Laser Speckle Contrast Images

Processing of Data to Analyze Microvascular Variations in Hand Skin of Healthy

Subjects

Anne Humeau-Heurtier

, J´er´emy Dexet

, Pierre Abraham

and Guillaume Mah´e

Universit´e d’Angers, LARIS - Laboratoire Angevin de Recherche en Ing´enierie des Syst`emes,

62 avenue Notre Dame du Lac, 49000 Angers, France

Universit´e d’Angers, CHU d’Angers, Laboratoire de Physiologie et d’Explorations Vasculaires,

UMR CNRS 6214-INSERM 1083, 49033 Angers cedex 01, France

Pˆole imagerie m´edicale et explorations fonctionnelles, CHU de Rennes, Universit´e de Rennes 1,

Hˆopital Pontchaillou, 35033 Rennes Cedex 9, France

Keywords:

Laser Speckle Contrast Imaging, Microvascular Blood Flow, Spatial Variation, Hand.

Abstract:

Diseases such as frostbite, Raynaud’s phenomenon, carpal tunnel syndrome, systemic sclerosis, hand burns,

hand ﬂaps or hand wounds lead to microvascular dysfunctions. These diseases can affect only one hand

and even sometimes only some ﬁngers. Therefore, in order to quantify the microvascular alterations, it can

be useful to compare microvascular perfusion of both hands (altered versus healthy hand). However, skin

microvasculature presents spatial variations and the possibility to directly compare perfusion values of the two

hands has not been studied yet. We therefore propose to quantify these spatial variations. For this purpose,

perfusion values from laser speckle contrast images (LSCI) of the two hands (dorsal faces) from seven young

healthy subjects are acquired simultaneously and then processed. The results show that the dorsal faces of

the two hands in our young healthy subjects present close perfusion values (average coefﬁcient of variation

less than 9% for our subjects). These ﬁndings are preliminary observations to quantify the microvascular

alterations in the above-mentioned diseases. The coefﬁcient of variation in diseased states remains to be

studied in order to see whether LSCI could be of interest to improve the diagnosis of hand skin pathologies.

1 INTRODUCTION

Microvascular dysfunction in hands is an early

stage in several pathologies. Thus, for diseases

such as frostbite (Foray, 1992), Raynaud’s phe-

nomenon (Brown, 2012; Gladue et al., 2013), carpal

tunnel syndrome (Shannon and Rizzolo, 2012), sys-

temic sclerosis (Wigley, 2009; Herrick, 2000), hand

burns (Meier et al., 2011; Barachini et al., 2004), hand

ﬂaps (Du et al., 2011; Soderstrom et al., 1999) or hand

wounds, hand skin microvasculatureis speciﬁcally af-

fected: morphological abnormalities of the capillaries

(enlarged loops, giant capillaries, ramiﬁcations, cap-

illary disorganization), microhemorrhages or lower

density (capillary loss) (DeAngelis et al., 2009). Al-

tered hand skin microvascular function could there-

fore be a surrogate marker for the above-mentioned

diseases. In order to quantify such vascular dysfunc-

tions, accurate and sensitive measurement techniques

have become a key issue.

Recently, the laser speckle contrast imag-

ing (LSCI) technique has been proposed for skin mi-

crocirculation monitoring (Humeau-Heurtier et al.,

2013a; Rege et al., 2012; Mahe et al., 2012b; Mahe

et al., 2012c; Miao et al., 2010; Draijer et al., 2009).

LSCI gives a 2D map of the microvascular blood

perfusion (see an example in Fig. 1). The princi-

ples of LSCI rely on the illumination of tissues (op-

tically rough surface) by a laser light. Raw speckle

reﬂectance images are collected with a CCD cam-

era. Motions (e.g., blood ﬂow) in the tissues alter the

speckle pattern over time. A faster blood ﬂow leads

to more blurring in the captured image than slower

ﬂow or no ﬂow. The degree of blurring is quanti-

ﬁed as the local speckle contrast value, with zero con-

trast (no speckle pattern) corresponding to high blood

ﬂow, and unity contrast (fully developed speckle pat-

tern) to no ﬂow. The speckle contrast K is deﬁned as

the ratio of the standard deviation to the mean inten-

sity hIi, as (Briers and Webster, 1996)

K =

hIi

, (1)

Humeau-Heurtier A., Dexet J., Abraham P. and Mahé G..

Laser Speckle Contrast Images - Processing of Data to Analyze Microvascular Variations in Hand Skin of Healthy Subjects.

