Relationship between the blood perfusion values determined by laser speckle imaging and laser Doppler imaging in normal skin and port wine stains
Accepted Manuscript Title: Relationship between the blood perfusion values determined by laser speckle imaging and laser Doppler imaging in normal ski...
Accepted Manuscript Title: Relationship between the blood perfusion values determined by laser speckle imaging and laser Doppler imaging in normal skin and port wine stains Author: Defu Chen Jie Ren Ying Wang Hongyou Zhao Buhong Li Ying Gu PII: DOI: Reference:
Please cite this article as: Chen Defu, Ren Jie, Wang Ying, Zhao Hongyou, Li Buhong, Gu Ying.Relationship between the blood perfusion values determined by laser speckle imaging and laser Doppler imaging in normal skin and port wine stains.Photodiagnosis and Photodynamic Therapy http://dx.doi.org/10.1016/j.pdpdt.2015.11.006 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Relationship between the blood perfusion values determined by laser speckle imaging and laser Doppler imaging in normal skin and port wine stains
Defu Chena, Jie Renb, Ying Wangb, Hongyou Zhaob, Buhong Lic and Ying Gub,*
a
School of Information and Electronics, Beijing Institute of Technology, Beijing 100081, China
b
Department of Laser Medicine, Chinese People’s Liberation Army General Hospital, Beijing 100853,
China c
Key Laboratory of OptoElectronic Science and Technology for Medicine of Ministry of Education,
Fujian Provincial Key Laboratory for Photonics Technology, Fujian Normal University, Fujian 350007, China
*
Corresponding author.
Ying Gu, M.D. Department of Laser Medicine, Chinese PLA General Hospital, Beijing 100853, P. R. China Tel: +86-010-6693-9394; Fax: +86-010-6822-2584 E-mail addresses: [email protected] (Ying Gu)
1
Highlights
Comparison of simultaneous LDI and LSI outputs during PORH for the first time.
Relationship between simultaneous LDI and LSI outputs during PORH was nonlinear.
Relationship between LDI and LSI outputs in port wine stains was linear.
Perfusion range should be considered when comparing LDI and LSI.
The measured skin characteristics should be considered when comparing LDI and LSI.
2
ABSTRACT
Objective: Laser Doppler imaging (LDI) and laser speckle imaging (LSI) are two major optical techniques aiming at non-invasively imaging the skin blood perfusion. However, the relationship between perfusion values determined by LDI and LSI has not been fully explored.
Methods: 8 healthy volunteers and 13 PWS patients were recruited. The perfusions in normal skin on the forearm of 8 healthy volunteers were simultaneously measured by both LDI and LSI during postocclusive reactive hyperemia (PORH). Furthermore, the perfusions of port wine stains (PWS) lesions and contralateral normal skin of 10 PWS patients were also determined. In addition, the perfusions for PWS lesions from 3 PWS patients were successively monitored at 0, 10 and 20 mins during vasculartargeted photodynamic therapy (V-PDT). The average perfusion values determined by LSI were compared with those of LDI for each subject.
Results: In the normal skin during PORH, power function provided better fits of perfusion values than linear function: powers for individual subjects go from 1.312 to 1.942 (R2=0.8967-0.9951). There was a linear relationship between perfusion values determined by LDI and LSI in PWS and contralateral normal skin (R2=0.7308-0.9623), and in PWS during V-PDT (R2=0.8037-0.9968).
Conclusion: The perfusion values determined by LDI and LSI correlate closely in normal skin and PWS over a broad range of skin perfusion. However, it still suggests that perfusion range and characteristics of the measured skin should be carefully considered if LDI and LSI measures are compared.
The human cutaneous microcirculation has been considered as a crucial surrogate marker of systemic microvascular function in a variety of injuries and diseases [1-3]. Over recent years, several laser-based techniques have been developed for assessing the microcirculatory blood perfusion [1, 2, 4, 5]. Among these techniques, laser Doppler and laser speckle contrast techniques play a preponderant role in the clinical and experimental studies [3]. Both these two techniques can be used to obtain the microcirculatory blood perfusion values, but the analytical methods applied to the two techniques are different [6, 7].
Laser Doppler technique analyze beat frequencies that are related to the mixing of Doppler-shift light scattered from the moving red blood cells, with the non-shift light scattered from the stationary tissue components [8, 9]. Laser Doppler flowmetry (LDF) is the initially developed laser Doppler technique and has been used for nearly 40 years. Single-point LDF provides good temporal resolution but poor reproducibility due to the regional heterogeneity of skin perfusion (especially in low capillary density skin area) [10]. This drawback was addressed by laser Doppler imaging (LDI). The conventional LDI obtain a two-dimensional image of tissue perfusion by rastering of a laser beam horizontally and vertically across the tissue surface, which may take several minutes [11]. In order to reduce the acquisition time, a new generation of commercial LDI technique called laser Doppler line scanner (LDLS), which raster a line of pixels in parallel at a single time step, was developed [8, 12]. The LDLS technique allows for a series of perfusion measurements that performed in parallel [2]. The improvement significantly reduces the acquisition time and thus it brings LDI technique one step closer to real time
5
imaging [2]. Recently, a high-speed LDI using an integrating CMOS sensor for full-field scanning has been also developed, but the instrument can only measure the area of about 50 cm2 in size [13].
