MICROWAVE DIELECTRIC SPECTROSCOPY

OF LOW-VOLUME FRACTION HUMAN CANCER CELLS

EMBEDDED IN COLLAGEN GELS

Experimental Feasibility Study with an Open-ended Coaxial Probe

Stéphane Egot-Lemaire*, Pierre-Olivier Bagnaninchi*, Jacek Pijanka*

Josep Sulé-Suso** and Serguei Semenov*

Institute for Science and Technology in Medicine,*Keele University and **University Hospital of North Staffordshire

Guy Hilton Research Center, Thornburrow drive, Stoke-on-Trent, U.K.

Keywords: Cancer cells, SK-MES cell line, Collagen gels, Volume fraction, Dielectric spectroscopy, Complex

permittivity, Open-ended coaxial probe, Microwaves.

Abstract: This paper addresses and demonstrates the feasibility for microwave dielectric spectroscopy to detect small

volume fractions of SK-MES lung cancer cells embedded in collagen gels with an open-ended coaxial

probe. Measurements were performed on the frequency range 200 MHz – 2 GHz. For all the cell volume

fractions tested (1.4%-4.4%), a significant difference in complex permittivity was observed between

composite gels (containing cells) compared to gels alone. Statistically significant changes were especially

found in the real part of the permittivity, which decreased consistently when the volume fraction increased.

1 INTRODUCTION

The dielectric properties of biological tissues and

cells have been widely investigated for many

decades (Foster, 1996; Gabriel 1996). They are

characterized by the so-called complex permittivity,

expressing the polarization response of a material in

the presence of a time-varying applied electric field.

Measuring its complex response as a function of the

field frequency is referred to as dielectric

spectroscopy (DS).

The difference in the dielectric properties

between various biological tissues is exploited in

electromagnetic tomography techniques, such as

biomedical microwave imaging, a promising

imaging modality (Tofighi, 2001; Semenov, 2003;

Fear, 2005). One of its particular interests is the

ablility to detect tumours (Hagness, 1998; Bulyshev,

2001; Shao, 2005; Bindu, 2006). Indeed malignant

tissues have mainly been found to have significantly

different dielectric properties than the corresponding

normal tissues regarding both the real and the

imaginary parts of the complex permittivity

(Chaudary, 1984; Surowiec, 1988; Smith, 1986;

Joines, 1994; Sha, 2002). In this regard, DS studies

related to cancer have mostly dealt with bulk tissues.

In these kinds of studies bulk tissues are either

investigated in vitro or in vivo which have both

disadvantages. In vitro studies have mostly

investigated non-living tissues. Moreover it is

difficult to deal with in vivo human tissues mainly

because surgery is needed, and it does not

necessarily give information at cell level.

On the contrary, working on cell culture samples

is a good alternative way to get a better biophysical

knowledge of living cells. In this regard, a number

of DS experiments have been carried out on various

types of cell suspensions. The most commonly

investigated cell suspensions, whatever the volume

fraction (ranging from a few percent to roughly

70%), have logically been blood samples (Lisin,

1996; Chelidze, 2002; Bordi 2002; Jaspard, 2003;

Treo, 2005) and yeast suspensions (Claycomb,

2002). A few of them have especially focused on

measuring the dielectric properties of white blood

cancer cells in suspension which were shown to be

also different from those of normal cells (Polevaya,

1999; Ermolina, 2001). Furthermore, in vitro cell

culture samples embedded in microporous scaffolds

have also been successfully investigated for tissue

engineering purposes (Bagnaninchi, 2003 and 2004).

156

Egot-Lemaire S., Bagnaninchi P., Pijanka J., Sulé-Suso J. and Semenov S. (2008).

MICROWAVE DIELECTRIC SPECTROSCOPY OF LOW-VOLUME FRACTION HUMAN CANCER CELLS EMBEDDED IN COLLAGEN GELS -

Experimental Feasibility Study with an Open-ended Coaxial Probe.

In Proceedings of the First International Conference on Biomedical Electronics and Devices, pages 156-161

DOI: 10.5220/0001055401560161

 SciTePress

The method was proved to be a good way to monitor

cell growth and differentiation in scaffolds used in

tissue engineering. The aforementioned studies

dealing with cell culture samples were shown to be

able to retrieve cell signature. This term refers to the

dielectric properties of the main cell compartments,

such as membrane, cytoplasm but also nucleus.

