Dynamic contact-free continuous-wave diffuse optical tomography system for the detection of vascular dynamics within the foot M. A. Khalil*a, J. Hoib, H. K. Kimb, A. H. Hielschera,b,c a Dept. of Biomedical Engineering, Columbia University, New York, NY, 10027; bDept. of Radiology&ROXPELD8QLYHUVLW\, New York, NY, 10039; cDept. of Electrical Engineering &ROXPELD8QLYHUVLW\, New York, NY, 10027 ABSTRACT We present a dynamic contact-free continuous-wave diffuse optical tomography system for the detection and monitoring of peripheral arterial disease (PAD) in the foot. Peripheral Arterial Disease (PAD) is the narrowing of the functional area of the artery generally due to atherosclerosis. It affects between 8-12 million people in the United States and if untreated this can lead to ulceration, gangrene and ultimately amputation. Contact-Free imaging is highly desirable, due to the presence of ulcerations and gangrene in many patients affected by PAD. The system uses an electron multiplying charge coupled device (EMCCD) camera for detection to achieve a dynamic range of 86 dB with a frame rate of 1 Hz using 20 collimated source fibers and 2 wavelengths. We present first clinical results showing 3D images of total hemoglobin changes in response to a dynamic thigh cuff. Keywords: Peripheral arterial disease, dynamic imaging, diffuse optical tomography, vascular dynamics.

1

INTRODUCTION

1.1 Prevalence Peripheral Arterial Disease (PAD) is the narrowing of the functional area of the artery generally due to atherosclerosis. It affects between 8-12 million people in the United States and if untreated this can lead to ulceration, gangrene and ultimately amputation. The current diagnostic method for PAD is the ankle-brachial index (ABI). The ABI is a ratio of the patient’s systolic blood pressure in the foot to that of the brachial artery in the arm, a ratio below 0.9 is indicative of affected vasculature. However, this method is ineffective in patients with calcified arteries (diabetic and end-stage renal failure patients), which falsely elevates the ABI recording resulting in a false negative reading. Several groups have started to apply optical imaging techniques to address problems in PAD. Optical imaging techniques can non-invasively measure blood in tissue without the need for contrast agents or ionizing radiation. To date, promising initial results have been shown using Near Infrared Spectroscopy (NIRS) to detect PAD, as well as monitor wound healing within the foot [1, 2]. However, in NIRS, measurements are performed in a reflection-based geometry at isolated points on the surface of the tissue, which makes the technique heavily dependent on the location of the probe. We have shown that dynamic diffuse optical tomography (DDOT) has great potential for use in diagnosis and monitoring of PAD, unlike NIRS, DDOT provide the physician with an image of the blood flowing within the foot, providing for better spatial information about the perfusion within the foot [3]. Diffuse Optical tomography (DOT) is novel imaging technique in which red and near infrared (NIR) light (650-900 nm) is shone at different projections encompassing some tissue of interest in order to probe its optical properties. This wavelength range is unique because that absorption by the tissue is relatively low allowing it to penetrate deep within tissue and the NIR light is non-ionizing enabling it to be used frequently for monitoring. In addition, the major chromophores that absorb the light are oxy and deoxy hemoglobin enabling DOT to image the blood content within tissue without the use of a contrast agent. There are three different modes in which diffuse optical tomography can be conducted, time-domain imaging [4-9], frequency domain imaging [10-14], and continuous wave imaging [15-19]. All these methods have advantages and disadvantages in terms of hardware complexity, amount of information, and acquisition speeds. In continuous wave (CW) imaging, an un-modulated, DC light source is used to illuminate the tissue. Only the amplitude difference is collected and analyzed. This is the simplest of the three types of DOT imaging, however it provides significantly less information resulting in cross-talk between the effects of scattering and absorption. The simplicity of the electronics allows for much more rapid acquisition and makes CW imaging ideal of dynamic Optical Tomography and Spectroscopy of Tissue X, edited by Bruce J. Tromberg, Arjun G. Yodh, Eva Marie Sevick-Muraca, Proc. of SPIE Vol. 8578, 85781H © 2013 SPIE · CCC code: 1605-7422/13/$18 · doi: 10.1117/12.2003424 Proc. of SPIE Vol. 8578 85781H-1 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