Laser speckle contrast imaging (LSCI) is based on the spatial and/or temporal statistics of the speckle pattern [14, 15]. The traditional LSCI is a non-contact, full-field, real-time technique by calculating the spatial speckle contrast, which is defined as the ratio of the standard deviation of the intensity to the mean intensity over a square of 5 × 5 or 7 × 7 pixels window sliding in one image [16]. Consequently, LSCI has the disadvantage of a lack of spatial resolution. In order to overcome this limitation, a new form of LSCI called laser speckle imaging (LSI) based on the speckle temporal contrast, which is determined in 1 pixel over a sequence of images, has been developed [17-19]. As compared with the conventional LSCI, LSI maintains the full spatial resolution of CCD camera but with a lower temporal resolution [19].
Although laser Doppler and laser speckle contrast techniques were widely used for microvascular blood perfusion measurement, the relationship between perfusion values determined by the two techniques has not been fully explored [3, 17, 20-26]. Recently, more and more studies have been focused on elucidating the relationship [17, 20-26]. Since the laser Doppler and laser speckle techniques cannot provide perfusion values in absolute units, most recent studies tried to analyze the relationship between perfusion values determined by single-point/integrating-probe LDF and LSCI coupled with specific reactivity tests [3, 20-23]. Among the tests, the increase in blood perfusion of normal skin on the forearm above baseline levels following release from brief arterial occlusion, commonly known as post-occlusive reactive hyperemia (PORH), is one of the most commonly used [3, 20-22]. The perfusion values are rapidly changing after the release of occlusion during PORH. The obtained results showed that the relationship
6
between perfusion values determined by single-point/integrating-probe LDF and LSCI is non-linear in normal skin on the forearm during PORH. Nevertheless, the works only compared data from singlepoint/integrating-probe LDF with full-field LSCI. Comparing data from LDI with those of LSCI would have made more sense, as LDI are ‘regional’ and not single-point as LDF [27]. However, the acquisition of an image for the conventional LDI requires a few minutes when they depend on the scanning of the tissues, which make PORH difficult to measure with the LDI. As a result, the relationship between perfusion values determined by LDI and LSCI in normal skin during PORH has never been investigated.
Port wine stains (PWS) birthmarks are congenital and progressive vascular malformations histologically characterized by ecstatic capillaries predominantly in the upper dermis of human skin [28]. The incidence of PWS is estimated to be approximately 0.3-0.5 %. Vascular-targeted photodynamic therapy (V-PDT) is widely considered to be an effective method for the treatment of PWS [28, 29]. The perfusion in PWS can be assessed objectively and noninvasively by LSI [29-31] and LDI [32-34], and significant changes in PWS perfusion was observed during V-PDT [31, 32]. However, the perfusion values of PWS as assessed by LDI and LSI have never been directly compared.
The aim of this study was to investigate the relationship between perfusion values determined by LDI and LSI in normal skin on the forearm during PORH and in PWS and contralateral normal skin, respectively. Firstly, we compared the relationship of perfusion values with high dynamic changes simultaneously assessed by LDI and LSI in normal skin on the forearm during PORH. Furthermore, the perfusions of port wine stains (PWS) lesions and contralateral normal skin of PWS patients were also determined. In addition, the perfusions for PWS lesions from 3 PWS patients were successively
7
monitored at 0, 10 and 20 mins during V-PDT.
2. Participants and methods
2.1 Participants
Eight healthy non-smoking Chinese volunteers (aged form 24 to 29 years old) and thirteen Chinese patients (aged from 10 to 28 years old) clinically diagnosed with PWS were recruited in the study. This study was approved by the Ethics Committee of the Chinese PLA General Hospital. Informed consents were obtained from all the participants.
2.2 Laser speckle imaging
A LSI system was provided by the Wuhan National Laboratory for Optoelectronics (NLO) at Huazhong University of Science and Technology, which has been described in detail elsewhere [29]. Briefly, the LSI system mainly consists of a wheeled stage, a connection arm and a speckle detection head. The connection arm allows the speckle detection head to flexibly move horizontally and vertically within a certain range. The scan distance between the scanner head and the skin being measured was about 25 cm. The speckle detection head contains a continuous-wave laser source with a wavelength of 660 nm (50 mW) and a 12 bit charge-couple device (CCD) camera (Baumer TXG03, Germany; pixel size = 7.4 × 7.4 μm2). After laser light is scattered onto the skin, speckle signal was acquired by the CCD camera through an optical imaging system. A 660 ± 10 nm band-pass filter was placed in front of the CCD
8
camera to eliminate the extraneous light. The CCD exposure time in this study was 20 ms. The sampling frequency was 5 Hz. The raw data was processed by the computer subsequently based on the temporal speckle contrast analysis [18]. Finally, LSI parameter configurations and data analyses were performed by graphical software named “Real-time Laser Blood Flow Imaging” (RTLBI), which was developed by the NLO.
