A Technique for the Early Detection of Brain Cancer Using Circularly Polarized Reconfigurable Antenna Array

In this paper, a technique for the early detection of brain cancer was proposed. The technique depends upon the use of an antenna and a model for the human head. The reflection coefficient (S11) was found for two cases: with and without tumor in the head model. The large increase of S11 due to the presence of the tumor provides a good indication for tumor detection. It also gives an idea about the size of the tumor. The antenna used was a reconfigurable four-element linear array of squared microstrip patches. Two arrays were designed one circularly polarized, the other linearly polarized. The antenna operates at Industrial Scientific and Medical (ISM) frequency 2.4 GHz. It was designed on FR-4 (lossy) substrate of relative permittivity 4.3 and thickness of 1.6 mm. To feed the array, a corporate feeding network was designed. The reconfigurability of the array was achieved using three single pole double throw (SPDT) switches. Two models of the human head were used; a specific anthropomorphic mannequin (SAM) model, and a 3-D head model consisting of four different head layers: skin, fat, skull and brain. The simulation calculates the reflection coefficient (S11) with and without tumor for circularly polarized (CP) and linearly polarized (LP) linear array. Calculations were taken for four sizes of the array. The best results were obtained with the four-element circularly polarized array. An increase in S11 of 1188% was obtained. Tumors as small as 5 millimeters (four-layer model) and 2.5 mm (SAM model) can be detected. Specific absorption rate (SAR) was calculated and found to be within the safe limit. A circularly polarized four-element linear antenna array was fabricated. The measured S11 and radiation pattern are in excellent agreement with simulated ones.


I. INTRODUCTION
There is no doubt that, cancer is considered as one of the most serious diseases. It is the growth of abnormal cells in human body. According to the National Cancer Institute, there are over 100 types of cancer [1], [2].
Brain cancer is one of the most dangerous types of cancers. This is because brain is the most critical and complex organ of the human body, as it controls the nerves leading to all other organs. Brain tumours are categorized as primary or secondary. A primary brain tumour originates in the brain.
The associate editor coordinating the review of this manuscript and approving it for publication was Mohammad Zia Ur Rahman .
A secondary brain tumour, occurs when cancer cells spread to brain from another organ, such as lung or breast [3]. Brain tumor is diagnosed by several methods such as skull x-rays, biopsy, computer tomography (CT) or magnetic resonance imaging (MRI) on head [4]. Recently, microwave tomography and radar-based imaging techniques are other ways to detect and diagnose tumors. These are low cost, and portable methods. The difference in the dielectric properties between healthy and non-healthy brain tissues is the basis for diagnosis in microwave systems [5]- [7].
The main purpose of this paper is to present a technique for the early detection of brain cancer. A reconfigurable circularly polarized patch antenna array which is simple, low cost and efficient was designed. Detection of the tumor is based upon measurement of S 11 due to single, double, triple and quadruple antenna elements. This is accomplished by using three SPDT switches [8]. Two models of the head were used; SAM model and a model consists of four layers; skin, fat, skull, and brain [9]. S 11 was computed for two cases: with and without tumor. The difference between them was used for the detection of the tumor. Tumors tissues have higher dielectric constant compared to normal tissues. The specific absorption rate (SAR) was computed and found to be less than the safety level [10], [11].
A circularly polarized four-element linear antenna array was fabricated. The measured S 11 and radiation pattern are in excellent agreement with simulated ones. Simulations were carried out using both finite element and finite integral techniques.
The paper is organized as follows. Section 2 includes design of the proposed antenna, design of the corporate feeding network, description of the SPDT switch, the fabricated antenna and the measured S 11 and radiation pattern. In Section 3 simulation results of circularly polarized and linearly polarized arrays are presented. The difference in reflection coefficient ( S 11 ) between the two cases: with and without tumor is given. The results due to the SAM model are given in section 4. Section 5 includes the calculated values of SAR. Finally, in section 6 the conclusions are presented.

II. PROPOSED DESIGN A. ANTENNA GEOMETRY
The CP reconfigurable patch antenna array on a head phantom for brain tumour detection is shown in Fig. 1. The squared patch antenna is designed on FR4-substrate with relative permittivity (ε r ) = 4.3 and thickness 1.6 mm with loss tangent (δ) = 0.025. The dimensions of the array are 200mm × 78mm. The substrate is above a rectangular metallic ground plane with thickness 0.035mm. The side length of the square patch is 30 mm. The antenna is fed by a corporate feed network. When the corners of the patch are truncated [17]- [19] the antenna radiates circularly polarized wave. The truncated part is a right isosceles triangle with side 5mm.

