Page 387 - Biomedical Engineering and Design Handbook Volume 2, Applications

BREAST IMAGING SYSTEMS: DESIGN CHALLENGES FOR ENGINEERS 365

measuring perturbations in the local conductance produced by small lesions, large lesions (with size
greater than that of the probe) can also be missed since the conductance and capacitance of the
imaged region can then appear uniform. For these reasons, it has been recommended that ultrasound
be performed prior to EIS (Malich and Fuchsjager, 2003). In a study comparing ultrasound, MRI,
and EIS, Malich et al. obtained 81 percent and 63 percent sensitivity and specificity for EIS in 100
mammographically suspicious lesions. The authors attributed the high false-positive rate to artifacts
produced when imaging superficial skin lesions, and resulting from poor contact or air bubbles
(Malich et al., 2001a).

12.4.6 Other Techniques
Techniques for using optical radiation for breast imaging (sometimes referred to as optical mammog-
raphy) have been under development for decades. Due to the fact that the optical absorption coefficient
of water drops by nearly 2 orders of magnitude in the frequency range between 600 and 1000 nm
relative to either longer or shorter wavelengths, and in this range the absorption coefficient of oxy
hemoglobin and deoxy hemoglobin are both about one-tenth of their values in the visible or infrared
(Nioka and Chance, 2005), biologic tissue is significantly more translucent for near-infrared (NIR) light
than for shorter wavelength visible light. However, within this spectral band, known as the NIR win-
dow, elastic scattering of photons dominates absorption [in the breast, the mean free path for absorp-
tion is approximately 25 cm, while that for scattering is ~0.001 cm (Nioka and Chance, 2005)]. Optical
transport of NIR radiation in the breast is therefore modeled as a diffusion process, and this type of
imaging is referred to as diffuse optical imaging (DOI) or diffuse optical tomography (DOT), if 3D
images are reconstructed.
It was also realized, however, that the hemoglobin signals provide information regarding the
source of oxygen in the tissue, while the cyt c ox signals indicate the intracellular availability of
oxygen foroxidative phosphorylation. This ability of recognizing the source/sink relationship greatly
enhances the value of NIR spectrophotometry (NIRS) for research andclinical purposes.
NIR breast imaging can be described as either endogenous or exogenous. Endogenous imaging
relies on the differences in the IR absorption of oxy hemoglobin and deoxy hemoglobin and on the
differences in vasculature between normal and malignant tissues (associated with angiogenesis).
Exogenous imaging utilizes injected NIR-excitable fluorescent dyes (such as indocyanine green) to
increase contrast between tumors and surrounding healthy tissue.
Photoacoustic tomography is a type of optical imaging that relies on the fact that the absorption
of pulsed NIR radiation by tissue leads to heating, followed by rapid thermoelastic expansion and
the subsequent generation of broadband thermoelastic waves. Photoacoustic imaging uses short NIR
pulses to deposit energy in the breast, but instead of imaging the transmitted NIR radiation, the pho-
toacoustic waves that are emitted by the irradiated tissue and propagate to the surface are detected,
typically by an array of ultrasound transducers (Ku et al., 2005; Manohar et al., 2004; Wang et al.,
2002). The image is formed by differences in the NIR absorption characteristics of different tissues.
As in transmission optical imaging, photoacoustic imaging can be enhanced by the use of optical
contrast agents (Alacam et al., 2008; Kim et al., 2007; Ku and Wang, 2005).
Optical mammography offers the potential advantages of being a low-cost, nonionizing diagnos-
tic adjunct to x-ray mammography. There is, however, evidence that laser irradiation can produce
long-term physical changes in biological tissue that result in a reduction of optical absorption
(Gondek et al., 2006). The principle obstacles to both structural and functional optical breast
imaging are the significant levels of scattering of optical photons in biological tissue, rendering the
image reconstruction task mathematically ill-posed. It is now generally acknowledged that optical
mammography will probably never achieve the submillimeter spatial resolution of other breast imaging
modalities such as x-ray, ultrasound, or MRI. Thus most current efforts are geared toward the detection
of the presence of particular physiologic states (e.g., hypoxia, angiogenesis, hypermetabolism) rather
than the acquisition of a high-resolution structural map. In particular, multiwavelength NIR imaging
(Srinivasan et al., 2007) and dynamic DOT imaging (Alacam et al., 2008; Boverman et al., 2007;
Nioka et al., 2005) are being developed by a number of investigators.