2.3 Laser Doppler imaging
The commercially available line scanning LDI system (moorLDLS2-IR, Moor instrument, Cambridge, UK) was employed in the study. In this system, a divergent laser line with a wavelength of 785 ± 10 nm is directed via a scanning mirror to sweep across skin or other tissue surfaces. The Doppler shifted light from moving blood cell and the non-shifted light from stationary tissue is directed by the same scanning mirror and other optics onto a 64 element linear array to obtain the perfusion measurement in parallel, and then the data was processed to build a 2-dimensional color coded blood perfusion image. The processing frequency bandwidth was from 30 Hz to 15 kHz. The scan distance between the scanner head and the skin being measured was fixed at 15 cm, as recommended in the manufacturer’s manual. The measurement can be performed in fast line scanning mode or in single image scanning mode. In the fast line scanning mode, the laser line remains in the same position and blood perfusion changes along the line are continuously recorded by the linear array detector with a sampling frequency of 40 Hz. The laser line scanning mode can be used to record dynamic changes in perfusion during PORH. In the single image scanning mode, the system scans a laser line over skin and the desired image pattern is built up line by line. The image size was 15 × 12 cm2 with the resolution of 256 × 256 points. In this study, the
9
slow scan speed was chosen in order to achieve the best image quality for PWS perfusion measurements, and the total scan time was estimated to be 20 s. The field of view of the LDI system is similar to that of the LSI system.
2.4 Measurement protocols
Measurements were performed in a temperature-controlled room (24 ± 1 oC) without any air movements. All the subjects had an acclimatization period of more than 20 min before measurement. In order to compare the perfusion values that measured by both LDI and LSI, two recording procedures were performed. Firstly, the perfusion measurements were performed in normal skin on the volar surface of the forearm of the 8 healthy volunteers during PORH. The acquisitions with the LDI and LSI systems were started simultaneously. In order to record dynamic changes in perfusion during the PORH, the LDI adopted the fast line scanning mode. The PORH protocol include 2 min baseline period, 3 min of ischemia period, and then followed by 3 min of reperfusion. Ischemic conditions were obtained by inflating the cuff at suprasystolic pressure (220 mmHg) [21]. Secondly, the perfusion measurements were performed in PWS and contralateral normal skin of 10 PWS patients resting in a comfortable supine position. The PWS and contralateral normal skin were imaged successively with the LDI and LSI systems. LDI adopted the single image scanning mode for PWS perfusion measurements.
Finally, 3 PWS patients were treated with V-PDT. The PWS patients were intravenously injected with a domestically produced photosensitizer HiPorfin (Chongqing Huading Modern Biopharmaceutics Co., Ltd. Chongqing, China) with a dosage of 2-3 mg/kg of body weight. Immediately after HiPorfin injection,
10
the PWS area were irradiated with the 532 nm laser light source at a power density of 100 mW/cm 2, and the irradiation time was 20 mins. During V-PDT, the PWS perfusion were imaged successively with the LDI and LSI at 0, 10, and 20 mins.
2.5 Data processing and analysis
The LDI and LSI data were digitized and stored in a computer, and analyzed offline with signal processing software (moorLDLS laser Doppler line scanner research version 2.2 for LDLS measurements, Moor Instruments Ltd.; RTLBI (Real-time Laser Blood Flow Imaging, Wuhan National Laboratory for Optoelectronics, Wuhan, China) software for LSI measurements). All the measured data were processed and graphed using OriginPro 9.1 software (OriginLab Corporation, MA USA). APULDI and APULSI stand for average perfusion units that determined by LDI and LSI, respectively. Region of interest (ROI) was defined as the region in which an average of all the pixel perfusion values was performed, and time of interest (TOI) was defined as the duration over which the perfusion data are averaged [20, 23, 35]. In the normal skin on the forearm during PORH, in order to directly compare perfusion value of LDI with that of LSI at the same time point, 6 rectangular ROIs (~100 mm 2) were chosen on the LSI data of each subject. The average of all the pixel perfusion values in the 6 LSI ROIs was computed to obtain a time-evolution LSI perfusion signal for each subject, while the average of all the pixel perfusion values along the divergent laser line for the LDI data was computed to generate a time-evolution LDI perfusion signal. As shown in Fig. 1 (B), 3 TOIs of 10 s, 3 TOIs of 10 s and 5 TOIs of 10 s were chosen independently for the rest, vascular occlusion and post-occlusive hyperemia period for the obtained perfusion data of LDI and LSI on the same time intervals, respectively. The average
11
values were calculated for each TOI. In the PWS and contralateral normal skin, 10 rectangular ROIs of 100 ± 20 mm2 (the maximum number of ROIs can be selected in one image depending on the RTLBI software) were chosen on the flat region of the PWS and contralateral normal skin. Furthermore, the 10 ROIs should cover the broad range of perfusions, which including the low, medium, high perfusion on the flat area. Finally, the average perfusion values of LDI were compared with those of LSI for each subject.
3. Results
3.1 Perfusion values of normal skin simultaneously assessed by LDI and LSI during PORH
Fig. 1 shows an example of LDI and LSI perfusion data in normal skin during PORH assessed simultaneously by LDI and LSI, for a typical subject. As shown in Fig. 1 (A) and (B), the first 2 min and successive 3 min regions correspond to the rest (no cuff) and to the biological zero-flow baselines (vascular occlusion), respectively. The transient increase at the time-point of 5 min, visible for both LDI and LSI perfusion data, correspond to the post-occlusive reactive hyperemic reaction occurring after release of the cuff. In order to directly compare LDI and LSI perfusion values, the normalized perfusion values were also shown in Fig. 1 (C). The changes and trends of perfusion values for LDI and LSI during PORH are consistent. Finally, based on the calculation of average perfusion value of TOIs (as shown in Fig. 1 (B)) for LDI and LSI data, the LSI perfusion data plotted against the LDI perfusion data was showed in Fig. 1 (D).