B. FEEDING NETWORK
The corporate feed network and antenna patches are drawn in Fig.1. The input impedance of each patch [20] is given by: and, where, W 1 is the side length of the squared patch (30 mm), and λ 0 is the free space wave length equals 125 mm corresponding to the resonance frequency 2.4 GHz. Refer to ''(1),'' the input impedance of the patch equals 250 . The quarter wave length transformer T 1 matches the patch to 100 . The characteristic impedance [21] is obtained by: where, Z in is the characteristic impedance of the transmission line and Z out is the input impedance of the patch at resonance. The characteristic impedance of the quarter wave length transformer T 2 which connects the 100 and 50 transmission lines is The width (W 2 ) of the transmission line with characteristic impedance (Z o ) is given by [22]:  where, and The lengths of the transmission lines can be found from [24]: where, The widths of the transmission lines at 50 , 70.7 , 100 , and 158 input impedance are shown in Table 1.
The input impedance of the antenna array and corporate feed network was computed. Simulations were carried out using CST and HFSS simulations tools. Fig. 2(a) shows the real part of the input impedance which is 50 at the design frequency 2.4 GHz. The imaginary part of the impedance is shown in Fig.2 (b) which is zero at the operating frequency, so perfect matching is achieved. Additionally, three single pole double throw switches (SPDT) S 1 , S 2 , and S 3 are used to reconfigure the antenna array. The number of operating patches of the array is controlled by the switches. The circuit diagram of the (SPDT) is illustrated in Fig. 3, D 1 and D 2 are PIN diodes. To link ports 1 and 2, D 1 must be OFF and D 2 ON. Thus, when D 2 is ON (short circuit) the input impedance to the λ 4 line will be open preventing signal from reaching port 3. Similarly; to connect ports 1 and 3, D 2 is OFF and D 1 is ON. Therefore, in the corporate feeding network port 2 represents the patch and port 3 is set to 250 to replace the input impedance of the patch.
The switches cause a small shift in resonance frequency. They also cause small loss in power which affects efficiency.
The head phantom contains four layers skin, fat, skull, and brain with side lengths 110mm, 108 mm, 101 and 94 mm respectively [9], [27]. The cancerous tumor is considered spherical with relative permittivity 70 and electric conductivity 2.3 [24], [25]. The dielectric properties of human model tissues are illustrated in Table. II.
The axial ratio is shown in Fig.4. It can be shown from this figure that the axial ratio obtained by CST simulation is

C. EXPERIMENTAL
The full size (four-element) array gave a large increase of S 11 of 1188%. Therefor it was decided to fabricate this array.  The full size array does not need switches. The fabricated array is shown in Fig. 5. The radiation pattern and S 11 were measured in free space. Antenna measurement system of Geozondas [26] was used (Fig. 6), which is a pulse time domain measurement system. The measured and simulated S 11 are shown in Fig. 7. Measurement shows resonance at 3 GHz. Therefore, the simulation was extended to 4 GHz and showed resonance at 3.75 GHz. The designed antenna is intended to work at 2.4 GHz, therefore the antenna behaviour away from this frequency is out of our interest. The impedance bandwidth of the antenna is determined by taking the 10-dB bandwidth of S 11 which is equal to 88.3 MHz in simulation and 90.9 MHz in measurement.
Moreover, the measured and simulated radiation patterns are shown in Fig.8. Very good agreement between simulation and measurement can be noticed.

III. SIMULATION RESULTS AND DISSCUSSION A. CP ANTENNA ARRAY SIMULATION
In this subsection, the CP reconfigurable antenna array with head phantom is simulated. The changes of reflection coefficient ( S 11 ) is calculated to distinguish between head phantom with and without tumor as; The number of patches operating in the array is controlled by the SPDT switches. Four sizes of the array are investigated and simulated. The return loss is calculated for each array size put upon human head model with and without tumor. The results are summarized in Table 3.