DOI: 10.5220/0004892900550058

In Proceedings of the International Conference on Bioimaging (BIOIMAGING-2014), pages 55-58

ISBN: 978-989-758-014-7

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Figure 1: Laser speckle contrast image of a zone on the

dorsal face of the hand from a healthy subject.

where σ

refers to the spatial standard deviation of the

speckle intensity. The speckle contrast K is a function

of the exposure time T of the camera and is related to

the autocovariance of the intensity ﬂuctuations in a

single speckle (Fercher and Briers, 1981)

(T)

hIi

ThIi

(τ)dτ, (2)

whereC

(τ) is the autocovarianceof the intensity ﬂuc-

tuations in a single speckle. The contrast value is

used to form the microvascular blood perfusion im-

age (Draijer et al., 2009).

In diseases such as frostbite, Raynaud’s phe-

nomenon, carpal tunnel syndrome, systemic sclero-

sis, burns, ﬂaps and wounds, skin microvascular al-

terations and vascular dysfunctions may affect only

one hand (the diseased one). In order to quantify such

alterations and dysfunctions, it can be useful to com-

pare skin perfusion values of the diseased hand to skin

perfusion values of the healthy hand. However, due

to spatial variations of perfusion in skin, the question

relative to the possibility to directly compare LSCI

perfusion values of the two hands still has to be an-

swered. Thus, in the case of no disease, are perfusion

values of both hands similar? Are there variations

from one hand to the other? If yes, is it possible to

quantify them?

In order to quantify spatial variations between the

perfusion values of two hands from healthy subjects,

we herein propose to process laser speckle contrast

images acquired simultaneously in the two hands of

voluntary subjects.

2 MATERIALS AND METHODS

2.1 Measurement Procedure

Seven Caucasian healthy subjects participated in the

study (six men, one woman; 22.6 ±5 years old). All

the subjects provided written, informed consent prior

to participation and the study was carried out in accor-

dance with the Declaration of Helsinki. Each subject

was sitted confortably in a quiet room with controlled

temperature and without any air movements (Mahe

et al., 2012a). They were asked to position the ventral

face of each hand on a table. LSCI data of both hands

were acquired in laser speckle perfusion units (LSPU)

with a PeriCam PSI System (Perimed, Sweden) hav-

ing a laser wavelength of 785 nm and an exposure

time of 6 ms. The distance between the laser head to

skin was set around 24 cm (Mahe et al., 2011), which

gave an average image resolution of 0.6 mm. Laser

speckle contrast images were acquired and stored for

5 min with a sampling frequency of 18 Hz on a com-

puter and analyzed off-line.

2.2 Image Processing Procedure

After the acquisitions, ﬁve pixels on each hand were

randomly chosen on the ﬁrst image of the LSCI se-

quence. These pixels are thereafter noted as P

, with i

varying between 1 and 5. Then, for each pixel P

, the

amplitude value was followed with time on the image

sequence (5 min) to obtain a time-evolution signal,

noted thereafter as sP

(t). Around each of these ﬁve

pixels P

, ﬁve square regions of interest (ROI), each

of different size, were also determined: 3.2 mm

pixels×3 pixels), 29.1 mm

(9 pixels×9 pixels),

81.0 mm

(15 pixels×15 pixels), 190.4 mm

(23

pixels×23 pixels), 345.9 mm

(31 pixels×31 pix-

els) (see (Humeau-Heurtier et al., 2013b)). Pixel val-

ues inside each ROI were averaged and followed with

time on the image sequence (5 min) to obtain time-

evolution signals, noted thereafter as sROI

(t) for

the ROI j around the pixel P

, j corresponding to the

size of the square value (3, 9, 15, 23, or 31). Then, for

each hand, the ﬁve time-evolution signals sP

(t) were

averaged in time to obtain ﬁve LSCI perfusion val-

ues which were averaged together. The same was per-

formed for the ﬁve time-evolution signals sROI

(t).

For each hand and for each of the seven subjects,

we therefore had one LSCI perfusion value corre-

sponding to the ﬁve pixels and one LSCI perfusion

value corresponding to each of the ﬁve ROI sizes (see

Fig. 2).

The coefﬁcient of variation (CV) was computed

to compare the LSCI perfusion values between each

hand. The CV was calculated according to the proce-

dure proposed by Hopkins (Hopkins, 2000). For this

purpose, the perfusion value from the ﬁve pixels P

one hand was compared to the perfusion value from

the ﬁve pixels P

of the other hand. The same was

BIOIMAGING2014-InternationalConferenceonBioimaging

Table 1: Average perfusion values in LSPU for the 7 subjects for each ROI size and for the two hands. ROI sizes are in pixels

× pixels (see text for details).

Size of ROI Pixel 3×3 9×9 15×15 23×23 31×31

Left hand 65.0 64.6 64.0 63.8 64.3 65.9

Right hand 67.6 67.4 67.1 67.5 67.8 69.1

Average 66.3 66.0 65.6 65.6 66.1 67.5

Table 2: Coefﬁcient of variation (CV, in percentage) between the perfusion values of the two hands of healthy subjects. ROI

sizes are in pixels × pixels (see text for details).