12
In Fig. 2, the LSI perfusion data plotted against LDI perfusion data for all the 8 subjects were also shown. The relationship between perfusion values determined by LDI and LSI was evaluated by fitting power function and linear function to the data for each subject individually. The residual sum of squares was smaller for power fitting than linear fitting for each subject. This indicates that the power function fitted the data of perfusion values better than the linear function: the powers for individual subjects increase from 1.312 to 1.942, and R2 varies from 0.8967 to 0.9951. Furthermore, the line of best fit for linear fitting crossed the ordinate at negative values for all the 8 subjects. Based on the above analysis, in normal skin on the forearm during PORH, the relationship between perfusion values determined by LDI and LSI was non-linear.
3.2 Perfusion values of port wine stains and contralateral normal skin directly assessed by LDI and LSI
Fig. 3 (A), (B) and (C) show LDI CCD camera photograph, LDI perfusion and LSI perfusion images of a typical PWS in the face, respectively. In order to fully and accurately evaluate the perfusion of PWS and contralateral normal skin, the perfusion images were acquired from three different angles. The two imaging systems incorporate a similar color scale to depict varying degrees of perfusion (i.e. blue: low perfusion, red: high perfusion). It is obvious that the overall distribution of perfusion determined by LDI
13
is well correspondent with LSI. However, there still have some minor differences in relative distribution of perfusion intensity. The result also illustrated that the perfusion values of the PWS were larger than the contralateral normal skin for the subject. In order to directly compare the LDI perfusion image with the LSI perfusion image, 10 rectangular ROIs of 100 ± 20 mm2 were chosen on the flat region of the PWS and contralateral normal skin for both LDI and LSI perfusion image. Fig. 3 (D) shows the relationship between the average perfusion values of the ROIs determined by the LDI and LSI.
Fig. 4 compares perfusion values determined by LSI and LDI obtained from scans of the 10 PWS patients. The perfusion value ranges in PWS and contralateral normal skin were 161.9-887.2 APULDI for LDI and 3.0-15.9 APULSI for LSI, which were significant larger than 11.3-268.7 APULDI for LDI and 0.4-6.2 APULSI for LSI in normal skin on the forearm during PORH. It can be observed that the scattering of the data for PWS and contralateral normal skin was also larger than that of the data for normal skin on the forearm during PORH. The relationship between perfusion values determined by LDI and LSI was also evaluated by fitting power function and linear function to the data for each subject individually. Power function cannot provide better fits than linear function for seven of the ten PWS patients, and the powers for fitting power function to data were close to 1. Non-linear relationship between the perfusion values could not be observed. There was linear relationship between the perfusion values determined by LDI and LSI in PWS and contralateral normal skin, and R2 value varies from 0.7308 to 0.9623.
14
3.3 Perfusion values of port wine stains directly assessed by LDI and LSI during V-PDT
Fig. 5 (A), (B) and (C) show LDI CCD camera photograph, LDI perfusion and LSI perfusion images of a typical PWS at 0, 10, and 20 mins during V-PDT, respectively. The overall level of perfusion was significantly increasing at 10 mins and then slowly decreasing at 20 mins. It is clear that the overall distribution of perfusion determined by LDI is well correspondent with LSI at various time points. The average perfusion values of the entire treatment area were plotted versus time to show the trend in perfusion of PWS during V-PDT, as shown in Fig. 5 (D). The average perfusion values measured by LDI and LSI first increased, and then decreased for all the 3 subjects. However, differences were found for the magnitude of increase (or decrease) in perfusion values measured by LDI and LSI.
Fig. 6 further compares the average perfusion values determined by LSI and LDI at 0 mins, 10 mins, and 20 mins in PWS during V-PDT for each of 3 PWS patients. There was linear relationship between the perfusion values determined by LDI and LSI, and R2 values are 0.8037, 0.9629 and 0.9968, respectively.
15
4. Discussion
In the present work, we have studied the relationship of perfusion values determined by LDI and LSI over a broad range of skin perfusion. In the normal skin during PORH, power function provided better fits of perfusion values than linear function. There was a linear relationship between the perfusion values determined by LDI and LSI in PWS and contralateral normal skin of 10 PWS patients, and in PWS of 3 PWS during V-PDT.
Since the traditional laser scanning-based LDI can’t be used for dynamic monitoring of high-frequency blood perfusion fluctuations during PORH, the relationship between LDI and LSCI signals combined with the PORH has never been investigated. In order to accurately record kinetics of the PORH, laser line scanning mode of LDI was employed in our study. In the fast line scanning mode, the laser line remains in the same position and blood perfusion changes along the line are continuously recorded with a sampling frequency of 40 Hz. Several previous studies showed that the LDF with the sampling frequency of 20 Hz can be fast enough to record the blood perfusion changes during the PORH [20, 21]. Therefore, fast line scanning mode of LDI employed in our study must be fast enough to record the very dynamic changes of skin perfusion during PORH. Furthermore, the spatial variability of skin perfusion can be also minimized by averaging the perfusion values from the laser line for LDI and averaging perfusion values from ROIs for LSI, respectively. In the study, the relationship of perfusion values determined by the LDI and LSI was determined to be non-linear in normal skin on the forearm during PORH. Our results are in consistent with those of others who study the relationship between single-point or integrating-probe LDF values and full-field LSCI [20-22].