imaging. With instrumentation [15, 16, 19, 20] geared specifically to the fast data acquisition rates required for dynamic imaging, thiss technique hass increasingly been b applied to other clinicaally relevant arreas, such as brreast imaging [15, 20], arthritis detecction [17] and cancer research in small anim mals [21, 22]. s that DD DOT has greatt potential, how wever we had ddifficult imaginng some patiennts due to XXX X, which In XXX we showed may mitigatee the use of PA AD in post-surg gical monitoring g. g makiing the imagingg process very difficult. Based on results ggenerated PAD patientss often have ulcerations and gangrene under specifiic aim 1, I startted to develop a novel dedicaated foot imagiing system. Thhe major novellties and improovements include (a) a contact-free imaging interface (b), Threee-dimensionall imaging capaabilities (c) caamera-based ddetection. After compleeting the 30-paatient pilot stud dy it became apparent a that inn order for thiss to be a cliniccal imaging syystem the fibers should d not have to be b individually adjusted. Thiis is both time consuming, aand introduces a large sourcee of error within the im maging protoco ol as an air tisssue boundary y between the fiber and the foot will introoduce artifacts into the results. With h this in mind d we began con nstructing a co ontact free imaaging system uusing collimatted source fibeers and a high-end elecctron multiplyiing charge coup pled device cam mera intended for operation w within the nearr infrared range.

2 2.1

METHOD DS

Overv view

L

E

> > >

Figurre 6 This figure outlines o the interaactions between th he major componeents within the im maging system. T Two laser drivers regullate the intensity and a temperature off the two waveleng gth of light used. T These two waveleengths are input innto a 2x20 Optical Switcch, which de-mulltiplexes the light among 20 collim mated optical fiberrs. Using an RS2 32 serial port andd LabVIEW user interfface controls the optical o switch, enaabling it to distrib bute the light and alternate the two wavelengths at eaach location. The opticcal fibers then shin ne light on differeent locations within n the surface of thhe foot. The lightt is scattered and absorbed and the transmitted light is bou unced of a 45 degrree protected silverr coated mirror intto an EMCCD cam mera which stores an image of light W user interface. transmitted from each source illumination. The camera is controlled using a GigE interface ussing the LabVIEW

2.2

Inputt Unit

The input un nit is compriseed of two 350 0mW CW laseer diodes at 6880 and 860nm m wavelengthss (Intense Inc). These wavelengths were chosen to maximize th he separation between absorpption and scatteering [23]. Thhese diodes havve a built ( unit an nd come in a high heat loaad fiber packaage (HHLF). T These laser dioodes are in thermoeleectric control (TEC) regulated in intensity i and temperature by y two Thorlabs OEM laser drrivers. The laseer drivers are aattached to a baackplane via a 64-pin input/output jack, the backp plane, which in turn wires th the inputs and outputs betweeen the diodess and the

Proc. of SPIE Vol. 8578 85781H-2 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

drivers. It alsso provides po ower to the laser drivers via an Internationnal Power openn frame DC poower supply (M Marlin P. Jones). This power p supply is i capable of providing 3A lo oads on both a positive and negative 12V raail enabling it tto supply the laser driv vers at peak inteensity. velengths are de-multiplexed d d unto 20 grad ded-index multtimode fiber ppatch cables. This fiber wass chosen The two wav because of itts low attenuaation in the NIIR (1-5 dB/km m). These souurce patch cablles begin and terminate withh angled physical conttact fiber optic connectors (FC/APC to FC//APC). A very fast microelecctromechanicall system (MEM MS) 2x20 multimode optical o switch distributes d the two waveleng gths among opptical fibers. Thhis switch is ccapable of swittching at 5ms while maintaining m a lo ow insertion lo oss of < 1.0 dB B and a cross taalk of -60dB between the inddividual channels. The switch operaates as if it iss 2x1 switch attached a seriallly to a 1x20 switch. Thee 2x1 switch ttoggles betweeen the 2 wavelengths and the 1x20 switch toggless between the 20 2 different soource fibers thaat illuminate thhe foot (Figuree 1). The nnected to the computer by an a RS232 seriaal port and opeerated throughh LabVIEW sooftware. The seerial port switch is con commands are a sent to the switch via a LabVIEW L useer interface, thhis allows for tthe scanning oof the light am mong the different fibeers as well as th he alteration beetween the two o wavelengths of light. The switch operatees by time multtiplexing the two waveelengths upon a single locatiion then proceeeding to movee onto the nextt location (i.e. Source1 Waveelength1, Source1 Wav velength2, Sou urce2 Waveleng gth1, Source2 Wavelength2, W ...etc.).

>

2X20

Optical Switc _C

u

o N X

a-i

co

V O.