Size of ROI Pixel 3×3 9×9 15×15 23×23 31×31

CV for subject 1 7.9 8.2 6.9 7.1 7.2 7.9

CV for subject 2 6.7 5.4 4.0 6.7 6.8 5.0

CV for subject 3 10.3 10.3 6.7 3.9 2.4 2.2

CV for subject 4 5.9 5.1 5.8 5.1 5.4 6.6

CV for subject 5 8.2 8.4 8.8 7.1 4.6 3.8

CV for subject 6 5.4 5.2 3.5 3.9 4.1 5.1

CV for subject 7 16.2 16.2 11.4 10.1 10.1 9.7

Average CV 8.7 8.4 6.7 6.3 5.8 5.8

pixel P

(1 ≤ i ≤ 5)

ROI around P

(t)

averaging in time

one perfusion

value for each pixel P

sROI

(t)

averaging in time

one perfusion

value for each ROI size and for

each pixel P

5 min

laser speckle contrast image

Figure 2: Schematic representation of the signals processed.

Only one hand is represented here. On our recordings, the

two hands were present on each image.

performed for each ROI size.

3 RESULTS AND DISCUSSION

The results obtained for the analysis of the spatial

variations are shown in Tables 1 and 2. From Ta-

ble 1 we observe that the average perfusion value is

66.1 LSPU when the two hands are considered to-

gether (average perfusion value between 65.6 LSPU

and 67.5 LSPU). When the two hands are considered

separately, we observe that the average perfusion val-

ues are verysimilar for both hands butthat the average

perfusion value is slightly higher for the right hand

than for the left hand, whatever the ROI size (see Ta-

ble 1). In that case, the standard deviation varies from

11.5 LSPU and 15.4 LSPU. The slightly higher value

for the perfusion in the right hand compared to the

one of the left hand may be due to the fact that all

our subjects were right-hander. To analyze these re-

sults, we computed the standard deviation of the per-

fusion values obtained on a white surface (same mea-

surement procedure as for the hands). For the same

ROI sizes as for the hands, we obtain standard devia-

tion values of the difference between the points in the

white surface between 0.7 LSPU and 4.3 LSPU (0.7

LSPU for a ROI of 31×31 pixels and 4.3 LSPU for a

ROI of the size of the pixel). For the white surface,

the standard deviation decreases for larger ROI sizes

since we have less noise in the perfusion values when

more data points are used for the analysis. These stan-

dard deviations on the white surface are much lower

than the ones found for the hands.

Moreover, from Table 2 we note that the CV be-

tween the perfusion values of the two hands varies be-

tween subjects but is always lower than 16.2% for a

ROI of the size of a pixel for a subject, and lower than

9.7% for a ROI size of 31 pixels×31 pixels for a sub-

ject. The average CV for the 7 subjects is low (< 9%)

for each size of ROI analyzed (see Table 2). Further-

more, the larger the ROI size (between 3.2 mm

and

345.9 mm

), the lower the average CV value.

All these results show that, despite the spatial

variations in skin microcirculation, the dorsal face of

the two hands in healthy subjects present close per-

fusion values (average CV < 9%). These ﬁndings

lead to the suggestion that, for patients with diseases

affecting one hand (e.g., frostbite, Raynaud’s phe-

nomenon, carpal tunnel syndrome, systemic sclero-

sis, hand burns, hand ﬂaps or hand wounds), it may

be possible to quantify skin microvascular alterations

and vascular dysfunctions by comparing laser speckle

contrast images perfusion of the two hands.

LaserSpeckleContrastImages-ProcessingofDatatoAnalyzeMicrovascularVariationsinHandSkinofHealthy

Subjects

4 CONCLUSIONS

Spatial variations in perfusion values from the two

hands of healthy subjects are herein studied from laser

speckle contrast images. We show that the dorsal face

of the two hands in healthy subjects present close

perfusion values (CV < 9.7% for a ROI size of 31

pixels×31 pixels). The possible increase of CV in

diseased states remains to be studied in order to see

whether LSCI could be of interest to improve the di-

agnosis of diseases such as frostbite, Raynaud’s phe-

nomenon, carpal tunnel syndrome, systemic sclerosis,

hand burns, hand ﬂaps or hand wounds.

ACKNOWLEDGEMENTS

The authors would like to thank Fredrik Salomonsson

from Perimed for its useful comments regarding the

analysis of the data on white surface.

REFERENCES

Barachini, P., Vezzoni, G. M., Palombo, C., Franzoni, F.,

and Bigalli, G. (2004). Skin blood ﬂow pattern in

burns outcomes. Burns, 30:312–316.