These so-called cell signatures were shown to be

distinct for different cell lines or types. This

property allows for cell separation of cancer cells

from normal cells by dielectrophoresis (Gascoygne,

1997). In these different studies, measurements have

covered different parts of the frequency spectrum,

extending from the extremely low frequencies to the

lower part of the microwave range depending on the

measurement method used.

The motivations of the present study are mainly

threefold. The first aim would be to get a better

biophysical understanding of the differences in

dielectric properties between epithelial lung cancer

cells and the corresponding normal ones. Secondly,

DS could also be used as a method to analyse the

response and effectivnesses of various anti-cancer

treatments such as chemotherapy drugs on in vitro

cell culture samples. Some authors have already

used DS in the radio and microwave frequency

range to do so (Santini, 1991 and 1995; Hübner

2005; Duncan, 2006). Thirdly, this study is a

preliminary step to investigate the capability of DS

to be used on intraoperative tissue biopsies for on-

line assessment of tissue resection effectiveness.

To address these issues, our approach is an

adjunct approach of cell suspensions and allows to

have an in vitro realistic biological 3D model for

living lung epithelial cells. As a relatively dense

matrix, a collagen gel is a good model for

investigation. Collagen is the major component of

the natural extracellular matrix (Pietrucha, 2005) and

has already been used in other biomedical studies

involving similar lung cancer cell culture samples

(Yang, 2004). We used low volume fractions of cells

in order to have an in vitro system that could detect

small number of cells so this could have the clinical

application of detecting tumours when they are still

small enough to undergo radical treatment.

This paper deals with the preliminary

investigation of the dielectric properties of lung

cancer cell samples embedded in collagen gels in the

lower part of the microwave range (UHF). The cells

in question belong to a human epithelial lung cancer

cell line, namely SK-MES cell line. This paper

especially addresses the feasibility of distinguishing

low-volume fraction of these cells from the matrix in

which they are embedded in this part of the

frequency spectrum.

The first section describes in details the materials

and experimental set-up used and the second section

gives and discusses some obtained results.

2 MATERIALS AND METHODS

We mainly conducted 6 different experiments, on

6 different cell volume fractions: 1.4%, 1.9%, 2.7%,

3.2%, 3.8%, 4.4% corresponding respectively to

about 4, 5.3, 7.4, 9, 10.6, 12.3 million cells per gel.

They were counted with the aid of a grid-counting

chamber (Hycor Kova Glasstic). The volume

fraction was estimated by supposing the cells are

spherical. Their diameter (mean: 18 μm ± standard

deviation: 2.2 μm) was measured on a slide with a

light microscope with a computerized ruler.

For each of the 6 experiments, two sets of

7 collagen gels were prepared and measured. The

first set did not include any cells and the second set

included the same volume fraction of SK-MES cells

for a given experiment.

2.1 SK-MES Cell Culture

SK-MES cells (ECACC, UK) were cultured in

175 cm

-cell culture flasks and incubated at 37°C

and 5% CO

. Each culture vessel contained

complete culture medium composed of high-glucose

Dulbecco’s Modified Eagle’s Medium supplemented

by 10% volume of foetal calf serum and other

standard components according to the provider’s

instructions and previous studies (Yang, 2004).

2.2 Collagen Gels Preparation

Collagen type I gels were prepared according to the

supplier’s (BD Biosciences) instructions. To ensure

the viability of SK-MES cells reported in (Yang,

2004), the concentration of collagen was 1.5 mg/mL.

Cells were added and mixed with the gel at a

temperature of 4°C when collagen is in liquid form.

Gels were allowed to set by incubation at 37°C for 3

hours.

Each gel was prepared in a cylinder well and had

a diameter of 19 mm and a height of 3 mm. These

dimensions are justified in the next section.

Therefore each gel had a volume of 0.85 mL.

MICROWAVE DIELECTRIC SPECTROSCOPY OF LOW-VOLUME FRACTION HUMAN CANCER CELLS

EMBEDDED IN COLLAGEN GELS - Experimental Feasibility Study with an Open-ended Coaxial Probe

157

2.3 Experimental Set-up

2.3.1 Experimental Material

Dielectric spectroscopy was performed using a

vector network analyser (model 8753E, Agilent

Technologies) operated on the frequency range

200 MHz – 2 GHz connected to a flanged open-

ended coaxial probe (dielectric probe, Agilent

model 85070) via a coaxial cable.

