A novel miniaturized radiometer front-end for early breast cancer detection using ultra-wideband flexible antenna

ABSTRACT


INTRODUCTION
Breast cancer represents a significant public health issue and an event of great significance in the life of women, this disease occupies first place in terms of mortality and incidence.According to statistics published in 2022 [1].The International Agency for Research on Cancer (IARC) has estimated the number of cases of breast cancer in the last 5 years to be 2.261 million (58.5% rate), and 684,996 deaths (17.7% rate).In Morocco, its incidence continues to increase, approximately one in ten women are at risk of developing breast cancer during their lifetime [2], this is an alarming number that increases with age.Researchers have shown that screening breast cancer once every 1 to 2 years is not sufficient to detect breast cancer since patients with fast-growing breast cancer account for a quarter of all breast cancer patients, so it is appropriate to combine screening with other non-invasive investigative methods [3].
The passive breast imaging technique based on the microwave radiometer is a method that allows measuring the intensity of natural electromagnetic radiation of biological tissues of the human body, it's harmless and without danger for the patients and the medical staff, this intensity is proportional to the  [4]- [7], indeed, the change in temperature (case of thermal abnormality) is caused by an increase in the metabolism of cancer cells, the generation of specific heat in the tumor is proportional to the growth rate of the tumor.Passive microwave radiometry is a very promising technique for early breast cancer detection, which also allows the detection of fast-growing tumors first [8], [9].
The objective of this work is to model a miniature passive microwave radiometer system for the early breast cancer detection, using a flexible ultra-wideband antenna operating in the S-band, and ergonomic with the breast.The technology adopted is microstrip in order to reduce costs and achieve high performance in detecting cancer cells.
This article is organized as follows: the section 1 presents the operation principle of the microwave radiometer with theoretical studies, thereafter we presented the model of healthy and abnormality of breast phantom that was used in our study, as well as the design and simulation of the miniaturized flexible antenna and its application to the breast phantom using computer simulation technology microwave (CST MWS) software [10].The following section 2 presents the design and simulation of the microwave radiometer front-end containing the bandpass filter, radio frequency (RF) power detector, and low noise amplifier (LNA).Section 3, describes the measurement results and interpretations of the overall microwave radiometer front-end system, and finally, in section 4, some conclusions and perspectives are proposed.

METHOD 2.1. Operating principle of the proposed microwave radiometer front-end
Radiometry is a non-invasive breast imaging modality [11], [12], non-toxic and relatively inexpensive, which is based on measuring the power of electromagnetic noise emitted by lossy materials using passive receivers [13].Several studies carried out in the literature concerning the architectures of the passive microwave radiometer front-end [14]- [16], have shown that the most used is the Dicke radiometer.This type of radiometer consists of a receiving part (antenna) which allows detecting the power of the electromagnetic noise, which is related to the thermal radiation self-produced in the breast tissue, a low noise amplifier, a bandpass filter and a detector to convert RF waves to direct current (DC) signals.
Figure 1 shows a schematic a depiction of a microwave radiometer front-end, connected to a flexible antenna to measure the thermal gradient on the breast surface.The concept of microwave thermal measurement is based on Planck's law.Planck [17] showed that spectral radiation is emitted by hot bodies, he formulated an equation relating to the electromagnetic radiation of a black body, as shown in (1): Where  is the speed of light, ℎ is Planck's constant,  is Boltzmann's constant,  is the blackbody spectral luminance,  is the temperature in (), and f is frequency (Hz).

Design of the receiving part of the radiometer 2.2.1. Breast phantom model
According to the literature [18], [19], the structural composition of the breast has a hemispherical shape, and for its physiological characteristics consist of four essential layers, namely skin, fat, mammary gland, and muscle (see Figure 2(a)).The Figure 2(b) shows the longitudinal sectional view of the finite element grids of the breast model.The breast was modeled as a hemisphere using CST MWS software as shown in Figure 3, adhering to the physical characteristics such as the relative dielectric permittivity, the electrical conductivity chosen in the S-band, and also geometric (thickness of each breast layer and the tumor radius) as listed in the Table 1.

