefore, the local temperature. This energy transfer is somewhat similar to the longitudinal (T1) MR relaxation, also known as spin-lattice relaxation, during which the energy absorbed from the radiofrequency coil is transferred to the lattice in the form of heat.
The sudden increase of the temperature produces thermal expansion. The increase in temperature is approximately 0.0001–0.1°C and corresponds to a pressure of only 10 Pa to 10 kPa. This tiny magnitude may be detected by an array of highly sensitive US transducers. The differential light absorption produces differential heating and, therefore, forms acoustic waves with different intensities. Thus, the image contrast is primarily determined by the different absorption property of the tissue.
Because US has a much higher penetrance in tissues than visible light or NIR light, exploration of relatively deep tissues could be permitted. However, photoacoustic imaging also suffers from depth limitation because it is based on the light propagation in tissue for excitation of chromophores. Modern instruments can achieve a spatial resolution of 0.150 mm at a 3-cm depth [39].
Photoacoustic tomography uses an image reconstruction algorithm. Typically, it consists of a pulsed laser and sensitive microphones or piezotransducers. For hemoglobin-rich tissues such as blood vessels, lasers at 500–600 nm are used. Cyanine dyes may also be used as exogenous contrast agents, absorbing in the NIR range (700–800 nm).
The first photoacoustic imaging prototype was proposed by Oraevsky et al. in 1996 and further developed to provide 3D images [40]. With the patient in the prone position, the uncompressed breast is suspended in a probe cup. An arc-shaped US transducer array with 64 elements is rotated around the breast to acquire a set of 2D images each with a depth resolution of 0.5 mm.
A different approach is the development of handheld photoacoustic devices, which have the advantage to be simple and portable and can be easily integrated in a US probe [4143].

Multiparametric Infrared Imaging

Real Imaging (Airport City, Israel) has developed a breast OI device using vasoconstriction for perturbation and infrared light in the range of 7.5–13.5 μm for detection. The device is composed of two optical heads separated by 60º. Patients are imaged while sitting stationary in an upright position in front of the device, with the upper part of the body exposed [44].
An initial period of thermal equilibrium precedes the imaging session in the dedicated room with controlled temperature set at 18–22°C. After thermal equilibrium is reached, a metabolic stress is induced by having the patient wear cold gloves (0–5°C), generating vasoconstriction (also in the breast). The cold gloves are removed after 1 minute. The vascular tree of both breasts returns to equilibrium by emitting infrared light due to the increase of blood flow [4548].
Because an association between BC and increased vascularity has been well documented with contrast-enhanced breast MRI, in particular for invasive cancers [4955], the 3D infrared map of both breasts can be used as a cancer-specific signature. The vascular maps are analyzed using multiparametric algorithms that test the difference between the two breasts, evaluating multiple factors such as vascular asymmetry, vascular distortion, and 3D spatial changes.
The device does not provide images. Results are presented as an index of probability that the patient has BC. The index scale ranges from –100 (normal) to 100 (abnormal). The risk model is based on training the parameters on a calibration set of clinically known cases to be initially collected at the installation site.

Hybrid Techniques

Breast OI has also been combined with other imaging modalities. Hereby, the conventional modality provides morphologic images of the breast that are exploited in the reconstruction of the optical properties of the breast tissue. In this way, low spatial resolution and diffuse blurring of reconstructed optical data can be overcome, and optics can add functional information. Morphologic and functional images are coregistered.
Bios (Vimodrone, Italy) has developed a 3D technique combining high-resolution US anatomic information with functional information provided by photoacoustics. A prototype is currently installed at the MD Anderson Cancer Center (Houston, TX) for a preliminary clinical evaluation [56]. In this prototype, the breast is freely pendant into a hemispherical cup filled with aqueous photoacoustic coupling fluid. The imaging module rotates in steps around the breast thereby acquiring a full spherical set of the photo-acoustic data.
The combination of OI and MRI is more challenging. To avoid interference of the optoelectronic components with the magnetic field, long fibers are needed to deliver and collect the light inside the MRI bore. Furthermore, the restricted space inside the bore limits the number of source and detector fibers that can be installed.
The first use of a hybrid OI/MRI system of the breast was reported by Ntziachristos et al. [57]. While at Massachusetts General Hospital (Boston, MA), investigators developed a combined OI/x-ray system for breast examination that uses the parallel-plate geometry [58].

