Imaging of vascular dynamics with in the foot using dynamic...

Imaging of vascular dynamics within the foot using dynamic diffuse optical tomography to diagnose peripheral arterial disease

M. A. Khalil*a, H. K. Kimb, J. W. Hoia, I. Kimc, R. Dayalc, G. Shrikhandec, A. H. Hielschera,b,d aDept. of Biomedical Engineering, Columbia University, New York, NY, 10027; bDept. of Radiology, Columbia University, New York, NY, 10039; cDept. of Vascular Surgery, New York Presbyterian, New York, NY, 10039, dDept. of Electrical Engineering, Columbia University, New York, NY, 10027

ABSTRACT

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. In this paper we present our results in a pilot study to deduce optical tomography’s ability to detect poor blood perfusion in the foot. We performed an IRB approved 30 patient study, where we imaged the feet of the enrolled patients during a five stage dynamic imaging sequence. The patients were split up into three groups: 10 healthy subjects, 10 PAD patients and 10 PAD patients with diabetes and they were imaged while applying a pressure cuff to their thigh. Differences in the magnitude of blood pooling in the foot and rate at which the blood pools in the foot are all indicative of arterial disease. 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 is associated with risk factors including, smoking, hypertension, hyperlipidemia, hypercholestorlemia and diabetes. Symptoms of the disease vary depending on the degree of occlusion, initially the patient will experience pain while walking known as claudication. In more severe cases, this pain can occur during rest and if left untreated can lead to ulceration, gangrene and ultimately amputation. Early diagnosis is essential as lifestyle changes as well as cholesterol reducing drugs (statins) and blood thinners can be used to help control the disease [1, 2].

1.2 Current Diagnostic Techniques 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, which falsely elevates the ABI recording resulting in a false negative reading. The next diagnostic testing the duplex ultrasound scan, which is a combination of B-mode imaging and Doppler ultrasound. The technician will look for occlusions and listen for bruits (“wooshing” sounds) resulting from blood flow through narrowed arteries. Duplex is used to identify lesions within the leg, however the infra-popliteal arteries (below the knee) are small and the results become heavily user dependent. Furthermore, indentifying lesions does not provide direct information on the perfusion of the foot. More invasive imaging techniques such as x-ray computed tomography angiograms (CTA) and magnetic resonance angiography (MRA) are also used pre-surgically to determine the locations of lesions. However, these techniques require the use of contrast agents that are nephrotoxic which renders them dangerous to use for diabetic patients and patients with renal insufficiency [3-5]. Furthermore, these anatomical imaging methods do not provide information on the perfusion response within the foot. We believe dynamic diffuse optical tomography (DDOT) will provide effective way to view foot perfusion and aid in filling the gaps left by the traditional diagnostic techniques [6].

Optical Tomography and Spectroscopy of Tissue X, edited by Bruce J. Tromberg, Arjun G. Yodh, Eva Marie Sevick-Muraca, Proc. of SPIE Vol. 8578, 85781M

© 2013 SPIE · CCC code: 1605-7422/13/$18 · doi: 10.1117/12.2003399

Proc. of SPIE Vol. 8578 85781M-1

Downloaded From: http://proceedings.spiedigitallibrary.org/ on 07/23/2013 Terms of Use: http://spiedl.org/terms

L iïl ïìïiiï'ï'LIì

TIM

.

m-1 w -onlr-rl`;

IIE t

tITf IrerPt "VW:°ßiìí,

40. t 44 ® ;11101$ 14.1111111111iiii_++°II-iiih::

1.3 DynamDiffuse Opticis shone at dwavelength rtissue and thchromophoretissue withouimaging studmedical applgreat promise

2.1 InstruTo perform tsystem is cocombined inencompassinthe waveleng

The measuriUtilizing a sdesign gives foot. To obta1, using two the foot boun

mic Diffuse Op

cal tomographydifferent projerange is uniquehe NIR light es that absorb ut the use of a cdies that can plications such ae to be used wi

umentation

the imaging stmprised of tw

nto a single beng the foot. Thgths are separat

ng probe was sandal-shaped

each fiber theain the mesh foreferences a c

ndary.

ptical Tomogr

y (DOT) is novections encompe because that is non-ionizinthe light are o

contrast agent. provide functioas rheumatoid ithin a clinical

tudies we usedwo wavelengths

eam. This behis light shone ted using synch

Figure

designed to adesign, optica

e ability to tranor the image reoordinate syste

raphy

vel imaging tepassing some

absorption byg enabling it oxy and deoxyFurthermore,

onal informatioarthritis, breasimaging moda

2. M

d a digital dyns (760 and 830eam is de-mularound the foo

hronous detect

1. Imaging sys

accommodate al fibers are gunslate differenteconstruction aem can be mad

chnique in whtissue of inter

y the tissue is rto be used fr

y hemoglobin it can be geareon on tissue h

st cancer, brainality [7-12].