Microwave radiometer antenna design
In this work, the proposed patch antenna will be considered as a front-end sensor of the radiometric system that is used non-invasively and adapted to the anatomical structure of the breast, and must also ensure good contact so that antenna characteristics are not degraded.Before proceeding with antenna design, the requirements listed in the Table 2 must be respected.Based on our laboratory's research work on the antennas development [21]- [23] to benefit from previous experiences in this field, a new antenna structure has been developed for the early breast cancer detection.The antenna selected in this work is a circular, planar, flexible, miniature, ultra-wideband (UWB), and a microstrip feed, with a circle-shaped patch on the front face and a rectangular-shaped partial ground plane on the rear face.The structure shown in Figure 4 was modeled and simulated using CST MWS.The enhancement and evaluation patch antenna were carried out by numerous iterative simulations in the S-band.The choice of the S-band is justified by its good penetration of several centimeters into the breast, calculated according to the skin depth δm presented in (2) [24]: Where µ is the material permeability,  is the electrical conductivity, and  is the frequency.
The Table 3 shows the different optimized parameters of the proposed flexible antenna.Kapton polyimide was chosen as the antenna substrate because of its low tangent loss factor over the entire S-band equal to 0.02, and its dielectric constant ℰr=3.4.From Figure 5, we observe that we have a better adaptation for a frequency band (2.3 to 6 GHz) (< -10 dB), with a return loss S1,1=-48,689 dB for 0 =2.75 GHz, and an ultra-wideband (UWB=3.67GHz) from 2.33 GHz to 6 GHz, with a good voltage standing wave ratio (VSWR) which is equal to 1.0084 < for the resonant frequency 2.75 GHz of the proposed antenna (see Figure 6), which shows the improvement of results.Figure 7 shows the radiation pattern of the proposed patch antenna.
Based on the design of the breast model using CST MWS, two approaches were studied in terms of measuring the S1,1 parameter of the ultra-wideband flexible antenna, the first is a phantom with tumor and the second approach is a phantom without tumor in order to analyze simulation results.From these results allowed us to conclude whether or not cancerous tumors were present.For the first scenario of the healthy breast phantom, the S1,1 parameter had an impact compared to the results obtained from the flexible antenna positioned in free space (see Figure 8), indeed, the resonant frequency has been changed from 2.749 GHz to 2.108 GHz with a frequency variation equal to |∆ =  0 −  1 |=640 MHz.For the second scenario of the abnormality breast phantom, the change was rather observed in the value of the resonance frequency, indeed, the value of the said frequency varies from 2.108 GHz for the case of the flexible antenna positioned in contact with breast phantom without tumor, to 2.953 GHz for the case of the presence of the tumor, with a difference in frequency is equal to |∆ =  2 −  1 |=845 MHz (see Figure 8).It can be concluded that the antenna developed in this work exhibits a very different propagation behavior from that in free space, and therefore, it can detect cancer cells in the virtually created breast phantom.

283
The simulation result of the proposed filter in terms of reflection coefficient S1,1 and insertion loss S2,1 is shown in Figure 10.We can observe that the resulting bandwidth is equal to 254 MHz at a center frequency of 2.75 GHz, with the S1,1 of -48 dB and the S2,1 is equal to -0.36 dB.The positing of the two transmission zeros are situated at 2.65 GHz and 3.07 GHz, indicating a clear cutoff before and after the passband.The simulation results show that this filter has a reasonable reflection coefficient and insertion loss while offering overall filter dimensions of 7×14.8 mm 2 .The presented filter demonstrates a large bandwidth behavior and a fractional bandwidth (FBW) of 12%, which justifies that the proposed filter is feasible for our microwave radiometer front-end system.Figure 10.Return loss S1,1 and the insertion loss S2,1 of the proposed microstrip bandpass filter

Radio frequency-direct current power detector
The design of the RF-DC power detector by CST MWS is divided into three parts: diode, low pass filter, and adaptation network [26].Figure 11 shows the block diagram of the RF-DC power detector.The Table 5 compares the critical parameters for different schottky diodes.The Table 5 shows that the SMS 7630 schottky diode offers the best sensitivity for low forward powers among commercially available schottky diodes operating in the S-band.According to the studies carried out on the choice of the schottky diode, we proceeded to the modeling of its simulation program with integrated circuit emphasis (SPICE) model using CST MWS.Table 6 shows the SPICE model parameters of the SMS 7630.  Figure 12 shows the SPICE subcircuit model of the SMS 7630 that we developed by CST MWS.The first circuit developed is a schottky SMS 7630 diode circuit connected in series.The interest of the series diode structure is that it can detect low power levels emitted by the tumor.The substrate used is FR-4, with a a thickness of Hsub=1.6 mm and a relative permittivity εr=4.3.The developed circuit has 5 blocks, namely two RF and DC filters to block higher harmonics, a schottky diode, the load, and an impedance adapter located between the RF filter and the diode, whose role is to enable maximum power transfer.Figures 13(a) and (b) show respectively, the circuit design using CST MWS and its equivalent circuit, whose RF-DC detector parameter values are mentioned in Table 7.The response of RF-DC power detector (see Figures 14(a) and (b)) is optimized by using CST MWS, which corresponds to a low pass filter, whose cutoff frequency is 2.75 GHz with the rejected band varying from 2.75 GHz to 4 GHz.The result of the voltage standing wave ratio (see Figure 14(c)) is less than 2 in the frequency band (2.5-3 GHz) of the proposed detector, which shows a good improvement of the results.