860nmI;

>

nputer

Figurre 1 The MEMS 2x20 Optical Switch is responsiblee for de-multiplexxing the light from m the two waveleengths among 20 opticcal fibers that distrribute the light upon the surface of the foot. The swiitch is comprised of a 2x1 optical switch attached in serial to a 1x20 opticaal switch. The firsst switch selects which w wavelength to use and the seccond is responsiblle for distributing that wavelength w among g the source fiberss. The switch is programmed p to opperate by shining bboth wavelengths tthrough one fiber then continuing to the next n fiber location n.

ut from the 2x x20 Optical sw witch are each attached to ann FC/APC fixeed wavelengthh 780nmThe 20 sourcce fibers outpu aligned asph heric lens fiberr collimators. Collimators are a wavelengthh dependent, ttherefore sincee two wavelenngths are being output through the saame fiber a traadeoff must be made as to whhich wavelenggth the collimator is aligned ffor. The mators were ch hosen in orderr to allow thee minimum diistortion for bboth wavelenggths. The 780nm waveelength collim collimators in ncrease the sizze 680nm spot size, as the diistance increasses from the fibber and decreaase the spot sizze of the 860nm with distance. How wever, as the distance d betweeen the collimattor and the fooot is less than ssix inches this effect is des give a spott size of approx ximately 3mm.. negligible as both laser diod 2.3

Detecction Unit

The detection n unit is comprrised of a 2in x 6in, protected d silver flat miirror (Rainbow w Optics) and aan electron muultiplying charge couplled device (EM MCCD) cameraa. The light from the collimaated fibers is ttransmitted through the foot and then reflects off th he 45 degree protected p silverr mirror. The mirror m has a 99 .9% reflectivitty in the NIR w wavelengths. T The light

Proc. of SPIE Vol. 8578 85781H-3 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

reflects off th hese mirrors an nd onto a Princceton Instrumen nts ProEM bacck illuminated EMCCD cameera. This cameera has a high dynamic range of 16b bits (96dB), as well as, a fastt acquisition raate (34fps withhout binning). The EMCCD D camera connects to the t computer via v a Gigabit Ethernet E (GigE)) interface andd is operated ussing a LabVIE EW user interfaace. The camera is paaired with a 14mm focal leng gth wide-anglee camera lens tto allow for a 6-inch wide viiew from a disstance of less than 15 inches i from thee mirror. 2.4

Laserr Scanner

To obtain the 3D geometry y of the foot, as well as the source detecttor positions, P PhotoModeler Scanner is sofftware is used. This so oftware is capaable of extractiing measuremeents and 3D m models from photographs. Muultiple photogrraphs are taken around d the patients’ foot along witth proprietary ‘RAD coded ttargets’ (Figurre 2). These tarrgets are autom matically located and marked in ph hotographs ussing the Photo oModeler softtware. The loccations of thee RAD targetts in the photographs provides the software s with the t position off the camera annd the angle att which the picctures were takken. The PhotoModeleer software theen proceeds to generate a thrree-dimensionaal surface meshh of the foot. U Using this infoormation we can then n use GID sofftware to geneerate a 3D volume mesh, w which can be input into thee image reconstruction algorithm.

Figure 2 Surface Mesh of Foot Gen nerated using PhottoModeler Scanneer Software.

2.5

Measuring Setup

The measurin ng probe is sett up to illuminate the two waavelengths of llight onto 20 ppositions on topp of the foot. T The light propagates th hrough the foott and is reflecteed off a protectted silver coateed mirror onto an EMCCD caamera. The m mirror and the EMCCD camera are plaaced inside a large l black rectangular box (336 in x 12 in x 9 in). This bbox is designedd to keep ambient lightt out and to atttenuate any straay reflections that t may arise.. It is imperativve that there bee no stray lightt seeping into the box during d imaging g, because the transmitted lig ght intensities aare very low annd would be loost due to restriictions in the dynamic range. To mitiigate the effectts of stray in th he box, there noo gaps betweenn the sides andd the materials used are black and non n-reflective.

Proc. of SPIE Vol. 8578 85781H-4 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

Figure 3 SolidWorks CAD Drawing of the measuring probe, the large box on the bottom contains the camera and a 45-degree silver coated mirror. The smaller box on top holds the optical fibers and can be elevated up and down via the two guide rails so as to allow for taking pictures of the foot. The interface is set on four casters to allow for mobility within the clinical setting.