16
We further compare the perfusion values imaged by LDI and LSI in PWS and contralateral normal skin on the face. PWS are congenital vascular malformations histologically characterized by ecstatic capillaries predominantly located at varying depths within the mid to upper dermis of human skin [3638]. The enlarged blood volume may lead to higher blood perfusion in PWS compared to normal skin. Therefore, heterogeneous perfusion maps with a wide range of skin perfusion were observed. Additionally, an initial increase and then a decrease in perfusion in PWS during V-PDT were observed by LSI and LDI, which is consistent with the previous study by using LSI [31] and LDI [32], respectively. The overall distribution of perfusion determined by LDI is well correspondent with LSI before and during V-PDT. However, there still have some minor differences in relative distribution of perfusion intensity. The differences may be attribute to different measurement depth for LDI and LSI. Firstly, measurement depth depends on the wavelength of the laser light (780 nm for LDI versus 660 nm for LSI), and on the type of illumination being used (laser line for LDI versus large spot for LSI). In fact, according to the previous studies, the skin measurement depth has been determined as approximately 0.5-1 mm for laser Doppler measurement whereas it is about 300 µm for laser speckle systems when laser wavelengths close to 780 nm are used in normal skin [8, 39]. Secondly, measurement depth also depends on tissue properties such as the structure and density of the capillary beds, pigmentation, oxygenation, and so on [40]. Since the PWS skin has an abnormal density of blood vessels, the optical properties will be significantly different from the contralateral normal skin [38]. Differences were observed in the LDI and LSI techniques response to tissue optical properties [41]. As a result, the differences in relative distribution of perfusion intensity can be observed in the PWS and contralateral normal skin measurements. Some previous works also directly compared the perfusion values imaged by LDI and LSCI in optic nerve head [24], in burn scar [17] and in normal human skin [25, 26]. A good linear correlation was found between
17
the two techniques in burn scar perfusion assessment [17]. Millet et al. have also shown that skin blood perfusion measured with LSCI is linearly related to the LDI signal in normal skin over a wide range of perfusion [26]. Consistent with the previous studies, the relationship between perfusion values determined by LDI and LSI in PWS and contralateral normal was also linear.
The non-linear relationship between perfusion values determined by LDI and LSI in normal skin on the forearm during PORH may be attributed in part to that the speckle contrast technique is not linear with respect to either velocity (or perfusion) compared to the laser Doppler techniques [22, 42]. The speckle contrast plotted against the logarithm of the ratio of the correlation time (τc) to the integration time (T) gives an S-shaped curve [22, 42]. (τc is inversely proportional to some characteristic velocity, and T is a constant for a single exposure.) However, the central portion of the S-shaped curve is approximately linear over quite a large range [22]. This means that the speckle contrast is approximately linear with respect to relative changes in velocity within a certain range of perfusion values. It should be pointed out that the change in exposure time of the CCD camera will change the linear velocity response range [43, 44]. The previous study have investigated the influence of exposure time (10, 15 and 20 ms) on the linear velocity response range for the LSI used in this study with a moving white plate model [43]. The result showed that the LSI system could not distinguish the white plate speed of below 0.42 mm/s when using exposure times of 10 and 15 ms [43], which implies that the exposure time should not be less than 20 ms for the LSI system once low skin perfusion was measured [43]. Therefore, in the present study, the exposure time was set to 20 ms, which was also consistent with the exposure time used in our previous study for PWS lesions assessment [29]. However, in the lower perfusion range, the relationship between speckle contrast and velocity (or perfusion) is still non-linear. A recent theoretical comparison also
18
showed that the linearity of perfusion values from LDF is very well preserved in the full perfusion values range from 0 to 2 mm/s, whereas the perfusion values from LSCI have a non-linear relationship with velocity, especially in the low perfusion values range from 0 to 0.2 mm/s [45]. The average capillary blood cell velocities at the dorsal aspect of the proximal index finger were determined to be 0.47±0.37 mm/s during the resting state and 0.90±0.46 mm/s at peak blood cell velocity during PORH, respectively [46]. The perfusion values can be further decreased to biological zero by vascular occlusion. Therefore, the perfusion value range of normal skin on the forearm during PORH may include the low perfusion value range, in which the relationship between laser speckle and velocity (or perfusion) is non-linear. However, in practical applications, the non-linear relationship between the perfusion values determined by LDI and LSI may be weakened, which was due to the Brownian motion [45], the occasional opening and closing of local shunts, and so on. As a result, the non-linear relationship can be observed only under carefully controlled conditions (i.e. PORH). In our study, the perfusion values in PWS and contralateral normal skin were 161.9-887.2 APULDI for LDI and 3.0-15.9 APULSI for LSI, which were generally larger than 11.3-268.7 APULDI for LDI and 0.4-6.2 APULSI for LSI in normal skin on the forearm during PORH. The perfusion in PWS and contralateral normal skin were over a wide range and may beyond the low perfusion value range of non-linear relationship, and thus non-linear relationship between the perfusion values could not be observed.