Clinical Investigations

Lesion Characterization

Exploratory clinical studies were performed as case studies including small samples of patients. Initially, BCs were found to show a large increase of hemoglobin concentration and a reduction in SO2. However, SO2 is not so reduced in all BCs, depending on specific vascular network and angiogenesis. Moreover, some benign lesions also present with a hemoglobin concentration increase.
The clinical studies performed in the context of the European OPTIMAMM Project on optical mammography showed that, on average, tissue SO2 in BCs is not substantially reduced [59, 60], a result confirmed by others [6163]. On the other hand, OI probes intravascular oxygen concentration, not the tissue oxygen partial pressure, and therefore cannot detect tumor hypoxia.
Using the total hemoglobin as a criterion for diagnosis, Choe et al. [61] reported a sensitivity of 98% and a specificity of 90%, and Zhu et al. [64] obtained a sensitivity of 92% and a specificity of 93%. Mastanduno et al. [65] calculated the following ratio from MRI-guided diffuse optical tomography:
They reported an increase in specificity from 67% for MRI only to 89% when OI is added.
Importantly, these results were obtained using an a priori knowledge about the lesion existence and location (i.e., using OI as an adjunct to other techniques for detection).
Blood flow can be purposely altered by applying external pressure to the breast. The ComfortScan of DOBI Medical International (Shirley, MA) temporarily applies a pressure of 10 mm Hg using parallel-plate geometry. This pressure is expected to quickly reduce SO2 in malignant lesions compared with healthy tissue and benign lesions. Using dynamic OI in patients with a nonpalpable BI-RADS 4 or 5 lesions scheduled for biopsy, Athanasiou et al. [66] obtained an overall sensitivity of 73% but a specificity of only 38%; the false-negative results were mainly small (< 10 mm) infiltrating malignant lesions and ductal carcinoma in situ.
The results of endogenous tissue contrast techniques have not yet revealed a potential for clinical applications. Exogenous contrast agents offer the possibility to improve lesion detection and characterization. To date, results are based on a small number of cases, and further studies are required for defining the real power of OI for breast lesion detection and characterization.

Monitoring Tumor Response to Neoadjuvant Therapy

One potential application of diffuse OI is monitoring the effect of NAT. The first attempts to use OI in this setting date back to the early 2000s. Functional OI methods seem more suited than morphologic imaging techniques [67]. Moreover, compared with dynamic contrast-enhanced MRI or PET, NIR spectroscopy is technically less demanding and does not need the administration of contrast agents or radiotracers.
Using diffuse OI, Cerussi et al. [68] found that pathologic complete responders to a 3-month course of NAT (doxorubicin hydro-chloride [Adriamycin, Pharmacia] and cyclophosphamide [Cytoxan, Bayer HealthCare]) could be distinguished from nonresponders as soon as after the first week of therapy; at that time, the concentration of deoxyhemoglobin and water dropped by 27% and 11%, respectively, in responders, whereas these values did not show any significant change in nonresponders.
A clinical study using the hybrid OI/US handheld probe of the University of Connecticut showed that responders and nonresponders to NAT are characterized by differences in the pretreatment tumor hemoglobin level. Pathologic complete and near-complete responders presented with significantly higher tumor concentrations of both oxy- and deoxyhemoglobin than subjects with modest or no response [69]. A well-cut differentiation was obtained by Jiang et al. [70] for 19 patients when the percentage change in tumor total hemoglobin within the first cycle of treatment was taken as the criterion for predicting response.
Schaafsma et al. [71] evaluated patients with ErbB-2 (also known as HER2/neu)–negative BC. They found less pronounced differences in total hemoglobin between tumor and healthy tissue. The 18 responders showed a statistically significant decrease in oxyhemoglobin concentration after the first treatment cycle compared with the four nonresponders. The predictive value for differentiating responders from nonresponders was comparable to that based on measurements of tumor volume carried out with dynamic contrast-enhanced MRI after the third of six treatment cycles but outperformed MRI according to Response Evaluation Criteria in Solid Tumors (RECIST) guidelines [71].
The data seem promising, but there are still too many differences in patient groups, NAT regimens, measurement protocols, and data analysis to allow reliable comparison among results.