The base of the rectangular box is made of a black solid aluminum breadboard (Thorlabs), the breadboard is compatible with many optical components and acts as a solid foundation for the imaging probe. The bottom of the box also contains four rubber caster wheels, which together are capable of withstanding more than 200 lbs. These wheels allow the probe to be more portable, to better accommodate use in a clinical setting. Off the aluminum breadboard are four 9in guide rails, which form the foundation of the box. Between each of these rails is a solid sheet of black acrylic used as the sides of the box. The black acrylic used was heavily polished and despite its color seemed to be rather reflective; so black felt material was pasted on the inside of the box to assure that there was no reflection occurring. The top piece of the rectangular box is a 34in x 12in piece of black acrylic that covers the entire top of the breadboard. This top piece contains a 2in x 6in slit to allow light that has transmitted through the foot to pass. A CAD drawing of the measurement enclosure is shown in Figure 3. On top of this large rectangular box is a smaller black box that contains the collimated fibers. The collimated fibers are screwed into the top face of this smaller black box. The top face contains 20 threaded holes in a 3 line staggered pattern (Figure 4). The small black box capable of gliding vertically along two 20inch-guide rails, which allows the user to take photographs of the foot without having the patients move their feet. These photographs are used in combination with the PhotoModeler scanning software to obtaining the geometry of the foot. In addition, also allows for aligning the foot with the slit in the larger rectangular box.

Proc. of SPIE Vol. 8578 85781H-5 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

Figurre 4 The above sh hows the source fib ber configuration on top of the smalll black box. The ccollimator lens sysstem are threaded on to op so they screws into the holes sho own above, which are laser cut on ttop of the small bllack box. The peeak of illuminated beam m is at the center of o the holes. The configuration c has three t rows of fiberrs, which are stagggered to fit multipple fiber locations withiin the small 2inch by 6inch slit in thee larger box.

The small blaack box encom mpasses the foo ot and holds thee fibers in placce during the im maging protocool. The side off the box facing the paatient has black k rubber strips covering the foot f entrance, enabling patieents to place thheir foot insidee without introducing much m ambient light. The patiient places his//her foot throuugh the rubber strips and theen covers the slit in the top piece of the larger recttangular box. The light then n is shown from m fibers on topp of the small black box, propagates through the foot f and through the slit into o the larger recctangular blackk box where itt reflects off thhe 45 degree pprotected silver coated mirror and intto the EMCCD D camera.

Figurre 5 Picture of thee built completed the black-box intterface is much m more atheistically ppleasing than the rright probes with screw w guided fibers alignment. It is frien ndly looking and requires significanttly less maintenannce than the fiber-bbased imaging.

ACKNO OWLEDGM MENT This work was w funded in n part by the Wallace H. Coulter C Foundaation, the Nattional Science Foundation G Graduate Research Fellowship, the National Science Foundatio on IGERT forr Optical Techhniques for Acctuation, Senssing, and Imaging of Biological B Systeems, and the Society of Vasccular Surgery.

Proc. of SPIE Vol. 8578 85781H-6 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