A large scattering data were observed in both Fig. 2 and Fig. 4. In particular, the scattering of the data for PWS and contralateral normal skin measurements is larger than that of the data for normal skin on the forearm during PORH. Firstly, this is at least in part probably due to different measured skin. Compared to the normal skin, the PWS skin has a higher perfusion heterogeneity due to the abnormal
19
density of blood vessels in the upper dermis of PWS skin [32]. Moreover, there is large heterogeneity of perfusion values between PWS patients. Secondly, large scattering of the data may be also due to the measurement depth. Therefore, the two instruments do not measure perfusion from the exact same skin layers, which may also explain part of the higher degree of perfusion variability observed especially in the heterogeneous PWS and contralateral normal skin measurements. Finally, the different measurement and data processing procedures may also have an influence on the scattering of the data. In normal skin during PORH, the continuously dynamic perfusion values were recorded in course of time. Some perfusion values were extremely low due to the vascular occlusion. However, in the PWS perfusion measurements, the discrete and relatively high perfusion values were observed due to the heterogeneous nature of the PWS skin. Furthermore, the choice of ROIs on LSI figures and the corresponding ROIs for the LDI data could not be completely matched in the same area for comparison of the perfusion values measured on the PWS skin. The slightly difference exist between the ROIs of LDI and LSI may lead to the scattering of the data. The choices of ROIs may slightly affect the relationship curve between the perfusion values determined by LDI and LSI.
A limitation of the study is that the average perfusion values derived from a single detection laser line of LDI were compared with the average perfusion values derived from the ROIs of LSI. Indeed, the acquisition of a perfusion image for the conventional LDI requires a few minutes, which make PORH difficult to measure with the LDI. Furthermore, it also not possible to place the detection heads of LDI and LSI juxtaposed measuring perfusion over the same skin while keeping the laser beams perpendicular to the skin. As a result, previous studies only compared data from single-point or integrating-probe (multi-points) LDF with full-field LSCI by far [3, 20-23]. In this regard, it may be also acceptable to
20
compare the average perfusion values derived from a single detection laser line of LDI with full-field LSI in normal skin during PORH in this study. Compared to the single-point or multi-points perfusion values from LDF, the variability of perfusion measurements can be further reduced by averaging the perfusion values along the detection laser line of LDI on the skin during PORH. However, the full-field laser Doppler perfusion imaging [13] is an emerging and promising technique that may be better compared with LSI than LDI or LDF, due to the full-field and near real-time perfusion imaging. Another limitation of this study is that the two different laser wavelengths (780 nm for LDI versus 660 nm for LSI) were applied in LDI and LSI, respectively. Therefore, the penetration depth of laser light may be different. However, the important parameter is measurement depth rather than the penetration depth. Even if the same wavelength was used, LDI and LSI could not measure exactly the same skin layers due to the different types of illumination and signal processing methods being used. Fortunately, the advantage of using two different wavelengths for LDI and LSI was that the laser of one system does not disturb the second system for the simultaneous acquisitions during PORH.
5. Conclusion
In conclusion, we have studied the relationship of skin perfusion values determined by LDI and LSI techniques over a broad range of skin perfusion. In normal skin on the forearm during PORH, the relationship between perfusion values determined by LDI and LSI was non-linear. Whereas, there was a linear relationship between the perfusion values determined by LDI and LSI in PWS and contralateral normal skin, and in PWS during V-PDT. Although LDI and LSI correlate closely, we suggest that perfusion range and characteristics of the measured skin should be carefully considered if LDI and LSI
21
measurements are compared.
Conflict of interest statement
The authors declare no conflict of interest in any form with respect to this article.
Acknowledgements
This study was supported by the National Natural Science Foundation of China (No. 61450005; 61036014; 61108078), the Industry Research Special Funds for Public Welfare Projects (No. 2015SQ00057) and the Fujian Provincial Natural Science Foundation (2014J07008). The authors would like to thank Prof. Pengcheng Li from National Laboratory for Optoelectronics at Huazhong University of Science and Technology for providing the laser speckle imaging system.
22
References
[1] Roustit M, Cracowski JL. Non-invasive assessment of skin microvascular function in humans: an insight into methods. Microcirculation. 2012;19:47-64. [2] Leahy MJ, Enfield JG, Clancy NT, O’Doherty J, McNamara P, Nilsson GE. Biophotonic methods in microcirculation imaging. Medical Laser Application. 2007;22:105-26. [3] Humeau-Heurtier A, Guerreschi E, Abraham P, Mahe G. Relevance of Laser Doppler and Laser Speckle Techniques for Assessing Vascular Function: State of the Art and Future Trends. IEEE Trans Biomed Eng. 2013;60:659-66. [4] Daly SM, Leahy MJ. ‘Go with the flow ’: A review of methods and advancements in blood flow imaging. J Biophotonics. 2013;6:217-55. [5] Allen J, Howell K. Microvascular imaging: techniques and opportunities for clinical physiological measurements. Physiol Meas. 2014;35:R91-R141. [6] Briers JD. Laser Doppler, speckle and related techniques for blood perfusion mapping and imaging. Physiol Meas. 2001;22:R35-R66. [7] Briers D, Duncan DD, Hirst E, Kirkpatrick SJ, Larsson M, Steenbergen W, et al. Laser speckle contrast imaging: theoretical and practical limitations. J Biomed Opt. 2013;18:066018-. [8] O'Doherty J, McNamara P, Clancy NT, Enfield JG, Leahy MJ. Comparison of instruments for investigation of microcirculatory blood flow and red blood cell concentration. J Biomed Opt. 2009;14:034025. [9] Leahy MJ, de Mul FF, Nilsson GE, Maniewski R. Principles and practice of the laser-Doppler perfusion technique. Technol Health Care. 1999;7:143-62.