B. LINEARLY POLARIZED (LP) ANTENNA ARRAY
Linear polarization is obtained when the square patches have no truncation at the corners. The resonance frequency becomes 2.3 GHz. The wave length λ o becomes 129 mm. The input impedance of the patch Z in becomes 259.5 . The characteristic impedance of the quarter wave transformer T 1 changes to 161 . The characteristic impedance of the quarter wave transformer T 2 stays at 70.7 . Hence, the width of the quarter wave transformer T 1 can be calculated from equations (5-7) and changed to 0.138 mm. The simulation results of S 11 for the four sizes of the array are shown in Table 3. There is a great difference between the results of the S 11 in circular polarization and linear polarization. This means that circularly polarized array is much better than the linearly polarized array for the detection of brain tumor. Table 3. summarizes the results of S 11 for CP and LP and the percentage increase of S 11 for CP over LP. Fig. 9 shows S 11 with the size of the array for CP and LP. The CP arrays give a large difference in S 11 between the two cases: with and VOLUME 9, 2021  without tumor. Circular polarization gives much better results than linear polarization. Fig. 10 shows variation of S 11 with tumor size. Reflection increases as the tumor size increases. This enables us not only to detect the presence of tumor but also to predict the size of the tumor.

IV. VALIDATION OF THE RESULTS WITH SAM HEAD MODEL
In this section SAM head model in CST is used to find S 11 . SAM head model and CP array are shown in Fig.10(a). The CP antenna array is put above the head model. The results of S 11 are presented for different sizes of the tumor. The differences in the reflection coefficient S and the shift in resonant frequency F are summarized in Table 4. Fig. 11 (b) shows the simulated results of the reflection coefficient at each stage of tumor. It can be noted that, tumors as small as 2.5 mm can be detected.

V. SPECIFIC ABSORBITION RATE (SAR)
In this section SAR for 10 grams of tissue are computed for the head phantom with CP array. The value of SAR is important for human body safety. SAR can be calculated from the following formula [11]: where σ is the electric conductivity of tissue (S/m), E is the electric field intensity (V/m) and ρ is tissue mass density (Kg/m 3 ). According to IEEE C95.1, the limit value of SAR is 2W/kg. Table 5 records the value of SAR for the four sizes of the array. There is little increase in SAR due to the presence of the tumour and it is still below the safety level.
A comparison between published works and the current work has been introduced in Table 6.
This table shows that researches depend upon one or more of the following parameters for the detection of tumor S 11 , SAR, current density, E-field, and H-field. Two papers refer to increase of current density of 1.92% and 0.5%. It is clear that the current density cannot be depended upon for tumor detection. Two papers refer to an increase of magnetic field of 170% and 3.9%. Two   papers refer to an increase of electric field of 79% and 0.04%.
Five papers refer to an increase in SAR of 138.7%, 118%, 200%, 88.9%, and 207%. It is not enough to mention the increase in SAR due to tumor. It is important to mention if the system is within the safe limits. Four papers refer to an increase in S 11 of 4.86%, 13%, 118%, and 7.76%. The proposed system in this paper shows an increase 0f 21.5 dB in S 11 which corresponds to a percentage increase of 1188%. This shows the superiority of the proposed system.

VI. CONCLUSION
A reconfigurable four-element circularly polarized linear antenna array was designed for brain tumor detction. A corporate feeding network was designed to match the antenna to the feeding source. Reconfigurability of the array was achieved using three single-pole double throw switches. The switches used PIN diodes. S 11 was found for the antenna array on head phantom with and without tumor. S 11 was found for four sizes of the array. The difference S 11 between the two cases; with and without tumor was found to be 3.1, 10.58, 17.41, 21.5 dB. The last three values are large enough to enable detection of brain tumors. The difference S 11 enable detection of small tumors (5 mm).The SAM head model in CST was simulated and S 11 was computed. Tumors as small as 2.5 mm can be detected. This enables early detection of brain cancer.
A four-element linearly polarized linear array was designed and S 11 was found. S 11 for circular polarization was found to be larger than that for linear polarization by 377.78%, 412%, 360%, and 287.5% for the four sizes of the array. Circular polarization gives much better results. The values of SAR were calculated for the four sizes of the array and were found to be 0.097, 0.095, 0.091, and 0.081 W/Kg. These values are below the safe limit of exposure (2 W/Kg).
The four-element circularly polarized linear antenna array, but without the SPDT switches, was fabricated. The radiation pattern and S 11 were measured in free space. Very good agreement between simulation and measurement was noticed. He is currently working as a Researcher with the Electrical Engineering Department, King Saud University. He has published more than 20 international publications and holds a patent for chipless RFID tags. His research interests include reconfigurable antennas, chipless RFID tags, biomedical engineering systems, and RFID sensor. In December 2011, he was appointed as a Lecturer and later as an Associate Professor of the Department of Electronics and Electrical Communication Engineering, Menoufia University, in August 2018. His research interests include numerical techniques, CAD tools, programming languages, UWB antennas and systems, EBG structures, wireless sensor networks and cognitive radio, metamaterials, the IoT, mobile antennas, and image processing.