Assessment of Breast Cancer Probability

The first clinical experience with the multiparametric infrared imaging device produced by Real Imaging [72] was performed on a group of 118 women, 52 healthy women (control group) and 66 women with biopsy-proven BC with a tumor size of 5–45 mm (mean, 17.5 mm). A reader blinded to the reference standard correctly classified findings as suspicious in 60 of 66 patients with BC (42 tumors were < 2 cm [i.e., T1 stage tumors]); however, the findings for 16 of the 52 healthy women were false-positives (sensitivity, 91%; specificity, 72%; positive predictive value [PPV], 81%; negative predictive value, 86%).
As one would expect for a risk screening, specificity and PPV are not excellent. Indeed, because the primary goal of this technology is selection of patients at risk for BC, sensitivity is the overriding performance value given that BCs should not be missed. Of course, if the specificity were low, a large number of women would undergo unnecessary stress and examinations for further evaluation. However, the reported specificity is more than 70%, which is easily comparable to existing breast imaging modalities such as mammography, US, and MRI. Furthermore, because the system lacks the ability to localize a suspicious lesion, its potential clinical role would be to assist as a gatekeeper, triaging patients to undergo more sensitive studies. On the other hand, given the high sensitivity, negative OI findings could potentially allow a tailored screening pathway.

Evaluation of Breast Density

The possibility of quantitatively estimating breast density by means of OI has also been investigated. A good correlation was shown between optical parameters and percentage mammographic density [73]. Moreover, an optical index based on tissue composition and structure was found to be in agreement with BI-RADS density [74]. Collagen, a major component of connective tissue of extracellular matrix, was found to progressively and significantly increase with breast density. Collagen seems to play an initial role in breast carcinogenesis by alterations in stromal architecture in the preneoplastic stage and during the invasive growth [75], opening new directions for the assessment of BC risk through optical molecular investigations.

Conclusion

A structural, vascular, and molecular imaging study of the breast is possible using OI and is relatively easily performed. Optical techniques are relatively inexpensive, are simple to operate, and are generally well accepted by patients. In the past decades, a technologic development tried to bring OI closer to clinical practice, especially for a superficial organ such as the breast. There is still not sufficient power for OI to be used for BC detection as a screening tool, where mammography (and its evolution into tomosynthesis) for average-risk women and MRI for high-risk women remain the standard of care. The spatial resolution of breast OI is still poor, and the diagnostic accuracy of stand-alone OI is still too low. An interesting potential is to use OI as an adjunct modality to mammography, US, or MRI.
The two most interesting fields of OI application are currently the evaluation of BC probability and monitoring the effect of NAT. Multiparametric infrared imaging provides a risk assessment (i.e., the probability for a woman to have a BC), without specifying the BC side and site. If results are confirmed, the system could allow identification of women needing further workup or specific surveillance imaging. However, there is concern certainly about the effect this tehnique would have on a radiologist's workload in the context of routine screening.
The special advantage of OI in monitoring the NAT effect is that, in this context, it works on known BC at a known location. By measuring functional parameters in cancerous tissue without the need of contrast agent, OI may give more detailed and specific insights for different cancer types and treatment regimens.
In the past decades radiologists have seen major changes basically consisting of refinements and developments of already existing breast imaging modalities. Film-screen mammography evolved into digital mammography and tomosynthesis, now into contrast-enhanced dual-energy mammography. Elastosonography was added to standard US as a tool for improved lesion characterization. MRI is gaining new power from adding DWI for lesion characterization and detection. No completely new technologies have entered the clinical arena of breast imaging in the past 30 years. If this will happen, OI seems to be the candidate.