23
[10] Murray AK, Herrick AL, King TA. Laser Doppler imaging: a developing technique for application in the rheumatic diseases. Rheumatology. 2004;43:1210-8. [11] Rajan V, Varghese B, van Leeuwen TG, Steenbergen W. Review of methodological developments in laser Doppler flowmetry. Lasers Med Sci. 2009;24:269-83. [12] Holland AJA, Ward D, La Hei ER, Harvey JG. Laser Doppler Line Scan Burn Imager (LDLS-BI): Sideways move or a step ahead? Burns. 2014;40:113-9. [13] Leutenegger M, Martin-Williams E, Harbi P, Thacher T, Raffoul W, Andre M, et al. Real-time full field laser Doppler imaging. Biomed Opt Express. 2011;2:1470-7. [14] Basak K, Manjunatha M, Dutta PK. Review of laser speckle-based analysis in medical imaging. Med Biol Eng Comput. 2012;50:547-58. [15] Draijer M, Hondebrink E, Leeuwen T, Steenbergen W. Review of laser speckle contrast techniques for visualizing tissue perfusion. Lasers Med Sci. 2009;24:639-51. [16] Senarathna J, Rege A, Li N, Thakor NV. Laser Speckle Contrast Imaging: theory, instrumentation and applications. IEEE Rev Biomed Eng. 2013;6:99-110. [17] Stewart CJ, Frank R, Forrester KR, Tulip J, Lindsay R, Bray RC. A comparison of two laser-based methods for determination of burn scar perfusion: Laser Doppler versus laser speckle imaging. Burns. 2005;31:744-52. [18] Li P, Ni S, Zhang L, Zeng S, Luo Q. Imaging cerebral blood flow through the intact rat skull with temporal laser speckle imaging. Opt Lett. 2006;31:1824-6. [19] Cheng HY, Luo QM, Zeng SQ, Chen SB, Cen J, Gong H. Modified laser speckle imaging method with improved spatial resolution. J Biomed Opt. 2003;8:559-64. [20] Humeau-Heurtier A, Mahe G, Durand S, Abraham P. Skin perfusion evaluation between laser
24
speckle contrast imaging and laser Doppler flowmetry. Opt Commun. 2013;291:482-7. [21] Binzoni T, Humeau-Heurtier A, Abraham P, Mahe G. Blood Perfusion Values of Laser Speckle Contrast Imaging and Laser Doppler Flowmetry: Is a Direct Comparison Possible? IEEE Trans Biomed Eng. 2013;60:1259-65. [22] Tew GA, Klonizakis M, Crank H, Briers JD, Hodges GJ. Comparison of laser speckle contrast imaging with laser Doppler for assessing microvascular function. Microvasc Res. 2011;82:326-32. [23] Puissant C, Abraham P, Durand S, Humeau-Heurtier A, Faure S, Leftheriotis G, et al. Reproducibility of non-invasive assessment of skin endothelial function using laser Doppler flowmetry and laser speckle contrast imaging. PLoS ONE. 2013;8:e61320-e. [24] Yaoeda K, Shirakashi M, Funaki S, Funaki H, Nakatsue T, Abe H. Measurement of microcirculation in the optic nerve head by laser speckle flowgraphy and scanning laser doppler flowmetry. Am J Ophthalmol. 2000;129:734-9. [25] Forrester KR, Stewart C, Tulip J, Leonard C, Bray RC. Comparison of laser speckle and laser Doppler perfusion imaging: measurement in human skin and rabbit articular tissue. Med Biol Eng Comput. 2002;40:687-97. [26] Millet C, Roustit M, Blaise S, Cracowski JL. Comparison between laser speckle contrast imaging and laser Doppler imaging to assess skin blood flow in humans. Microvasc Res. 2011;82:147-51. [27] Humeau-Heurtier A, Abraham P, Mahe G. Linguistic Analysis of Laser Speckle Contrast Images Recorded at Rest and During Biological Zero: Comparison With Laser Doppler Flowmetry Data. IEEE Trans Med Imaging. 2013;32:2311-21. [28] Qiu H, Gu Y, Wang Y, Huang N. Twenty Years of Clinical Experience with a New Modality of Vascular-Targeted Photodynamic Therapy for Port Wine Stains. Dermatol Surg. 2011;37:1603-10.