The complex permittivity of the sample is

actually deduced by calculation from the

measurement of the complex reflection coefficient at

the tip of the probe. Indeed the reflection coefficient

is linked to the impedance or admittance seen at the

tip of the probe, which is itself directly related to the

complex permittivity of the sample by a suitable

model (implemented by Agilent’s software).

In theory, the model supposes that the sample is

semi-infinite (i.e. covers a half space) isotropic and

homogeneous. In practice, if the sample is not

homogeneous (our case), the result is an average

value weighted by the pattern of intensity of the

electric field (which is highest at the centre of the

probe tip). Besides, the sample is always of finite

size. In the probe supplier’s data sheet, the diameter

of the sample must be at least that of the probe, and

its minimal thickness is given by a simple formula

related to the permittivity, which in our case yields

about 2.5 mm. This is in good agreement with values

found in scientific papers to measure biological

tissues with probes of similar dimensions (Semenov,

2000; Hagl, 2003). Indeed, some researchers have

shown that measurement errors are small when the

sample thickness is at least as big as the outer

conductor radius of the probe (Fan, 1990; De

Langhe, 1994; Hoshina 2001), which is 1.5mm in

our case. We therefore decided to prepare 3mm-

thick collagen gels.

Nevertheless, a small sample size can be

problematic in the microwave range especially when

its permittivity is high and its losses relatively low,

because cavity resonances could occur. This

phenomenon explained by electromagnetic cavity

theory has been pointed out by some investigators

using the same kind of probe (Grant, 1989; De

Langhe, 1994; Sheen, 1999). To avoid potential

resonance effects, we chose 2 GHz as the upper

frequency (until which we did not observe any

resonance effect).

Besides, the frequency range of the probe using a

network analyser and the supplier’s software is

guaranteed from 200 MHz to 20 GHz. As a result,

we chose the frequency range 200 MHz – 2 GHz.

As shown in Figure 1, gel samples were actually

put onto the probe and fitted its dimension, as the

outer diameter of the probe flange is 19mm. Several

attempts were made to measure the gels from the top

with the probe upside down (compared to Figure 1),

but the repeatability of the measurement was very

poor. Moreover, the gels measured from the top got

squashed. On the contrary, putting the gel onto the

probe keeps its integrity and allows for a good

control of the contact between the gel and the probe,

and improves the repeatability of the measurement.

Figure 1: Diagram of the experimental set-up.

2.3.2 Experimental Method

For each experiment, 7 gels with cells, and 7 gels

without cells (‘controls’) were made and measured.

Each gel was measured 7 times in order to address

the repeatability issue. Prior to measurement, the

system was calibrated using a standard procedure, in

which the standards are air, a short circuit and

deionised water. The latter was measured in a

250 mL beaker to avoid resonance phenomena, in

accordance with (Blackham, 1997).

The measurement method was as follows. The

wells containing the gels were taken out of the

incubator and placed in a water bath at 37°C. Each

gel was then put onto the probe, which was at room

temperature. To achieve as good repeatability as

possible, a few minutes were necessary to get very

stable results. On the one hand, while the gels cool

down towards room temperature, we observed that

the real part of the permittivity ε’ increased slightly,

and the imaginary part ε’’ decreased slightly on the

whole chosen frequency range. On the other hand,

when left several tens of minutes at room

temperature, they start to dry out and we observed

that ε’ started to decrease and ε’’ started to increase

on the whole chosen frequency range. Measurements

were taken at the time when the two aforementioned

Network

analyser

Open-ended

coaxial probe

coaxial cable

stand

collagen gel sampleNetwork

analyser

Open-ended

coaxial probe

coaxial cable

stand

collagen gel sample

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

158

phenomena compensate each other. Thus, excellent

repeatability was achieved: at the very most, the

fractional error (standard deviation divided by the

mean of 7 measurements on the same gel) was

0.25% for ε’ and 1% ε’’. If measurements are taken

before stabilization, these figures become

respectively 0.6% and 2.5% at the very most.

3 RESULTS AND DISCUSSION

The measurement reproducibility was assessed for

each volume fraction tested by calculating the

fractional error (standard deviation divided by the

mean, both calculated on 7 gels measured 7 times,

i.e. on 49 measurements) as a function of frequency.

