LIU Yan, YANG Xue, ZHAO Jing, LI Gang, LIN Ling*

1. State Key Laboratory of Precision Measuring Technology and Instruments, Tianjin University, Tianjin 300072, China

2. Tianjin Key Laboratory of Biomedical Detecting Techniques & Instruments, Tianjin University, Tianjin 300072, China

3. School of Chinese Medicine Engineering, Tianjin University of Traditional Chinese Medicine, Tianjin 300193, China

Study on Internal Information of the Two-Layered Tissue by Optimizing the Detection Position

LIU Yan1,2, YANG Xue1,2, ZHAO Jing3, LI Gang1,2, LIN Ling1,2*

1. State Key Laboratory of Precision Measuring Technology and Instruments, Tianjin University, Tianjin 300072, China

2. Tianjin Key Laboratory of Biomedical Detecting Techniques & Instruments, Tianjin University, Tianjin 300072, China

3. School of Chinese Medicine Engineering, Tianjin University of Traditional Chinese Medicine, Tianjin 300193, China

As to most methods of detecting the inner information of inhomogenous tissue, a significant issue is that the detection position is ambiguous because of the complexity of human tissue structure and discrepancies among individuals. This paper studies the best source-detector distance (SDSbest) to detect internal information of a fat-muscle tissue with spatially resolved diffuse reflectance spectra. In order to weaken the measurement error caused by the discrepancies among individuals and multiple backscattered photons, and according to the transmission model of light in complex biological tissue, then we added the constraint condition——two ideal “banana shape”paths——to define the effective photon ratio(SNR), which was used to select the best source-detector separations (SDSbest), and the results from Monte Carlo simulation modified by adding constraint condition were statistically analyzed, and we regard the SNR as a basis and analyze the relationship between the fat thickness (hf)，the absorption coefficient of a fat layer (μaf)，the absorption coefficient of a muscle layer (μam) and the source-detector distance (SDS), andhfis used as the independent variable to develop a linear regression model to predict SDSbest. The result showed thatμafandμamhave no effect on SDSbestwhen 0

Fat-muscle tissue; Inner information of tissue; Best Source-detector Distance; Spatially diffuse reflectance spectra; Monte Carlo simulation

Introduction

Spectral analysis technique has been widely used to analyze compositions of the scattering liquids such as milk[1], blood, etc, which is of great importance to food safety[2], disease detection[3]and non-invasive measurement of tissue oxygenation[4-5]. Nevertheless, in recent years, one strategy to gain internal information of the measured tissue from reflected photons is using multiple source-detector separations (SDSs)[6], which can obtain a wealth of optical and structural information, however, a significant challenge is that the detection position is ambiguous.

Near-infrared spectroscopy is a non-invasive optical technique to monitor tissue oxygenation, but the thickness of the fat layer has great difference between people or people subjects, that is, the individual discrepancy which is the most difficult problem may influence the measured spectrum, thus it is significant and necessary to eliminate the interference of the individual discrepancy[7]; in addition, previous publications[8]have supported that multiple backscattered photons has lost the accurate information, subsequently, although some photons may travel through the target layer, they are still considered as non-effective photons because it depends on the amount of scattering acts in the investigated media, especially for thick tissues, where multiple scattering is severe. Moreover, the multiple backscattered photons[9]are often ignored in recent years. Therefore, here, to minimize the individual discrepancies and the interference of the multiple backscattered photons, the effective photon ratio (SNR) was defined as a basis for selecting the best detection position.

The aim of the present work is to analyze the relationship between the fat layer thickness, the absorption coefficient of the fat layer, the absorption coefficient of the muscle layer and the best detection position based on the modified Monte Carlo simulation which is developed by Lihong V. Wang[10]. By selecting the best source-detector distance can achieve the selected measurement for the specific target layer and improve the efficiency. This not only provides the early auxiliary function for researching the internal information, but also conducive to the further study on more complex biological tissue, and can help identify the precise position.

1 Spatially Resolved Diffuse Reflectance Spectra

To make sure the effective information from the muscle layer, as shown in Figure.1, for example, we define the two typical “banana shape”paths as ideal boundary: (1) pathⅠensures that the incident photons go through the effective area of the target layer, its path length ispi1; (2) path Ⅱ ensures that the photons are limited scattering photons, its path length ispi2. When the two conditions above are satisfied, in other words, wherepi1

(1)

(2)

(3)

Wherebiis the half of the source-detector separations (SDS).

According to the elliptic arc length formula,ai1，ai2，biare used respectively as a semi-major axis and semi-minor axis to calculate the length of path Ⅰ and Ⅱ, as shown in Eq. (4) and (5).

(4)

(5)

Fig.1 Fat-muscle schematic for Monte Carlo simulation,lrepresents the space between the fibers.hfis the thickness of the fat layer

Whereiis the receiver fiber at positioni.

2 Modified Monte Carlo Simulation

Monte Carlo is a flexible approach to calculate of a large number of random trajectories of photons migrating in the scattering medium basing on its optical properties.

In this paper, we introduced the constraint conditionpi1

(6)

Where,Nnis non-effective photons.

3 Results and Discussion

In order to obtain the diffuse reflectance, tissue as illustrated in Fig.1 was measured by inserting several independent optical fibers held by a fixed support with a separation distancelamong them. The optical parameters of the fat and muscle layer are from previous publications[11-12], and the thickness of the muscle layer is not limited. From Eq. (3),kis a variable, whereas the aim of this paper is to study the relationship between the fat layer thickness (hf)，the absorption coefficient of a fat layer (μaf)，the absorption coefficient of a muscle layer (μam) and the source-detector distance (SDS), so we focused on a particular case ofk=3, and the optical parameters are stated in Table 1, and a packet of 108photons is launched per simulation, and repeat 6 times per simulation.

Table 1 Optical parameters of the fat and muscle layers at 760 nm

Experiences can be separated into the following steps:

(1)hf=0.1 cm, the relationship between SDS and SNR is demonstrated in Fig.2(a) withμafspecified in the range [0.04, 0.05, 0.06, 0.07, 0.08] cm-1; and withμamspecified in the range [0.22, 0.23, 0.24] cm-1, the relationship between SDS and SNR is demonstrated in Fig.2(b);

(2) In the same way, the value ofhfincreases by 0.1 cm untilhfequals to 0.6 cm, as shown in Fig.2(c)—(v);

(3) According to Fig.2(a)—(v),hfis used as the independent variable to develop a linear regression model to predict SDSbest.[Fig.3];

(4) Randomly selecthf=0.12, 0.22 cm to predict the precision of the model [Fig.4 (a),(b) and Fig.5 (a),(b)].

From Fig.2, when 0.1 cm

Fig.2 The graphs on the left study the influence ofμafon the relationship between SDS and SNR whenhfis 0.1, 0.2～0.6 cm respectively; the graphs on the right study the influence ofμamon the relationship between SDS and SNR whenhfis 0.1, 0.2～0.6 cm respectively

In this paper,hfis used as the independent variable to develop a linear regression model to predict SDSbest., as shown in Eq. (7), the correlation coefficientR2is 0.991 8.

SDSbest=1.097hf+0.258 5

(7)

Randomly selecthf=0.12, 0.22 cm, the prediction error is 0.030 14 and 0.020 16 respectively, the error can be controlled within 5%.

Fig.3hfis used as the independent variable to develop a linear regression model to predict SDSbest.

Fig.4 (a) represents the influence ofμafon the relationship between SDS and SNR whenhf=0.12 cm; (b) represents the influence ofμamon the relationship between SDS and SNR whenhf=0.12 cm

Fig.5 (a)represents the influence ofμafon the relationship between SDS and SNR whenhf=0.22 cm; (b) represents the influence ofμamon the relationship between SDS and SNR whenhf=0.22 cm

4 Conclusion

The present paper addresses experimental design considerations, in particular the selection of source-detector separations used. The constraint condition is introduced into the MC simulation to obtain the effective photons (Ns), and the non-effective photons (Nn), and the effective photon ratio (SNR). Simulations are used to characterize the link between the fat layer thickness (hf)，the absorption coefficient of a fat layer (μaf)，the absorption coefficient of a muscle layer (μam) and the source-detector distance (SDS). The result suggested thatμafandμamhave no effect on SDSbestwhen 0

[1] Bogomolov A, Melenteva A. Chemometrics and Intelligent Laboratory Systems，2013, 126(15): 129.

[2] Poliana M Santors, Edenir R Pereira-Filho, Luis E Rodriguez-Saona. Agricultural and Food Chemistry，2013, 61(6): 1205.

[3] Georg Hennig, Herbert Stepp, Ronald Sroka. Applied Optics(USA)，2013, 52(5): 1066.

[4] Lorilee S L Arakaki, Kenneth A Schenkman, Wayne A Ciesielski. Analytica Chimica Acta, 2013, 785: 27.

[5] Michelle L Davis, Thomas J Barstow. Respiratory Physiology & Neurobiology, 2013, 186(2): 180.

[6] Kung-Bin Sung, Kuang-Wei Shih, Fang-Wei Hsu. Journal of Biomedical Optics，2014, 19(7): 077002-1.

[7] Li Chenxi, Zhao Huijuan, Zheng Jiaxiang. Acta Optica Sinica, 2012, 32(7): 0717001-1.

[8] Wu Kai-jie, Li Gang, Lin Ling. Journal of Tianjin University, 2012, 32(3): 755.

[9] Wang Liyun, Li Gang, Li Zhe. Spectroscopy and Spectral Analysis，2014, 34(4): 1026.

[10] Wang L, Jacques S L, Zheng L. Computer Methods and Programs in Biomedicine, 1995, 47: 131.

[11] Ye Yang, Olusola O Soyemi, Michelle R Landry. Optics Express, 2005, 13(5): 1570.

[12] Lin Ling, Chen Yaqin, Li Gang. Chinese Journal of Biomedical Engineering, 2003, 23(5): 419.

*通訊聯(lián)系人

O657.3

A

基于漫反射光譜法下層組織信息的最佳探測位置的研究

劉 妍1,2，楊 雪1,2，趙 靜3，李 剛1,2，林 凌1,2*

1. 天津大學精密測試技術(shù)及儀器國家重點實驗室，天津 300072

2. 天津大學天津市生物醫(yī)學檢測技術(shù)與儀器重點實驗室，天津 300072

3. 天津中醫(yī)藥大學中醫(yī)藥工程學院，天津 300193

由于人體組織結(jié)構(gòu)的復雜性和個體差異性，采用光譜分析技術(shù)實現(xiàn)組織內(nèi)部信息的檢測方法中，仍存在著檢測位置不明確的問題。采用多位置空間分辨漫反射光譜法對脂肪肌肉組織中內(nèi)部信息的最佳探測位置進行了研究。由于目標組織為肌肉組織，為了消除由于不同人體皮下脂肪層厚度的差異以及目標層中多次后向散射光對測量結(jié)果引起的較大誤差，并根據(jù)光在復雜生物組織中的傳輸模型，通過增加約束條件——兩條理想的“香蕉型”路徑—來定義有效光子比SNR，并以SNR作為評價最佳探測位置SDSbest的標準，對優(yōu)化后的Monte Carlo仿真結(jié)果進行統(tǒng)計分析，分別研究了脂肪層厚度hf、脂肪層吸收系數(shù)μaf和肌肉層吸收系數(shù)μam與光源探測器間距SDS之間的關(guān)系，并以hf為自變量與SDSbest建立線性回歸模型。研究表明，(1)當0

脂肪肌肉組織；內(nèi)部組織信息；最佳檢測位置；空間漫反射光譜；Monte Carlo模擬

2015-07-07，

2015-11-20)

10.3964/j.issn.1000-0593(2016)10-3434-08

Received：2015-07-07; accepted：2015-11-20

*Corresponding author e-mail: linling@tju.edu.cn