METHODS

namic diffuse o0nm), which altiplexed amonot and upon extion techniques

tem set-up with

the various shuided towards t lengths to m

a photograph isde to identify th

hich red and nerest in order torelatively low requently for menabling DOT

ed for fast acquhemodynamicsn imaging and s

S

optical tomograre modulated ng 16 source xiting the tissus [14].

measuring prob

hapes and sizethe foot via

make contact ws taken of the pthe locations th

ear infrared (NIo probe its opallowing it to

monitoring. IT to image theuisitions speedss. This technolsmall animal im

raphy system dwith a low frfibers placed

ue detector fibe

be.

es of feet seenspring-loaded

with the non-unpatient’s foot suhe sources and

IR) light (650-ptical propertieo penetrate deeIn addition, the blood contens allowing for logy has beenmaging and ha

described in [1requency (5-7k

at different lers collect the l

n in a clinicalshoulder screw

niform curvaturuch as shown idetectors and

-900 nm) es. This ep within he major nt within dynamic

n used in as shown

13]. This kHz) and locations light and

l setting. ws. This re of the in Figure segment



2.2 Image Reconstruction To generate the two-dimensional reconstructions of the optical properties in the foot, a diffusion-theory-based PDE-constrained multispectral image reconstruction scheme was employed [10]. This method solves the forward problem (boundary radiance at each wavelength) and the inverse problem (spatial distribution of chromophores concentrations) simultaneously using a reduced Hessian sequential quadratic programming (rSQP) method [10, 15]. This scheme directly reconstructs the spatial distributions of the oxy and deoxy-hemoglobin concentrations in the foot. Note that the differences in [HbO2] and [Hb] obtained through reconstruction is relative to baseline which is assumed to be given by [HbO2] = 23.43μM and [Hb] = 14.69μM, throughout the foot. A radial basis function (RBF)-type regularization scheme is employed to obtain quality images by reducing image noise such as artifacts near the foot surface. More details about this code can be found in [15].

2.3 Measurement Protocol The data presented were a result of a pilot study performed at the New York-Presbyterian Hospital–Columbia (NYP). The institutional review board (IRB) of the NYP approved the human subject protocol and written consents were obtained from all patients. During the imaging protocol the subjects were asked to place their foot on the patient interface (Figure 1) while sitting upright in a chair. A total of 34 fibers (14 source and 20 detection fibers) encompassed the foot, forming a coronal cross-section at the mid-metatarsal level. This location was chosen because it contains the major arteries of the foot the dorsalis pedis and the major branches of the posterior tibial artery. Physicians can probe these arteries when measuring a patient’s ABI. Furthermore, it is a common location for diabetic foot ulcers to occur and the vasculature in that region is too small to diagnose even with the high resolution anatomical imaging modalities such as CTA and MRA. After positioning the probe around the patient’s foot, the instrument automatically determined and stored the ideal gain settings for each channel at every source position. To illicit a controlled vascular response a pressure cuff was applied to the upper thigh. Patients are familiar with leg cuffs from ABI measurements, making this dynamic protocol a natural extension of the existing diagnostic procedures. The protocol consists of five stages. First a baseline measurement was taken while the patient is seated at rest for approximately one minute. Second, the pressure cuff is inflated to 60mmHg around the thigh. This induces venous occlusion, allowing arteries to supply blood the foot but preventing the veins from returning it to the heart, causing the blood to pool in the foot. The pressure was maintained for one minute after which it was rapidly released. During the third stage of the sequence, the foot was left to recover for one minute. In the fourth portion of this protocol a 120mmHg was applied to the thigh inducing greater venous occlusion for one minute. Then the pressure was released enabling the foot was left to recover. This five-stage protocol was applied three times for each subject to show repeatability.

