! Abstract—Optical imaging has the potential to play a major role in breast cancer screening and diagnosis due to its ability to image cancer characteristics such as angiogenesis and hypoxia. A promising approach to evaluate and quantify these characteristics is to perform dynamic imaging studies in which one monitors the hemodynamic response to an external stimulus, such as a valsalva maneuver. It has been shown that the response to such stimuli shows MARKED differences between cancerous and healthy tissues. The fast imaging rates and large dynamic range of digital devices makes them ideal for this type of imaging studies. Here we present a digital optical tomography system designed specifically for dynamic breast imaging. The instrument uses laser diodes at 4 different near-infrared wavelengths with 32 sources and 128 silicon photodiode detectors. I.INTRODUCTION Optical Tomography (OT) is a novel medical imaging modality that involves the delivery of near-infrared (NIR) light into tissue and measuring the transmitted and reflected light intensities at multiple positions along the surface [1,2]. Using multiple wavelengths one can reconstruct the data and gain insight into the distribution of chromophores such as oxy- and deoxy-hemoglobin, lipid concentration and water [3,4]. Using information about the chromophore distributions, OT can detect changes in metabolic activity and vascularization related to growing tumors [3,5]. As a result, OT has the potential to impact the early detection and diagnosis of breast cancer. In recent years there have been a number of clinical studies published from research groups around the world that use optical tomography for the detection of breast tumors. For example, Rinneberg et al and Grosenick et al, both with the Physikalische- Technische Bundesanstalt, have published clinical results using a dual wavelength time- domain optical mammography system [6,7]. Culver et al have developed a combined frequency domain and continuous-wave clinical system that uses six wavelengths [8], and Enfield et al published results from a three- ! Manuscript received April 15, 2008. This work was supported in part by the Technology Transfer Incentive Program (TTIP) of the New York State Office of Science, Technology and Academic Research (NYSTAR) in partnership with NIRx Medical Technologies. M. L. Flexman, Y. Li, A. Bur, C. Fong, and J. Masciotti are with the Biomedical Engineering Department, Columbia University, New York, NY 10027 USA (phone: 212-342-0086; e-mail: mlf2129@columbia.edu). R. Al Abdi and R. Barbour are with the Department of Pathology, State University of New York - Downstate Medical Center, Brooklyn, NY 11203 USA (e-mail: Randall.Barbour@downstate.edu). A. H. Hielscher is with the Radiology Department and the Biomedical Engineering Department, Columbia University, New York, NY 10027 USA (e-mail: ahh2004@columbia.edu) dimensional time-resolved optical mammography study [9]. Most recently, Boverman et al presented a diffuse-optical- tomography study that explored hemoglobin levels in the breast [10]. Many of the existing optical tomography systems utilize analog technology. Generally, compared to digital systems, analog systems have more noise due to external noise sources such as power supplies, coupling from other channels, and electronic non-linearities [3]. To reduce these effects, lock-in detection algorithms are typically employed, which can separate the system response from the background noise by modulating the illumination sources and synchronously demodulating the detected signal [3,11]. It has been shown that lock-in detection algorithms that are implemented using digital devices are faster and have less electronic nonlinearities, less signal drift, and less gain errors than analog lock-in detection implementations [12]. In addition, digital systems have excellent scalability as well as improved temporal resolution and measurement accuracy. This paper discusses the design of a dynamic breast imaging device that makes use of the benefits of digital electronics. One of the advantages of this design is that it uses a large number of sources and detectors while still attaining fast imaging rates. Some of the challenges of this design include the timing of the light delivery and acquisition, the streamlined processing of such a large number of source and detector points, and the creation of a user interface that can seamlessly transition into the clinic. We have achieved considerably better dynamic range and can acquire data at higher speeds and with better precision than comparable analog electronics-based systems. II.SYSTEM DESIGN A.System Overview The system involves three main components: a light delivery unit, a control unit, and a detection unit. The input unit contains 4 laser diodes that deliver NIR light at 4 wavelengths to a switch that moves the light between 32 different sources. These sources simultaneously illuminate both the left and right breasts. The sources and detectors are brought into contact with the breast using a newly designed adaptive dual breast measuring head. Each measuring head uses eight, precision linear translating articulating arms that have a pivoting fiber-optic support that can contour to various breast geometries. The measuring head also brings 64 detectors into contact with each breast for a total of 128 detectors. As shown in Table I, with 32 sources, 128 detectors, and 4 wavelengths The Design and Characterization of a Digital Optical Breast Cancer Imaging System Molly L. Flexman, Member, IEEE, Yang Li, Andres M. Bur, Christopher J. Fong, James M. Masciotti, Member, IEEE, Rabah Al Abdi, Randall L. Barbour, and Andreas H. Hielscher, Member, IEEE 30th Annual International IEEE EMBS Conference Vancouver, British Columbia, Canada, August 20-24, 2008 978-1-4244-1815-2/08/$25.00 ©2008 IEEE. 3735 Authorized licensed use limited to: Columbia University. Downloaded on July 22, 2009 at 18:24 from IEEE Xplore. Restrictions apply.