At the very most, it reached 0.9% for ε’ and 5% for

ε’’. The latter was higher in the lower part of the

explored frequency range, where the conductivity is

particularly high. However, it could reach 3.3%

around 2 GHz.

For all the cell volume fractions tested, a

significant difference in permittivity was observed

between composite gels (containing cells) compared

to gels alone. An example of result for a volume

fraction of 4.4% is given on Figure 2 in the complex

plane (Cole-Cole diagram).

200 MHz

2 GHz

200 MHz

2 GHz

Figure 2: Comparison composite gels / pure gels in the

complex plane (mean +/- standard deviation: data obtained

on 7 gel samples of each type); volume fraction: 4.4%.

The real part ε’ was found to be lower when the gels

contained cells rather than when they did not (cf.

Figure 3), and also to consistently decrease with the

cell volume fraction. A statistically significant

difference between composites and pure gels was

found for all volume fractions. This was

demonstrated by a two-tailed Student’s t-test (p-

value lower than 0.05 or even 0.01 on the major part

of the frequency range). Moreover, the difference in

real part between two adjacent volume fractions was

also found to be statistically significant by a two-

tailed t-test despite the close proximity of the

observed variations. It is in good agreement with

(Bagnaninchi, 2004) stating that a variation of 0.5%

volume fraction is detectable for low cell volume

fractions. The p-values were even lower for

differences in volume fractions greater than 1%.

Figure 3: Comparison composite gels / pure gels. Real part

of permittivity (mean +/- standard deviation: data obtained

on 7 gel samples of each type); volume fraction: 4.4%.

Figure 4: Comparison composite gels / pure gels for the

imaginary part of permittivity (mean +/- standard

deviation: data obtained on 7 gel samples of each type);

volume fraction: 4.4%.

Regarding the imaginary part of the permittivity, a

group of 3 experiments (out of 6) showed an

increase in ε’’ when comparing composites and pure

gels, and the 3 others showed a decrease (when

considering the means). A two-tailed t-test proved

that 2 variations among them were not statistically

MICROWAVE DIELECTRIC SPECTROSCOPY OF LOW-VOLUME FRACTION HUMAN CANCER CELLS

EMBEDDED IN COLLAGEN GELS - Experimental Feasibility Study with an Open-ended Coaxial Probe

159

significant (one in each group; p>0.1 and 0.3), an

example of which is shown on Figure 4.

This result about the imaginary part ε’’ can be

commented on qualitatively as follows. This

suggests that ε’’ of the composite gels could actually

be of the same order of magnitude as ε’’ of pure gels

because the reproducibility fractional error on ε’’

(which could reach 5% in the lower frequencies) is

not negligible. As gels are quite conductive, ε’’ is a

sensitive parameter whose variability is not

negligible.

Quantitatively we did some modelling to explain

why ε’’ of the composites gels can either be a bit

lower or higher than ε’’ of pure gels. We used

effective medium approximations commonly utilised

with biological cells, such as Maxwell-Wagner or

Looyenga equations (Bordi, 2002; Asami, 2002). A

composite gel is considered as an effective medium

in which cells are inclusions. A basic single-shell

cell model, modelling the membrane and cytoplasm

by their respective permittivities and conductivities

was implemented. The latter were varied even

beyond their commonly accepted values: the

permittivity of the membrane and the cytoplasm

were respectively varied from 2 to 30 and from 30 to

70. The simulation easily shows that for small

volume fractions, the conductivity of the cytoplasm

is decisive: if it is lower (respectively greater) than

the one of pure gel, ε’’ is lower (respectively

greater) for composite than for pure gel. Thus,

owing to the variability of the cytoplasm properties,

ε’’ of composites gels can be a bit lower or higher

than ε’’ of pure gels. Hence, the conductivity of the

cytoplasm of the measured SK-MES cells could be

of the same order of magnitude as that of the

measured gels.

Another study (Bagnaninchi, 2003) carried out

on the same frequency range but with macrophages

put inside chitosan scaffolds filled with a similar

ionic culture medium (RPMI) showed different

results. The addition of cells induced an increase in

ε’ and a decrease in ε’’. However, the effective

dielectric behaviour depends on the particular

dielectric properties of each type of cells and of the

surrounding media. The former constitute the next

step of the study.

4 CONCLUSIONS - PROSPECTS

This study has proven the feasibility of detecting

small volume fractions of lung cancer cells

embedded in collagen gels by microwave dielectric

spectroscopy. The real part of the permittivity was

found to decrease with the presence of cells. The

imaginary part did not significantly show a

consistent variation.