25
[29] Ren J, Li P, Zhao H, Chen D, Zhen J, Wang Y, et al. Assessment of tissue perfusion changes in port wine stains after vascular targeted photodynamic therapy: a short-term follow-up study. Lasers Med Sci. 2014;29:781-8. [30] Huang Y-C, Ringold TL, Nelson JS, Choi B. Noninvasive blood flow imaging for real-time feedback during laser therapy of port wine stain birthmarks. Lasers Surg Med. 2008;40:167-73. [31] Qiu H, Zhou Y, Gu Y, Ang Q, Zhao S, Wang Y, et al. Monitoring microcirculation changes in port wine stains during vascular targeted photodynamic therapy by laser speckle imaging. Photochem Photobiol. 2012;88:978-84. [32] Chen D, Ren J, Wang Y, Gu Y. Laser Doppler line scanner for monitoring skin perfusion changes of port wine stains during vascular-targeted photodynamic therapy. Proc SPIE. 2014; 9268: 92681Y. [33] Mermod T, El Ezzi O, Raffoul W, Erba P, de Buys Roessingh A. Assessment of the role of LASERDoppler in the treatment of port-wine stains in infants. J Pediatr Surg. 2015;50:1388-92. [34] Troilius AM, Ljunggren B. Evaluation of port wine stains by laser Doppler perfusion imaging and reflectance photometry before and after pulsed dye laser treatment. Acta Derm Venereol. 1996;76:2914. [35] Rousseau P, Mahe G, Haj-Yassin F, Durand S, Humeau A, Leftheriotis G, et al. Increasing the "region of interest" and "time of interest", both reduce the variability of blood flow measurements using laser speckle contrast imaging. Microvasc Res. 2011;82:88-91. [36] Ortiz AE, Nelson JS. Port-wine stain laser treatments and novel approaches. Facial Plast Surg. 2012;28:611-20. [37] Sharif SA, Taydas E, Mazhar A, Rahimian R, Kelly KM, Choi B, et al. Noninvasive clinical assessment of port-wine stain birthmarks using current and future optical imaging technology: a review.
26
Br J Dermatol. 2012;167:1215-23. [38] Rajan V, Varghese B, Van Leeuwen TG, Steenbergen W. Influence of tissue optical properties on laser Doppler perfusion imaging, accounting for photon penetration depth and the laser speckle phenomenon. J Biomed Opt. 2008;13:024001. [39] Fredriksson I, Larsson M, Stromberg T. Measurement depth and volume in laser Doppler flowmetry. Microvasc Res. 2009;78:4-13. [40] Greco A, Ragucci M, Liuzzi R, Gargiulo S, Gramanzini M, Coda A, et al. Repeatability, Reproducibility and Standardisation of a Laser Doppler Imaging Technique for the Evaluation of Normal Mouse Hindlimb Perfusion. Sensors. 2012;13:500-15. [41] Forrester KR, Tulip J, Leonard C, Stewart C, Bray RC. A laser speckle imaging technique for measuring tissue perfusion. IEEE Trans Biomed Eng. 2004;51:2074-84. [42] Fercher AF, Briers JD. Flow visualization by means of single-exposure speckle photography. Opt Commun. 1981;37:326-30. [43] Zhang H. Research on laser speckle blood flow imaging system for clinical application [PhD thesis]. Wuhan: Huazhong University of Science & Technology; 2012. [44] Qiu JJ, Li PC, Ul'yanov SS, Zeng SQ, Luo QM. Enlarging the linear response range of velocity with optimum imaging parameters and modified data processing in laser speckle imaging. Proc SPIE. 2008; 6863:68630T. [45] Bi R, Dong J, Poh CL, Lee K. Optical methods for blood perfusion measurement-theoretical comparison among four different modalities. J Opt Soc Am A. 2015;32:860-6. [46] Stücker M, Baier V, Reuther T, Hoffmann K, Kellam K, Altmeyer P. Capillary Blood Cell Velocity in Human Skin Capillaries Located Perpendicularly to the Skin Surface: Measured by a New Laser
Figure 1 Typical normal skin perfusion (male, 28 y, non-smoker) measured by (A) LDI and (B) LSI as a function of time during a rest (2 min), vascular occlusion (3 min) and reactive hyperemia (3 min) sequence, for a typical subject. (C) Normalized skin perfusion values determined by LDI (black line) and LSI (red line), respectively. (D) The LSI perfusion data plotted against LDI perfusion data for the subject.
29
Figure 2 The LSI perfusion values plotted against LDI perfusion values for all the 8 subjects during PORH. The solid red line represents the fitting curve of power function. Each symbol represents one
subject. (Powers for each individual are: 1.312, 1.345, 1.437, 1.553, 1.605, 1.626, 1.753 and 1.942)
30
Figure 3 Typical perfusion images of PWS and contralateral normal skin as assessed by LDI and LSI: (A) LDI CCD camera photograph; (B) LDI perfusion image; (C) LSI perfusion image; (D) The LSI perfusion data plotted against LDI perfusion data. (The white rectangles marked in the photograph represent ROIs)
31
Figure 4 The LSI perfusion values plotted against LDI perfusion values for the 10 PWS patients. The solid red line represents the fitting curve. Each symbol represents one subject.
32
Figure 5 Typical perfusion images of PWS as assessed by LDI and LSI at 0 mins, 10 mins, and 20 mins during V-PDT: (A) LDI CCD camera photograph; (B) LDI perfusion image; (C) LSI perfusion image; (D) The LDI and LSI perfusion data at various time points for the studied 3 subjects.
33
Figure 6 The LSI perfusion values plotted against LDI perfusion values in PWS for the 3 PWS patients during V-PDT. The solid red line represents the fitting curve and each symbol represents one subject.