Design and numerical simulation of low noise amplifier
The important aspect of the microwave radiometer is the amplification of a low-amplitude signal.Some essential elements of the amplifier design are the study of its stability.When designing an amplifier, it's important to assess the stability of an amplifier over the frequency band used.There are simpler tests to assesing the stability of an amplifier [29].The first test is referred to the "-", it provides unconditional stability if the following conditions are achieved: There is another criterion has been formulated, which consists a single parameter to determine the unconditional stability, namely µ, which is defined by: We used two software, altium designer was chosen to design the low noise amplifier bearing the reference 'MAAL-011078' (see Figure 15(a)), and CST MWS to characterize it in terms of S-parameters and stability.According to the MACOM producer of 'MAAL-011078' (see Figure 15(b), we chose the S-parameter file with the extension '.s2p' relating to 'MAAL-011078', which we then imported into a linear array block 'touchstone' (see Figure 16).
Figures 17(a) and (b) respectively represent the simulation results of S-parameters obtained by CST MWS and those given by MACOM producer.It's quite clear that the different results obtained are identical, indeed, the gain value of 'MAAL-011078' calculated by CST MWS is 21.62 dB in the S-band, and in terms of adaptation, the S1,1 and S2,2 are always lower than -3 dB, which are similar to those given by MACOM.In terms of stability, Figure 18 presents the parameter values (, µ (or ),  (or delta)) for the MAAL-011078 component simulated using CST MWS.We notice that the MAAL-011078 respects the stability

. Global microwave radiometer front-end system
From the simulation and interpretation results of the components designed in the previous sections, we were able to develop the global system of the radiometer front-end, associated with the normal and abnormal breast phantom, using the co-simulation modules provided by the CST MWS, as shown in Figures 19 and 20.The amplification must be quite high, for this, we used two low noise amplifiers of 'MAAL-011078' from the MACOM producer.A second-order bandpass filter presented in the previous section is inserted just after the amplifier, it delivers maximum gain in the band (2.71 to 2.97 GHz).The RF power detector is inserted as the last block, it makes it possible to convert the RF power into DC power, and to deliver the analog signal in the output of the RF radiometer front-end (port 6 and port 12) proportional to level of the noise power at the input (port 1 and port 7).In addition, this signal is proportional to the measured temperature which is then converted using an analog-to-digital converter (ADC) contained in a microcontroller, and the digital signal is then sent to the computer for processing.In this section, we inserted inside the breast phantom a mass of cancerous cells, has a diameter of 5 mm with its electrical conductivity (=4 S/m as shown in Table 1) keeping the same structure of the frontal stage of the microwave radiometer presented previously.

Interpretation of simulation results
The receiver's two most critical characteristics are gain value and the verification of it's stability.Figure 21 shows the gain variation at the radiometer output for two scenarios (with and without tumor).The proposed scenarios gave the expected gain in the S-band frequency, indeed, the difference of the maximum gain |ΔGmax| measured between the abnormal and healthy phantoms is about 0.92 dB at 2.75 GHz.
Figure 22 summarizes the different simulations of the parameter's stability, we notice that the stability parameters  and  are greater than 1 in the S-band (see Figures 22(a) and (b)), and the delta is less than 1 (see Figure 22(c)).We can conclude that the results obtained justify the stability of the radiometer front-end associated with the breast phantom.

CONCLUSION
This article describes a global study of a model of the microwave radiometric system, that is capable of detecting the presence of an anomaly inside the mammary breast, using the CST MWS simulator.We were able to study the performance and the implementation of a new miniaturized microwave radiometer for early breast cancer detection, thanks to simulations carried out on the breast phantom for both scenarios (healthy and abnormal).The results obtained from this work are satisfactory in terms of reflection/transmission coefficients, stability, and gain.For the next works, we plan to realize this system, moreover, there are two main area that may be improved in this system, firstly, due to the need for a relatively low gain, using a single amplifier is viable option, which will minimize the number of elements in the circuit and then minimize its size.

Figure 1 .
Figure 1.Block diagram of radiometer system for breast cancer detection

Figure 4 .
Figure 4. UWB flexible antenna design using CST MWS

Figure 8 .Figure 9 .
Figure 8. Results of S1,1 parameter for the flexible antenna positioned in free space, in contact with the healthy and abnormality of breast phantom

Figure 11 .
Figure 11.Functional diagram of the system

Figure 19 .
Figure 19.Co-simulation of normal breast phantom associated with the radiometer front-end part

289 Figure 21 .Figure 22 .
Figure 21.Gain variation measured at the output of the radiometer front-end

Table 2 .
Antenna requirements Receive low-power signals from different depths of breast tissue Frequency S-band Operate with a minimum reflection coefficient Dimension Miniature Easy to integrate into the screening device Use Contact with breast Operate near the breast

Table 3 .
Proposed flexible antenna parameter values

Table 4 .
Microstrip bandpass filter dimensions A novel miniaturized radiometer front-end for early breast cancer detection using … (AchrafElouerghi)

Table 7 .
RF-DC power detector microstrip line parameter values