The prospects of this study are mainly threefold.

Firstly, further suitable modelling should be

developed to try to retrieve cell signature and

properties. Secondly, similar experiments should be

carried out with the corresponding normal lung

epithelial cells and results compared to this study.

Thirdly, other DS experiments should also be tried

to analyse the response and effectivnesses of various

anti-cancer treatments such as chemotherapy drugs

on in vitro cell culture samples.

ACKNOWLEDGEMENTS

This work was partially supported by Maxime Hanss

Prize (BBSRC – Alliance Française).

REFERENCES

Asami K., 2002. “Characterization of heterogeneous

systems by dielectric spectroscopy”. Progr. in Polym.

Sci., 27.

Bagnaninchi P-O et al., 2003. “Towards On-Line

Monitoring of Cell Growth in Microporous Scaffolds:

Utilization and Interpretation of Complex Permittivity

Measurements”. Biotech. Bioeng, 84, 3.

Bagnaninchi P-O et al., 2004. “Complex Permittivity

Measurement as a New Noninvasive Tool for

Monitoring In Vitro Tissue Engineering and Cell

Signature Through the Detection of Cell Proliferation,

Differentiation, and Pretissue Formation”. IEEE

Trans. on Nanobioscience, 3, 4.

Bindu G. et al., 2006. “Active Microwave Imaging For

Breast Cancer Detection”. Prog. In Electromagnetics

Research, PIER 58.

Blackham D.V., Pollard R.D., 1997. “An Improved

Technique for Permittivity Measurements Using a

Coaxial Probe”. IEEE Trans. Instrum. Meas., 46, 5.

Bordi F. et al., 2002. “Dielectric spectroscopy of

erythrocyte cell suspensions. A comparison between

Looyenga and Maxwell-Wagner-Hanai effective

medium theory formulations”. J. Non-Cryst. Sol., 305.

Bulyshev A.E. et al., 2001. “Computational Modeling of

Three-Dimensional Microwave Tomography of Breast

Cancer”. IEEE Trans. Biomed. Eng.,48, 9.

Chaudhary S.S. et al., 1984. “Dielectric Properties of

Normal & Malignant Human Breast Tissues at

Radiowave & Microwave Frequencies”. Indian J.

Biochem. and Biophys., 21.

Chelidze T., 2002. “Dielectric spectroscopy of blood”. J.

Non-Cryst. Sol., 305.

BIODEVICES 2008 - International Conference on Biomedical Electronics and Devices

160

Claycomb J.R. et al., 2002. “Nonlinear Dielectric

Spectroscopy of Living Cell Suspensions”. 2nd joint

EMBS/BMES Conference, Houston, USA.

De Langhe P. et al., 1994. “Design Rules for an

Experimental Setup Using an Open-Ended Coaxial

Probe Based on Theoretical Modelling”. IEEE Trans,

Instrum. Meas., 43, 6.

Duncan L. et al., 2006. “Assessment of the dielectric

properties of drug sensitive and resistant leukaemic

cells before and after ion channel blockers using

dielectrophoresis”. NSTI-Nanotech 2006.

Ermolina I. et al., 2001. “Study of Normal and Malignant

White Blood Cells by Time Domain Dielectric

Spectroscopy”. IEEE Trans. Dielectrics and Electrical

Insulation, 8, 2.

Fan S. et al., 1990. “Static analysis of an Open-Ended

Coaxial Line Terminated by Layered Media”. IEEE

Trans, on Instrum. Meas., 39, 2.

Fear E.C., 2005. “Microwave Imaging of the Breast”.

Tech. Cancer Res. Treatm., 4, 1.

Foster K.R., Schwan H.P., 1996. “Dielectric properties of

tissues” in C. Polk, E. Postow: Handbook of

Biological Effects of Electromagnetic fields, CRC,

Boca Raton, FL.

Gabriel S. et al., 1996. “The dielectric properties of

biological tissues: I. Literature survey - II.

Measurements in the frequency range 10 Hz to

20 GHz - III. Parametric models for the dielectric

spectrum of tissues”. Phys. Med. Biol., 41.

Gascoyne P. R. C. et al., 1997. “Dielectrophoretic

separation of cancer cells from blood”. IEEE Trans. on

industry applic., 33, 3.