REFERENCES [1] U. Wolf, M. Wolf, J. H. Choi, M. Levi, D. Choudhury, S. Hull, D. Coussirat, L. A. Paunescu, L. P. Safonova, A. Michalos, W. W. Mantulin, and E. Gratton, "Localized irregularities in hemoglobin flow and oxygenation in calf muscle in patients with peripheral vascular disease detected with nearinfrared spectrophotometry," Journal of Vascular Surgery 37, 1017-1026 (2003). [2] C. Casavola, L. A. Paunescu, S. Fantini, and E. Gratton, "Blood flow and oxygen consumption with near-infrared spectroscopy and venous occlusion: spatial maps and the effect of time and pressure of inflation," J Biomed Opt 5, 269-276 (2000). [3] M. A. Khalil, H. K. Kim, I. K. Kim, M. Flexman, R. Dayal, G. Shrikhande, and A. H. Hielscher, "Dynamic diffuse optical tomography imaging of peripheral arterial disease," Biomedical optics express 3, 2288-2298 (2012). [4] E. Lapointe, J. Pichette, and Y. Berube-Lauziere, "A multi-view time-domain non-contact diffuse optical tomography scanner with dual wavelength detection for intrinsic and fluorescence small animal imaging," Review of Scientific Instruments 83(2012). [5] S. Okawa, Y. Hoshi, and Y. Yamada, "Improvement of image quality of time-domain diffuse optical tomography with l(p) sparsity regularization," Biomedical optics express 2, 3334-3348 (2011). [6] W. R. Mo and N. G. Chen, "Design of an Advanced Time-Domain Diffuse Optical Tomography System," Ieee J Sel Top Quant 16, 581-587 (2010). [7] W. R. Mo and N. G. Chen, "Fast time-domain diffuse optical tomography using pseudorandom bit sequences," Optics express 16, 13643-13650 (2008). [8] S. Lam, F. Lesage, and X. Intes, "Time domain fluorescent diffuse optical tomography," P Soc Photo-Opt Ins 5578, 179-187 (2004). [9] H. J. Zhao, F. Gao, Y. Tanikawa, K. Homma, Y. Onodera, and Y. Yamada, "Anatomical and functional images of in vitro and in vivo tissues by NIR time-domain diffuse optical tomography," Jsme Int J C-Mech Sy 45, 1033-1039 (2002). [10] B. Pogue, M. Testorf, T. McBride, U. Osterberg, and K. Paulsen, "Instrumentation and design of a frequency-domain diffuse optical tomography imager for breast cancer detection," Optics express 1, 391-403 (1997). [11] A. G. Orlova, I. V. Turchin, V. I. Plehanov, N. M. Shakhova, I. I. Fiks, M. I. Kleshnin, N. Y. Konuchenko, and V. A. Kamensky, "Frequencydomain diffuse optical tomography with single source-detector pair for breast cancer detection," Laser Phys Lett 5, 321-327 (2008). [12] J. P. Culver, R. Choe, M. J. Holboke, L. Zubkov, T. Durduran, A. Slemp, V. Ntziachristos, B. Chance, and A. G. Yodh, "Three-dimensional diffuse optical tomography in the parallel plane transmission geometry: Evaluation of a hybrid frequency domain/continuous wave clinical system for breast imaging," Med Phys 30, 235-247 (2003). [13] B. W. Pogue, H. B. Jiang, K. D. Paulsen, and U. L. Osterberg, "Frequency-domain diffuse optical tomography of breast tissue: detector size and imaging geometry," P Ann Int Ieee Embs 19, 2745-2749 (1997). [14] B. W. Pogue, M. S. Patterson, H. Jiang, and K. D. Paulsen, "Initial Assessment of a Simple System for Frequency-Domain Diffuse Optical Tomography," Physics in medicine and biology 40, 1709-1729 (1995). [15] M. L. Flexman, M. A. Khalil, R. Al Abdi, H. K. Kim, C. J. Fong, E. Desperito, D. L. Hershman, R. L. Barbour, and A. H. Hielscher, "Digital optical tomography system for dynamic breast imaging," J Biomed Opt 16, 076014 (2011). [16] J. M. Lasker, J. M. Masciotti, M. Schoenecker, C. H. Schmitz, and A. H. Hielscher, "Digital-signal-processor-based dynamic imaging system for optical tomography," The Review of scientific instruments 78, 083706 (2007). [17] J. M. Lasker, C. J. Fong, D. T. Ginat, E. Dwyer, and A. H. Hielscher, "Dynamic optical imaging of vascular and metabolic reactivity in rheumatoid joints," J Biomed Opt 12, 052001 (2007). [18] A. Bluestone, G. Abdoulaev, C. Schmitz, R. Barbour, and A. Hielscher, "Three-dimensional optical tomography of hemodynamics in the human head," Optics express 9, 272-286 (2001). [19] C. H. Schmitz, H. L. Graber, H. Luo, I. Arif, J. Hira, Y. Pei, A. Bluestone, S. Zhong, R. Andronica, I. Soller, N. Ramirez, S. L. Barbour, and R. L. Barbour, "Instrumentation and calibration protocol for imaging dynamic features in dense-scattering media by optical tomography," Appl Opt 39, 6466-6486 (2000). [20] C. M. Carpenter, R. Rakow-Penner, S. Jiang, B. W. Pogue, G. H. Glover, and K. D. Paulsen, "Monitoring of hemodynamic changes induced in the healthy breast through inspired gas stimuli with MR-guided diffuse optical imaging," Med Phys 37, 1638-1646 (2010). [21] A. Hielscher, "Optical tomographic imaging of small animals," Med Phys 32, 2096-2096 (2005). [22] A. Y. Bluestone, M. Stewart, B. Lei, I. S. Kass, J. Lasker, G. S. Abdoulaev, and A. H. Hielscher, "Three-dimensional optical tomographic brain imaging in small animals, part 2: unilateral carotid occlusion," J Biomed Opt 9, 1063-1073 (2004). [23] A. Corlu, T. Durduran, R. Choe, M. Schweiger, E. M. Hillman, S. R. Arridge, and A. G. Yodh, "Uniqueness and wavelength optimization in continuous-wave multispectral diffuse optical tomography," Optics letters 28, 2339-2341 (2003).

Proc. of SPIE Vol. 8578 85781H-7 Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms