Rapid Prototyping of Neonatal Jaundice Detector Using ... - IEEE Xplore

2014 IEEE Conference on Biomedical Engineering and Sciences, 8 - 10 December 2014, Miri, Sarawak, Malaysia

Rapid Prototyping of Neonatal Jaundice Detector Using Skin Optics Theory Zulfadhli Osman, Afandi Ahmad-IEEE Member and Azlan Muharam-IEEE Student Member Abstract— When newborns present with jaundice, various clinical assessment need to be executed including evaluation of the history, physical examination, and assessment of severity of jaundice. This project concerns with the quantification of neonatal jaundice level using Simulink model-based design. It is important to address that the proposed method is non-invasive, hence suitable for infants. In this project, the concentration level of bilirubin in dermis layer of skin has been detected using skin optic theory with light emitting diode (LED) and photodiode as a detector. To evaluate the proposed system, a mock skin soaking is used to replace the real infant’s skin. At the system level, Arduino Uno has been used as a hardware platform, whilst MATLAB and Simulink have been fully utilised to implement the system’s architecture. An evaluation of the proposed systems reveals its capability to detect, process and display real-time results with three (3) levels of jaundice, including normal, mild and critical. Index Terms— Bilirubin, Jaundice

I. INTRODUCTION New born baby or infant have high risk to face health problems. This is due to the organs and metabolism just started to develop. One of common health problem synonym with infants is jaundice, which the condition of excessive bilirubin serum in blood stream with the peak bilirubin levels occurring on day five (5) of life [1], [2]. This condition appears because there is a problem with the immature infant’s liver [3], [4]. Too high number of bilirubin pigment or hyperbilirubinemia can brought to neurological disorder, irreparable neurologic dysfunction and even death [5]. To diagnose jaundice, most hospitals in Malaysia using invasive method, by taking a specific amount of infant’s blood and test it in the laboratory. This procedure may give trauma to infant; hence it is not an easy job and has risk of This work was supported in part by Universiti Tun Hussein Onn Malaysia (UTHM). Software used in this project was sponsored by Mathworks for Innovate Malaysia Design Competition 2014. *Z. Osman is with Bioelectronics and Instrumentation Research Laboratory (BioEI), Microelectronic and Nanotechnology Shamsuddin Research centre (MiNTS-SRC), Universiti Tun Hussein Onn Malaysia (UTHM), P. O. Box 101, 86400 Batu Pahat, Johor, Malaysia. (e-mail: [email protected]) A. Ahmad is a leturer at Department of Computer Engineering, Faculty of Electrical and Electronics Engineering, Universiti Tun Hussein Onn Malaysia (UTHM), P. O. Box 101, 86400 Batu Pahat, Johor, Malaysia. (e-mail: [email protected]) A. Muharam is with the VLSI Architecture and Systems Design Research Laboratory, Microelectronic and Nanotechnology Shamsudin Research Centre (MINT-SRC), Universiti Tun Hussein Onn Malaysia (UTHM) Beg Berkunci 101 Parit Raja Batu Pahat Johor 86400, (corresponding author provide phone: 010-4002649; e-mail: [email protected]).

978-1-4799-4084-4/14/$31.00 ©2014 IEEE

blood loss [6]. Furthermore, it makes the infant’s skin become sore because bilirubin serum needs to monitor at frequent interval. Therefore, there is a need for non-invasive method needs to quantify the level of jaundice. It is also suggested by the Neonatal Institute of Child Health and Human Development to have a system for bilirubin fraction measurement which deposits in skin and contribute to jaundice [5]. To portray the importance of this research, literature reviews show that this topic is still immature and further research needs to be executed. Hamza et al. [5] used semiconductor diode laser and diode pumped solid-state lasers laser diodes with 532 and 457 nm, respectively as light sources which directed to infant’s skin, then the light reflected back and absorbed by photodetector. The used of diode laser is considered less economics for prototyping development and also harmful for infant’s skin. In [7], infrared (IR) has been used to absorbed the light from light emitting diode (LED) that pass through a yellow liquid with different concentration. Two (2) specific LEDs with wavelength of 455 and 575 nm have been used by Kudavalley et al. [8] to measure bilirubin concentration in blood samples. The wavelengths were processed by National Instrument (NI) data acquisition (DAQ) card which located in the opposite side. On the other hand, Mansor et al. [9] propose a monitoring system for jaundice using colour detection method. With digital camera and MATLAB software, bilirubin level on the skin has been detected based on different skin’s colour and it is considered a software-based implementation, hence very time consuming process. This research attempts to propose a hardware-based system to quantify jaundice level using artificial bilirubin standard solution (ABSS) [10] that has been widely practiced in laboratory technology. The ultimate aim of this research is to develop a low-cost, accurate, and non-invasive hardware-based system. The remainder of this paper is organised as follows. Section II describes about the experimental setup, systems architecture and its implementation. Experimental results and analysis are described in Section III. Finally, concluding remarks are given in Section IV.

328


Part 1

Part 2

Part 3

Fig. 3: Bilirubin detection systems. Part 1: ADC. Part 2: Bilirubin concentration calculation. Part 3: Output decision system

I. METHODS The general goal of this research is concerned with the design and implementation of jaundice level quantification systems, with more emphasis on skin optic theory. In this research, there were four (4) phases of work packages have been realised as follows: basic experimental setup, initial data analysis, development of the detection systems, circuit design and prototyping of the systems.

Inc has spectral responsivity in range of 300 to 1100 nm with peak irradiation responsivity at 470 nm.

A. Basic Experimental Setup In this experiment, sprague-dawley (SD) rats’s skin with eight (8) weeks old (male) based on skin soaking method has been used. It is worth noting that the SD rat is considered reasonable and practical to predict human skin permeability [11]. SD rats’ skin was then being carefully shaved and soaked in ABSS for three (3) minutes with ten different concentrations that have been prepared earlier. In preparing the ABSS [10], the following guidelines have been followed:

Fig. 1. Basic experimental setup

B. Initial Data Analysis Based on experiment conducted, the voltmeter identified changes in voltages value when different mock skin with different ABSS was tested. Fig 2 that has been plotted using MATLAB illustrates the relationship between these two (2) parameters, ABSS concentration and voltage. To fully utilised the functionality of MATLAB, this graph has applied the best fitting curve and the following equation has been automatically generated.

Materials 1) Methyl red powder Manufacturer: Denver-hill Chemicals

y = −42.776 * x + 88.197

To prepare a 10mg/dL of ABSS, 10 milligrams (mg) of methyl red powder with 1 dL of acetic acids have been mixed. With the experimental setup as depicted in Fig. 1, gallium nitride (GaN) type bright blue LED (455 nm), manufactured by Kingbright Electronic Co. and a high-sensitivity low-noise light-to-voltage optical converter (TSL257), which combines photodiode and transimpedance amplifier on a single monolithic complementary metal oxide semiconductor (CMOS) integrated circuit is used to obtain voltage range of light reflectance from mock skin. The photodiode manufactured by Texas Advanced Optoelectronic Solution

(1)

ABSS Concentration

2) Acetic Acid Manufacturer: Bendosen Molar : 60.05g/mol

Voltage (V)

Fig. 2. Relationship between the ABSS concentration with output voltage from photodiode.

329


C. Detection System To develop the detection systems, three (3) sub-blocks have been designed as shown in Fig. 3. For the first part, it acts as an analog to digital converter (ADC) system. Pin A4 of Arduino Uno was assigned as an analog input terminal. The analog voltage is converted to digital data using ADC equation as follows: ADC Equation = analog voltage value * 0.0049

Power

USB Cable

(1) LED

Then, the converted digital data has been passed through ‘IF Block’ to limit the system decision. If any data less than value of 1.4, it was then considered as a constant value of 1.4. Whilst for any value greater than 2, are considered to be a constant value of 2. These limits value depends on maximum and minimum voltage value from graph shown in Fig. 2. From Part 1, then data will be sent to Part 2 that performs the calculation of bilirubin concentration based on Eq. (1). Finally, in Part 3 the bilirubin value has been sent through ‘IF Block’ for display purposes. The output of particular bilirubin value is assigned as digital output value (High = 1) for next stage operation. Table I shows the output terminal decision of Arduino Uno to display the results. OUTPUT TERMINAL PIN ASSIGNMENT AND BILIRUBIN CONCENTRATION

Concentration ≤ 10 10 < Concentration ≤ 20 Concentration > 20

Number of Arduino Uno Pin 9 10 11

Arduino Uno

Black box Light path Mock skin

Fig. 4. Overall bilirubin detection system.

PIC 16F877A

After that, the digital signal from the Arduino Uno has been passed to the microcontroller (PIC 16F877A). At this stage, the digital signal being processed to display the level of bilirubin and the LED will shows its response. An overview for the overall bilirubin detection system is given in Fig. 4. Meanwhile, the prototype of this research is shown in Fig. 5.

Photodiode

Fig. 5. Prototype of the system.

II. RESULT & DISCUSSION

A.

D. Circuit Design & Systems Prototyping In this research, the bilirubin detection system begins with the projection of blue light from LED to the mock skin in the black box. The reflected light from mock skins gives particular voltage value at photodiode. Then, the voltage value has been sent to and processed by the Arduino Uno to determine its bilirubin concentration.

LED

Black box

To verify that the proposed biological sample (SD rats’ skin) attributes are almost similar to the various level of jaundice, several validation processes have been taken placed. The first two (2) processes were related with validation of the samples that have been prepared, whilst the rest were about the functionality of the proposed prototype.

TABLE I

Bilirubin concentration (mg/dL)

LCD

Spectrophotometer Test

The ABSS solutions were tested using spectrophotometer (Shimadzu: UV-1800). The aim of the test is to observe the wavelength of ABSS and the results are illustrated in Fig. 6. It appears that the ABSS solutions have great absorbance at the range of 400 to 550 nm. Hence, it is proved that the solutions are significant to be used throughout this research due to the range of wavelength peak absorption similar to the real bilirubin serum.

LCD

LED

Fig. 6. Absorbance and wavelength of ABSS with different concentrations.

B.

330

Laboratory Test


To measure the bilirubin serum concentration invasively, the ABSS was then being tested using bilirubin meter (Cosmedi: Bil-100). It is noted that if the solution’s accuracy less than 95%, then the solution need to be re-prepared. This is due to the ultimate aim of this test which is to prepare the most accurate concentration of ABSS solution based on desired value. C.

Bilirubin Detection System

Based on Fig. 2 in Section II, the output voltages from photodiode were varies with ABSS concentration. The voltage values for 10 samples of ABSS concentration have been taken five (5) times for data accuracy. It is always noted that the expected range of responsivity for bilirubin serum lies between 400 to 500 nm. Thus, it is expected that the higher concentration of ABBS, the lower the light intensity reflected back to environment which brought to the lower of voltage output from photodiode. In brief, the output voltage of photodiode is directly proportional to light intensity [12]. It shows that, the capability of the photodiode used in this research to quantify the jaundice level in the form of light intensity and gives particular voltage output. The output voltage value was then being processed in the Arduino Uno to determine its bilirubin concentration and later the value has been sent to the microcontroller. At this stage, the digital signal was being processed to display the level of jaundice and LEDs will appear according to its level. Three (3) levels of jaundice, which are normal, mild and critical, will be displayed on the display part. A summary of voltage range from photodiode output and jaundice level displayed on liquid crystal display (LCD) of the prototype are tabulated in Table II. TABLE II SUMMARY OF RESULTS FOR JAUNDICE LEVEL QUANTIFICATION ABBS concentration (mg/dL) Concentration ≤ 10 10 < Concentration ≤ 20 Concentration > 20

Voltage range (volts) 2.06 – 1.83 1.82 – 1.62 1.61 – 0.00

LED colour Green Yellow Red

III. CONCLUSION This research has explored the potentiality combination of biomedical and electronic engineering in solving medical problem. To date, the more widespread of these combinations contributes for better societal impact. Throughout this research, rapid prototyping of neonatal jaundice detector using skin optic theory has been successfully proposed. It is also proved that Simulink software were able to communicate with Arduino Uno and process the jaundice level. It is also important to address that the proposed prototype was non-invasive, hence relevant for new born baby. ACKNOWLEDGMENT The authors would like to thank to the Innovates Malaysia and Dr Azizan Aziz, a pediatric specialist at Hospital Sultanah Nora Ismail (HSNI), Batu Pahat for supporting this research work. REFERENCES [1] [2]

[3]

[4]

[5] [6] [7]

Jaundice level Normal Mild Critical

[8] [9]

It is worth mentioning that after the Arduino Uno has been successfully process the voltage range, only one (1) signal of active high ‘1’ was generated at one (1) time, either at pin 9, 10 or 11 of Arduino Uno digital output terminal. The digital signal was then being manipulated by the microcontroller and the system architecture in microcontroller has been designed using Flowcode software: a programming method based on flowchart. In terms of user’s functionality, the proposed prototype also has its reset button to reset the entire system for different quantification process.

[10] [11]

331

W. E. Nelson, R. E Behrman, H. B. Jenson, R. M. Kliegman, Nelson Textbook of Pediatries. 16th Edition. New Jersey: W.B Sauders Co., 2000, pp.515 – 519. K. B. Kolman, K. M. Mathieson, C. Frias, “Comparison of transcutaneous and Total Serum Bilirubin,” in Newborn Hispanic Infants at 35 or more weeks of gestation , J Am Board Fam Med 2007, 20, pp.266-271. A. Roggan, M. Friebel, K. Dorschel, A. Hahn, G. Muller, “Optical Properties of Circulating Human Blood in the Wavelenght Range 400 – 2500 nm”. Journal of Biomedical Optics, 4(1); pp.36-46. 1999. I. S. Saidi, S. L. Jacques, M. Keijzer, F. K. Tittel, “Optical fiber probe monitor for neonatal bilirubinemia.” Proceeding of the 11th International Conference of the IEEE Engineering in Medicine and Biology, pp.12001-12012, 1989. M. Hamza, M. H. S. El-Ahl, A. M. Hamza, “New laser source for clinical treatmentand diagnostic of neonatal jaundice” in Proceeding of SPIE Vol.4254, pp.187-193, 2001. V. Tuchin, “Tissue optics light scattering methods and instrument for medical diagnosis” in The International Society for Optical Engineering, USA; 2000. H. Baharuddin, M. S. Sulong, A. Joret, T. A. Rahman, N. H. Ismail, “Bile pigments detection via IR sensor” in Proc. of EnCon, Kuching, Sarawak, 2010, pp.1-3. S. Kudavelly, P. Keswarpu, S.A. Balakrishnan, “A Simple and accurate method for estimating bilirubin from blood” in IEEE Journal. 2011. M. N. Mansor, S. Yaacob, M. Hariharan, S. N. Basah, A. Jamil, M. L Mohd Khidir, M. N. Rejab, K. M. Y. Ku Ibrahim, A. K. Junoh, and S. A. Saad. Jaundice in newborn monitoring using color detection method. International Workshop on Information and Electronic Engineering (IWIEE). Elsevier. 2012. R. Sood, “Preparation of calibration curve” in Medical Laboratory Technology, 5th ed., Jaypee Brothers Publisher, 2005, pp. 396. H. Takeuchi, Y. Mano, S. Terasaka, T. Sakurai, A. Furuya, H. Urano, K. Sugibayashi, Usefulness of Rat Skin as a Substitute for Human