Grant J.P. et al., 1989. “A critical study of the open-ended

coaxial line sensor technique for RF and microwave

complex permittivity measurements”. J. Physics E:

Sci. Instrum., 22.

Hagl D.M. et al., 2003. “Sensing Volume of Open-Ended

Coaxial Probes for Dielectric Characterization of

Breast Tissue at Microwave Frequencies”. IEEE

Trans. Microw. Theo. Tech., 51, 4.

Hagness S.C. et al., 1998. “Two-dimensional FDTD

analysis of a pulsed microwave confocal system for

breast cancer detection: Fixed-focus and antenna array

sensors”. IEEE Trans. Biomed. Eng., vol. 28.

Hoshina et al., 2001. “A Numerical Study on the

Measurement Region of an Open-Ended Coaxial

Probe Used for Complex Permittivity Measurement”.

IEEE Trans. Magnetics, 37, 5.

Hübner Y. et al., 2005. “Parallel measurements of drug

actions on Erythrocytes by dielectrophoresis, using a

three-dimensional electrode design”. IEE Proc.

Nanobiotechnol., 152, 4.

Jaspard F. et al., 2003. “Dielectric properties of blood: an

investigation of haematocrit dependence”. Physiol.

Meas., 24.

Joines W.T. et al., 1994. “The measured electrical

properties of normal and malignant human tissues

from 50 to 900 MHz”.

Med. Phys., 21, 4.

Lisin R. et al, 1996. “Time domain dielectric spectroscopy

study of human cells. I. Erythrocytes and ghosts”.

Biochim. Biophys. Acta, vol. 1280.

Pietrucha K., Marzec E., 2005. “Dielectric properties of

the collagen–glycosaminoglycans scaffolds in the

temperature range of thermal decomposition”.

Biophys. Chemistry, 118.

Polevaya Y. et al., 1999. “Time domain dielectric

spectroscopy study of human cells. II. Normal and

malignant white blood cells”. Biochim. Biophys. Acta,

1419.

Santini M.T. et al., 1991. “Effects of lonidamine on the

membrane electrical properties of Ehrlich ascites

tumor cells”. FEBS, 291, 2.

Santini M.T. et al., 1995. “A dielectric Relaxation Study

on the Effects of the Antitumor Drugs Lonidamine and

Rhein on the Membrane Electrical Properties of

Ehrlich Ascites Tumor Cells”. Anticancer Res., 15.

Semenov S.Y. et al., 2000. “Microwave Spectroscopy of

Myocardial Ischemia and Infarction. I. Experimental

Study”. Ann. Biomed. Eng., 28.

Semenov S.Y. et al., 2003. “Microwave Tomography for

Detection/Imaging of Myocardial Infarction. I.

Excised Canine Hearts”. Ann. Biomed. Eng., 31.

Sha L. et al., 2002. “A review of dielectric properties of

normal and malignant breast tissue”. IEEE

SoutheastCon conference, Columbia SC, USA.

Shao W. et al., 2005. “UWB Microwave Imaging for

Breast Tumor Detection in Inhomogeneous Tissue”.

Conf. IEEE EMBS.

Sheen N.I., Woodhead I.M., 1999. “An Open-ended

Coaxial Probe for Broad-band Permittivity

Measurement of Agricultural Products”. J. Agricult.

Eng. Res., 74.

Smith S.R. et al., 1986. “Dielectric properties of VX-2

Carcinoma Versus Normal Liver Tissue”. IEEE Trans.

Biomed. Eng., 33, 5.

Surowiec A.J. et al., 1988. “Dielectric Properties of Breast

Carcinoma and the Surrounding Tissues”. IEEE Trans.

Biomed. Eng., 35, 4.

Tofighi M.R., Daryoush A.S., 2001. “Near field

microwave brain imaging”. Electron. Lett., 37, 13.

Treo E.F. et al., 2005. “Hematocrit Measurement by

Dielectric Spectroscopy”. IEEE Trans. Biomed. Eng.,

52, 1.

Yang Y. et al, 2004. “Monitoring of lung tumour cell

growth in artificial membranes”. Biosensors and

Bioelectronics, 20.

MICROWAVE DIELECTRIC SPECTROSCOPY OF LOW-VOLUME FRACTION HUMAN CANCER CELLS

EMBEDDED IN COLLAGEN GELS - Experimental Feasibility Study with an Open-ended Coaxial Probe

161