Footnote

G. Di Leo received a consulting fee from Real Imaging.

References

1.
Grosenick D, Rinneberg H, Cubeddu R, Taroni P. Review of optical breast imaging and spectroscopy. J Biomed Opt 2016; 21:091311
2.
Nguyen FT, Zysk AM, Chaney EJ, et al. Intraoperative evaluation of breast tumor margins with optical coherence tomography. Cancer Res 2009; 69:8790–8796
3.
Curatolo A, McLaughlin RA, Quirk BC, et al. Ultrasound-guided optical coherence tomography needle probe for the assessment of breast cancer tumor margins. AJR 2012; 199:[web]W520–W522
4.
Nguyen FT, Zysk AM, Chaney EJ, et al. Optical coherence tomography: the intraoperative assessment of lymph nodes in breast cancer. IEEE Eng Med Biol Mag 2010; 29:63–70
5.
Savastru D, Chang EW, Miclos S, Pitman MB, Patel A, Iftimia N. Detection of breast surgical margins with optical coherence tomography imaging: a concept evaluation study. J Biomed Opt 2014; 19:056001
6.
Cutler M. Transillumination of the breast. Ann Surg 1931; 93:223–234
7.
Geslien GE, Fisher JR, DeLaney C. Transillumination in breast cancer detection: screening failures and potential. AJR 1985; 144:619–622
8.
Tuchin VV. Tissue optics: light scattering methods and instruments for medical diagnosis, 2nd ed. Bellingham, WA: SPIE Publications, 2007
9.
Wang LV, Wu HI. Biomedical optics: principles and imaging, 1st ed. Hoboken, NJ: Wiley-Interscience, 2007
10.
Berezin MY, Achilefu S. Fluorescence lifetime measurements and biological imaging. Chem Rev 2010; 110:2641–2684
11.
Wang LV. Photoacoustic imaging and spectroscopy. Boca Raton, FL: CRC, 2009
12.
Knappe V, Frank F, Rohde E. Principles of lasers and biophotonic effects. Photomed Laser Surg 2004; 22:411–417
13.
Stuker F, Ripoll J, Rudin M. Fluorescence molecular tomography: principles and potential for pharmaceutical research. Pharmaceutics 2011; 3:229–274
14.
Poellinger A, Persigehl T, Mahler M, et al. Near-infrared imaging of the breast using omocianine as a fluorescent dye: results of a placebo-controlled, clinical, multicenter trial. Invest Radiol 2011; 46:697–704
15.
Sokolov K, Follen M, Richards-Kortum R. Optical spectroscopy for detection of neoplasia. Curr Opin Chem Biol 2002; 6:651–658
16.
Kukreti S, Cerussi AE, Tanamai W, Hsiang D, Tromberg BJ, Gratton E. Characterization of metabolic differences between benign and malignant tumors: high-spectral-resolution diffuse optical spectroscopy. Radiology 2010; 254:277–284
17.
Ramanujam N, Mitchell MF, Mahadevan A, Thomsen S, Silva E, Richards-Kortum R. Fluorescence spectroscopy: a diagnostic tool for cervical intraepithelial neoplasia (CIN). Gynecol Oncol 1994; 52:31–38
18.
Richards-Kortum R, Sevick-Muraca E. Quantitative optical spectroscopy for tissue diagnosis. Annu Rev Phys Chem 1996; 47:555–606
19.
Tromberg BJ, Pogue BW, Paulsen KD, Yodh AG, Boas DA, Cerussi AE. Assessing the future of diffuse optical imaging technologies for breast cancer management. Med Phys 2008; 35:2443–2451
20.
Vishwanath K, Klein D, Chang K, Schroeder T, Dewhirst MW, Ramanujam N. Quantitative optical spectroscopy can identify long-term local tumor control in irradiated murine head and neck xenografts. J Biomed Opt 2009; 14:054051
21.
Kukreti S, Cerussi A, Tromberg B, Gratton E. Intrinsic tumor biomarkers revealed by novel double-differential spectroscopic analysis of near-infrared spectra. J Biomed Opt 2007; 12:020509
22.
Moesta KT, Fantini S, Jess H, et al. Contrast features of breast cancer in frequency-domain laser scanning mammography. J Biomed Opt 1998; 3:129–136
23.
Grosenick D, Wabnitz H, Rinneberg HH, Moesta KT, Schlag PM. Development of a time-domain optical mammograph and first in vivo applications. Appl Opt 1999; 38:2927–2943
24.
Taroni P, Danesini G, Torricelli A, Pifferi A, Spinelli L, Cubeddu R. Clinical trial of time-resolved scanning optical mammography at 4 wavelengths between 683 and 975 nm. J Biomed Opt 2004; 9:464–473
25.
Colak SB, van der Mark MB, Hooft GW, Hoogenraad JH, van der Linden ES, Kuijpers FA. Clinical optical tomography and NIR spectroscopy for breast cancer detection. IEEE J Sel Top Quantum Electron 1999; 5:1143–1158
26.
Floery D, Helbich TH, Riedl CC, et al. Characterization of benign and malignant breast lesions with computed tomography laser mammography (CTLM): initial experience. Invest Radiol 2005; 40:328–335