3. RESULTS 3.1 Time Traces We present three select cases, one patient with PAD (ABI = 0.66), one patient with diabetes and PAD (ABI = 1.07) and one healthy subject (ABI = 1.00). Using the traditional ABI measurement it is possible to discern between the healthy control and the PAD patient, however the diabetic PAD patient is not distinguishable. Diabetic patients often have incompressible arteries due to calcifications which result in elevated ABI readings and leads to false negative diagnoses. Figure 2 shows the detector intensity measurements for a single source position for the healthy, PAD patient and Diabetic PAD patient respectively. We can view the five stages of the imaging procedure within the raw detector readings. Initially there is a baseline period then upon application of the 60mmHg pressure cuff venous occlusion is induced. The blood pumps from the heart into the foot but since the veins are occluded it does not return but pools within the foot. This causes greater attenuation within the foot as the hemoglobin absorbs the light. This causes a dip in the detected intensity. The pressure cuff is then released and the blood returns to the heart and the light intensity returns to initial value during rest. Then when the pressure cuff is reapplied with greater magnitude of 120mmHg there was more blood that pooled in the legs causing a greater magnitude drop in detector intensity. Upon release the signal returned to its baseline rest state.



r

3 4

C

r

1

I )eihelic

2

B

1

PAD: Raw Trace

3 4

3 4rime (min)

5

A

2 3

He

1 i 2

We observe and the diabephases 2-3 anthe affected impedance carterial comppatient suffer

Figure 2Illuminatwere filte60mmHglight absocuff. Thethe thigh

3.2 ImageUsing mo

change in totdynamics dusignificant dipressure, andpatient and ththe healthy vvenous occluin the affectecompliance.

a greater ampletic PAD patiennd 4-5). Furthevasculature pa

caused by plaqpliance. DDOTred from calcif

. Normalized Dion over time foered with a 50-pg cuff is applied orption. Upon rere are clear diffecuff.

e Reconstructi

odels of light ptal hemoglobinring the imaginifferences betwd before the rehe two PAD p

volunteer. This usion. Physioloed vasculature

litude drops wint (20% and 20ermore, there aatients during aque within theT was capable fied arteries, th

Detector Readinor (A) a Healthy point moving avto thigh. This im

elease of the cuferences in the m

ions

propagation in n within the foong sequence w

ween the three elease of the 1atients. The chsuggests that t

ogically we susas well as arter

ithin the health0%), before th

appears to be a application ande vascular lum

of discerning he traditional A

ngs (different dvolunteer, (B) P

verage filter. A mpedes blood frff the signal retu

magnitude drop d

tissue [10, 15ot for each timewe see that the

cases. Figure 320mmHg pres

hanges in [HbTthere is significspect that the drial stiffening d

hy subject (20%e release of theslower rate of

d release of thmen as well as

between theseABI measureme

detector positionPAD patient, andip in intensity

om returning to urns to baseline during thigh occl

5] we were able point during two time poin3 shows reconssure. These imT] are much lecantly less bloodifferential is ddue to calcifica

% and 45%) the 60mmHg andf decay and reche thigh cuff. Ws calcificationse three case stuent however re

ns shown as cnd (C) a Diabeticy is observed atthe heart and cathe same behavlusions as well a

le to obtain twthe imaging pr

nts prior to relenstruction resulmages show cess pronouncedod pooling in Pdue to the blooations within th

han the PAD pd the 120mmHcovery of the dWe suspect this within the arudies even thosulted in a fals

olored lines) foc PAD patient. Tt approximately auses it to pool

vior can be seen as the recovery r

wo-dimensionalrotocol. Lookinease of the thiglts before the rlear difference

d in the feet of PAD patients’ od flow impedhe arterial wal

patient (10% aHg thigh cuffs (detector readingis is due to blorterial wall deough the diabese-negative rea

for a Single SoThe individual tr

1 minute whenin the leg increawith the 120mm

rates from releas

l reconstructionng at the bloodgh cuffs show release of the 6es between thef the PAD patiefeet during a m

dance caused bl lowering the

and 15%) (between gs within ood flow ecreasing etic PAD ading.

ource races n the asing mHg se of

ns of the d volume the most

60mmHg e healthy ents than minute of y plaque vascular



2 3Time (min)

4

A

Weighted Aver

3

It

1 2

B

1 2T

igee[HbT]

i4 5i

A

60 r

A

3

I1

ime3

min)

nmHg 12

`

-

5

iabetic PAD: Wf

1 2

0 mmHg

sighted Average

3 4

e

me34mi n)

ibT(%)

26

17

9

0

r10

1

PAD:

2

Figure 3. patient. Th

To further qudynamic imaamount of bltotal hemoglconcordance 4(B), 4(C)) assertion thatarterial stiffe

Fig. 4. Arpatient. Wthe patient

We reported and a patient

2D Reconstructiohere is more hemoguantify the hemaging sequencelood within thelobin in the with the time

show slower ot plaque in thening due to cal

rea weighted AverWe recover the fivets with affected va

on dynamic dt with both PA

ons of the Total Hglobin present withmodynamic che and obtainede foot increasefoot during the-dependent deocclusion and

e arterial walls lcifications wit

rage Total Hemoge phases seen in thsculature.