Briers, J. D. and Webster, S. (1996). Laser speckle contrast

analysis (lasca): a non- scanning, full-ﬁeld technique

for monitoring capillary blood ﬂow. J. Biomed. Opt.,

1:174–179.

Brown, S. (2012). Diagnosis and management of patients

with raynaud’s phenomenon. Nurs. Stand., 26:41–46.

DeAngelis, R., Grassi, W., and M.Cutolo (2009). A grow-

ing need for capillaroscopy in rheumatology. Arthritis

Rheum., 61:405–410.

Draijer, M., Hondebrink, E., van Leeuwen, T., and Steen-

bergen, W. (2009). Review of laser speckle contrast

techniques for visualizing tissue perfusion. Lasers

Med. Sci., 24:639–651.

Du, Z., Zan, T., Li, H., and Li, Q. (2011). A study of blood

ﬂow dynamics in ﬂap delay using the full-ﬁeld laser

perfusion imager. Microvasc. Res., 82:284–290.

Fercher, A. F. and Briers, J. D. (1981). Flow visualiza-

tion by means of single-exposure speckle photogra-

phy. Opt. Commun., 37:326–330.

Foray, J. (1992). Mountain frostbite. current trends in prog-

nosis and treatment (from results concerning 1261

cases). Int. J. Sports Med., 13:S193–S196.

Gladue, H., Maranian, P., Paulus, H. E., and Khanna, D.

(2013). Evaluation of test characteristics for outcome

measures used in raynaud’s phenomenon clinical tri-

als. Arthritis Care Res (Hoboken), 65:630–636.

Herrick, A. L. (2000). Vascular function in systemic scle-

rosis. Curr. Opin. Rheumatol., 12:527–533.

Hopkins, W. G. (2000). Measures of reliability in sports

medicine and science. Sports Medicine, 30:1–15.

Humeau-Heurtier, A., Guerreschi, E., Abraham, P., and

Mahe, G. (2013a). Relevance of laser doppler and

laser speckle techniques for assessing vascular func-

tion: state of the art and future trends. IEEE Trans.

Biomed. Eng., 60:659–666.

Humeau-Heurtier, A., Mahe, G., Durand, S., and Abraham,

P. (2013b). Multiscale entropy study of medical laser

speckle contrast images. IEEE Trans. Biomed. Eng.,

60:872–879.

Mahe, G., Durand, S., Humeau, A., Leftheriotis, G.,

Rousseau, P., and Abraham, P. (2012a). Air move-

ments interfere with laser speckle contrast imaging

recordings. Lasers Med. Sci., 27:1073–1076.

Mahe, G., Haj-Yassin, F., Rousseau, P., Humeau, A., Du-

rand, S., Leftheriotis, G., and Abraham, P. (2011).

Distance between laser head and skin does not inﬂu-

ence skin blood ﬂow values recorded by laser speckle

imaging. Microvasc. Res., 82:439–442.

Mahe, G., Humeau-Heurtier, A., Durand, S., Leftheriotis,

G., and Abraham, P. (2012b). Assessment of skin

microvascular function and dysfunction with laser

speckle contrast imaging. Circulation: Cardiovasc.

Imaging, 5:155–163.

Mahe, G., Humeau-Heurtier, A., Durand, S., Leftheriotis,

G., and Abraham, P. (2012c). Impact of experimen-

tal conditions on non-contact laser recordings in mi-

crovascular studies. Microcirc., 19:669–675.

Meier, T. O., Guggenheim, M., Vetter, S. T., Husmann, M.,

Haile, S. R., and Amann-Vesti, B. R. (2011). Mi-

crovascular regeneration in meshed skin transplants

after severe burns. Burns, 37:1010–1014.

Miao, P., Rege, A., Li, N., Thakor, N. V., and Tong, S.

(2010). High resolution cerebral blood ﬂow imag-

ing by registered laser speckle contrast analysis. IEEE

Trans. Biomed. Eng., 57:1152–1157.

Rege, A., Senarathna, J., Li, N., and Thakor, N. V.

(2012). Anisotropic processing of laser speckle im-

ages improves spatiotemporal resolution. IEEE Trans.

Biomed. Eng., 59:1272–1280.

Shannon, H. and Rizzolo, D. (2012). Carpal tunnel syn-

drome: symptoms, diagnosis, and treatment options.

JAAPA, 25:22–26.

Soderstrom, T., Svensson, H., Koop, T., and Moller, K. O.

(1999). Processing of laser-doppler signals from free

ﬂaps. Technol. Health Care, 7:219–223.

Wigley, F. M. (2009). Vascular disease in scleroderma. Clin.

Rev. Allergy Immunol., 36:150–175.

BIOIMAGING2014-InternationalConferenceonBioimaging