27.
O'Sullivan TD, Cerussi AE, Cuccia DJ, Tromberg BJ. Diffuse optical imaging using spatially and temporally modulated light. J Biomed Opt 2012; 17:071311
28.
Erickson SJ, Godavarty A. Hand-held based near-infrared optical imaging devices: a review. Med Eng Phys 2009; 31:495–509
29.
Ntziachristos V, Ripoll J, Wang LV, Weissleder R. Looking and listening to light: the evolution of whole-body photonic imaging. Nat Biotechnol 2005; 23:313–320
30.
Graves E, Ripoll J, Weissleder R, Ntziachristos V. A submillimeter resolution fluorescence molecular imaging system for small animal imaging. Med Phys 2003; 30:901–911
31.
Patwardhan S, Bloch S, Achilefu S, Culver J. Time-dependent whole-body fluorescence tomography of probe bio-distributions in mice. Opt Express 2005; 13:2564–2577
32.
Ntziachristos V, Tung CH, Bremer C, Weissleder R. Fluorescence molecular tomography resolves protease activity in vivo. Nat Med 2002; 8:757–760
33.
Berezin MY, Lee H, Akers W, et al. Engineering NIR dyes for fluorescent lifetime contrast. Conf Proc IEEE Eng Med Biol Soc 2009; 1:114–117
34.
Nothdurft RE, Patwardhan SV, Akers W, Ye Y, Achilefu S, Culver JP. In vivo fluorescence lifetime tomography. J Biomed Opt 2009; 14:024004
35.
Bloch S, Lesage F, McIntosh L, Gandjbakhche A, Liang K, Achilefu S. Whole-body fluorescence lifetime imaging of a tumor-targeted near-infrared molecular probe in mice. J Biomed Opt 2005; 10:054003
36.
Maslov K, Stoica G, Wang LV. In vivo dark-field reflection-mode photoacoustic microscopy. Opt Lett 2005; 30:625–627
37.
Zhang HF, Maslov K, Wang LV. In vivo imaging of subcutaneous structures using functional photo-acoustic microscopy. Nat Protoc 2007; 2:797–804
38.
Li ML, Zhang HE, Maslov K, Stoica G, Wang LV. Improved in vivo photoacoustic microscopy based on a virtual-detector concept. Opt Lett 2006; 31:474–476
39.
Kim C, Favazza C, Wang LV. In vivo photoacoustic tomography of chemicals: high-resolution functional and molecular optical imaging at new depths. Chem Rev 2010; 110:2756–2782
40.
Ermilov SA, Khamapirad T, Conjusteau A, et al. Laser optoacoustic imaging system for detection of breast cancer. J Biomed Opt 2009; 14:024007
41.
Deán-Ben XL, Razansky D. Portable spherical array probe for volumetric real-time optoacoustic imaging at centimeter-scale depths. Opt Express 2013; 21:28062–28071
42.
Deán-Ben XL, Razansky D. Functional optoacoustic human angiography with handheld video rate three dimensional scanner. Photoacoustics 2013; 1:68–73
43.
Daoudi K, van den Berg PJ, Rabot O, et al. Hand-held probe integrating laser diode and ultrasound transducer array for ultrasound/photoacoustic dual modality imaging. Opt Express 2014; 22:26365–26374
44.
Real Imaging website. The product: Real Imager 8 (RI8). www.realimaging.com/ri-usa/technology/product.asp. Accessed February 8, 2017
45.
Williams KL, Williams FJ, Handley RS. Infra-red thermometry in the diagnosis of breast disease. Lancet 1961; 2:1378–1381
46.
Kennedy DA, Lee T, Seely D. A comparative review of thermography as a breast cancer screening technique. Integr Cancer Ther 2009; 8:9–16
47.
Gautherie M. Thermopathology of breast cancer: measurement and analysis of in vivo temperature and blood flow. Ann N Y Acad Sci 1980; 335:383–415
48.
Feig SA, Shaber GS, Schwartz GF, et al. Thermography, mammography, and clinical examination in breast cancer screening: review of 16,000 studies. Radiology 1977; 122:123–127
49.
Mahfouz AE, Sherif H, Saad A, et al. Gadolinium-enhanced MR angiography of the breast: is breast cancer associated with ipsilateral higher vascularity? Eur Radiol 2001; 11:965–969
50.
Carriero A, Di Credico A, Mansour M, Bonomo L. Maximum intensity projection analysis in magnetic resonance of the breast. J Exp Clin Cancer Res 2002; 21:77–81
51.
Sardanelli F, Iozzelli A, Fausto A, Carriero A, Kirchin MA. Gadobenate dimeglumine-enhanced MR imaging breast vascular maps: association between invasive cancer and ipsilateral increased vascularity. Radiology 2005; 235:791–797
52.
Sardanelli F, Fausto A, Menicagli L, Esseridou A. Breast vascular mapping obtained with contrast-enhanced MR imaging: implications for cancer diagnosis, treatment, and risk stratification. Eur Radiol 2007; 17:48–51
53.
Schmitz AC, Peters NH, Veldhuis WB, et al. Contrast-enhanced 3.0-T breast MRI for characterization of breast lesions: increased specificity by using vascular maps. Eur Radiol 2008; 18:355–364
54.
Sardanelli F. Vascular maps on contrast-enhanced breast MR images: a new approach requiring further investigation. Radiology 2008; 247:295–296
55.
Martincich L, Bertotto I, Montemurro F, et al. Variation of breast vascular maps on dynamic contrast-enhanced MRI after primary chemotherapy of locally advanced breast cancer. AJR 2011; 196:1214–1218
56.
ClinicalTrials.gov website. LOUIS-3D Breast Study. https://clinicaltrials.gov/ct2/show/NCT01755130. Accessed February 8, 2017
57.
Ntziachristos V, Ma X, Chance B. Time-correlated single photon counting imager for simultaneous magnetic resonance and near-infrared mammography. Rev Sci Instrum 1998; 69:4221–4233
58.
Fang Q, Carp SA, Selb J, et al. Combined optical imaging and mammography of the healthy breast: optical contrast derived from breast structure and compression. IEEE Trans Med Imaging 2009; 28:30–42
59.
Grosenick D, Wabnitz H, Moesta KT, Mucke J, Schlag PM, Rinneberg H. Time-domain scanning optical mammography. Part II. Optical properties and tissue parameters of 87 carcinomas. Phys Med Biol 2005; 50:2451–2468
60.
Spinelli L, Torricelli A, Pifferi A, Taroni P, Danesini G, Cubeddu R. Characterization of female breast lesions from multi-wavelength time-resolved optical mammography. Phys Med Biol 2005; 50:2489–2502
61.
Choe R, Konecky SD, Corlu A. Differentiation of benign and malignant breast tumors by in-vivo three-dimensional parallel-plate diffuse optical tomography. J Biomed Opt 2009; 14:024020
62.
Cerussi A, Shah N, Hsiang D, Durkin A, Butler J, Tromberg BJ. In vivo absorption, scattering, and physiologic properties of 58 malignant breast tumors determined by broadband diffuse optical spectroscopy. J Biomed Opt 2006; 11:044005
63.
Fang Q, Selb J, Carp SA, et al. Combined optical and x-ray tomosynthesis breast imaging. Radiology 2011; 258:89–97
64.
Zhu Q, Hegde PU, Ricci A Jr, et al. Early-stage invasive breast cancers: potential role of optical tomography with US localization in assisting diagnosis. Radiology 2010; 256:367–378
65.
Mastanduno MA, Xu J, El-Ghussein F, et al. MR-guided near-infrared spectral tomography increases diagnostic performance of breast MRI. Clin Cancer Res 2015; 21:3906–3912
66.
Athanasiou A, Vanel D, Fournier L, Balleyguier C. Optical mammography: a new technique for visualizing breast lesions in women presenting non palpable BIRADS 4–5 imaging findings: preliminary results with radiologic-pathologic correlation. Cancer Imaging 2007; 7:34–40
67.
Roblyer D, Ueda S, Cerussi A, et al. Optical imaging of breast cancer oxyhemoglobin flare correlates with neoadjuvant chemotherapy response one day after starting treatment. Proc Natl Acad Sci U S A 2011; 108:14,626–14,631
68.
Cerussi A, Hsiang D, Shah N, et al. Predicting response to breast cancer neoadjuvant chemotherapy using diffuse optical spectroscopy. Proc Natl Acad Sci U S A 2007; 104:4014–4019
69.
Zhu Q, DeFusco PA, Ricci A Jr, et al. Breast cancer: assessing response to neoadjuvant chemo-therapy by using US-guided near-infrared tomography. Radiology 2013; 266:433–442
70.
Jiang S, Pogue BW, Kaufman PA, et al. Predicting breast tumor response to neoadjuvant chemotherapy with diffuse optical spectroscopic tomography prior to treatment. Clin Cancer Res 2014; 20:6006–6015
71.
Schaafsma BE, van de Giessen M, Charehbili A, et al. Optical mammography using diffuse optical spectroscopy for monitoring tumor response to neoadjuvant chemotherapy in women with locally advanced breast cancer. Clin Cancer Res 2015; 21:577–584
72.
Sella T, Sklair-Levy M, Cohen M, et al. A novel functional infrared imaging system coupled with multiparametric computerised analysis for risk assessment of breast cancer. Eur Radiol 2013; 23:1191–1198
73.
Blackmore KM, Knight JA, Walter J, Lilge L. The association between breast tissue optical content and mammographic density in pre- and post-menopausal women. PLoS One 2015; 10:e0115851
74.
Taroni P, Pifferi A, Quarto G, et al. Noninvasive assessment of breast cancer risk using time-resolved diffuse optical spectroscopy. J Biomed Opt 2010; 15:060501
75.
Luparello C. Aspects of collagen changes in breast cancer. J Carcinogene Mutagene 2013; S13:007

Information & Authors

Information

Published In

American Journal of Roentgenology
Pages: 230 - 238
PubMed: 28379746

History

Submitted: August 24, 2016
Accepted: February 9, 2017
Version of record online: April 5, 2017

Keywords

  1. breast
  2. fluorescence molecular tomography
  3. optical imaging
  4. photoacoustic imaging

Authors

Affiliations

Giovanni Di Leo
Radiology Unit, IRCCs Policlinico San Donato, Via Morandi 30 20097 San Donato Milanese, Milan, Italy.
Rubina Manuela Trimboli
Department of Biomedical Sciences for Health, Università Degli Studi di Milano, Milan, Italy.
Tamar Sella
Department of Diagnostic Imaging, Hadassah Hebrew University Medical Center, Jerusalem, Israel.
Francesco Sardanelli
Radiology Unit, IRCCs Policlinico San Donato, Via Morandi 30 20097 San Donato Milanese, Milan, Italy.
Department of Biomedical Sciences for Health, Università Degli Studi di Milano, Milan, Italy.

Metrics & Citations

Metrics

Citations

Export Citations

To download the citation to this article, select your reference manager software.

There are no citations for this item

View Options

View options

PDF

View PDF

PDF Download

Download PDF

Media

Figures

Other

Tables

Share

Share

Copy the content Link

Share on social media