diffuse optical AD and diabete

Hemoglobin Conchin the healthy vo

hange within thd a area weighd when the thihigh occlusionetector readingrecovery rateis causing blothin the arteria

globin Signal Healhe raw detector tim

4. COimaging (DDO

es. DDOT wa

entration [HbT] flunteer than the twhe feet we calchted average sigh cuff was apns is greatest gs. Furthermors than the heaod flow impedl wall is loweri

thy for (A) Healthme traces and obse

ONCLUSIOOT) results obas used to show

for Healthy volunwo patient cross seculated the sumsignal (Figure pplied to the fo

within the hre, the patientsalthy volunteerdance caused lering the vascula

hy Volunteer, (B) erve more blood p

ON btained for onew the hemodyn

nteer, PAD patienections shown belom of the [HbT4). These tim

foot (phases 2 ahealthy volunts with diseaser. This correlaess blood to poar compliance.

a PAD patient, anpooling within the

e healthy volunnamic respons

nt and a Diabetic ow it. T] in each fram

me traces show and 4). The incteer (Figure 4d vasculature ates with our ool in the foot

nd (C) a Diabetic healthy volunteer

nteer, one PADes observed w

PAD

me of the that the

crease in 4(A)), in (Figures previous and that

PAD

r than

D patient within the

foot while providing cross-sectional images that correspond to the foot vasculature. We found differences between all



three cases in the magnitude of the detector intensity drop during thigh cuff occlusion and the weighted average change in [HbT] signal obtained from the image reconstructions. In addition, DDOT was capable of discerning between the diabetic patient’s vasculature, despite their arterial calcifications, which render the traditional diagnostic methods inapt. These preliminary results show that DDOT has the potential to aid in the diagnosis and monitoring of PAD. Furthermore it has the potential to fill the diagnostic gap that currently exists within the diabetic patient population.

ACKNOWLEDGMENT This work was funded in part by the Wallace H. Coulter Foundation, the National Science Foundation Graduate Research Fellowship, the National Science Foundation IGERT for Optical Techniques for Actuation, Sensing, and Imaging of Biological Systems, and the Society of Vascular Surgery.

REFERENCES [1] A. T. Hirsch, M. H. Criqui, D. Treat-Jacobson, J. G. Regensteiner, M. A. Creager, J. W. Olin, S. H. Krook, D. B. Hunninghake, A. J. Comerota, M. E. Walsh, M. M. McDermott, and W. R. Hiatt, "Peripheral arterial disease detection, awareness, and treatment in primary care," JAMA : the journal of the American Medical Association 286, 1317-1324 (2001). [2] W. R. Hiatt, "Medical treatment of peripheral arterial disease and claudication," The New England journal of medicine 344, 1608-1621 (2001). [3] D. M. N. M. Al-Qaisi, D.H. King, S. Kaddoura and M. Hamady "Imaging of peripheral vascular disease " Reports in Medical Imaging 2 25-34 (2009). [4] A. Krnic, N. Vucic, and Z. Sucic, "Duplex scanning compared with intra-arterial angiography in diagnosing peripheral arterial disease: three analytical approaches," VASA. Zeitschrift fur Gefasskrankheiten 35, 86-91 (2006). [5] A. W. Pollak, P. T. Norton, and C. M. Kramer, "Multimodality imaging of lower extremity peripheral arterial disease: current role and future directions," Circulation. Cardiovascular imaging 5, 797-807 (2012). [6] 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). [7] A. H. Hielscher, "Optical tomographic imaging of small animals," Current opinion in biotechnology 16, 79-88 (2005). [8] U. J. Netz, J. Beuthan, and A. H. Hielscher, "Multipixel system for gigahertz frequency-domain optical imaging of finger joints," The Review of scientific instruments 79, 034301 (2008). [9] 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). [10] 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). [11] D. K. Joseph, T. J. Huppert, M. A. Franceschini, and D. A. Boas, "Diffuse optical tomography system to image brain activation with improved spatial resolution and validation with functional magnetic resonance imaging," Appl Opt 45, 8142-8151 (2006). [12] H. Dehghani, B. R. White, B. W. Zeff, A. Tizzard, and J. P. Culver, "Depth sensitivity and image reconstruction analysis of dense imaging arrays for mapping brain function with diffuse optical tomography," Appl Opt 48, D137-143 (2009). [13] 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). [14] 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). [15] H. K. Kim, M. Flexman, D. J. Yamashiro, J. J. Kandel, and A. H. Hielscher, "PDE-constrained multispectral imaging of tissue chromophores with the equation of radiative transfer," Biomedical optics express 1, 812-824 (2010).



Imaging of vascular dynamics with in the foot using dynamic...

Documents

Transcript of Imaging of vascular dynamics with in the foot using dynamic...