VOLUMETRIC INDUCTION PHASE SHIFT DETECTION SYSTEM FOR DETERMINING TISSUE WATER CONTENT PROPERTIES

Abstract

A method and apparatus of determining the condition of a bulk tissue sample, by: positioning a bulk tissue sample between a pair of induction coils (or antennae); passing a spectrum of alternating current (or voltage) through a first of the induction coils (or antennae); measuring spectrum of alternating current (or voltage) produced in the second of the induction coils (or antennae); and comparing the phase shift between the spectrum of alternating currents (or voltages) in the first and second induction coils (or antennae), thereby determining the condition of the bulk tissue sample.

Claims

1. A method of determining the condition of a bulk tissue sample, comprising: positioning a bulk tissue sample between a pair of induction coils or antennae; passing a spectrum of alternating current or voltage through a first of the induction coils or antennae; measuring a spectrum of alternating current or voltage produced in the second of the induction coils or antennae; and comparing the phase shift between the spectrum of alternating currents or voltages in the first and second induction coils or antennae, thereby determining the condition of the bulk tissue sample.

2. The method of claim 1, wherein the first and second induction coils or antennae do not contact the bulk tissue sample.

3. The method of claim 1, wherein determining the condition of the bulk tissue sample comprises: detecting at least one condition from the group consisting of edema, ischemia, bleeding, dehydration, water accumulation in the bulk tissue sample, extravasation, and disease.

4. The method of claim 1, wherein the bulk tissue sample is selected from the group consisting of brain tissue, lung tissue, heart tissue, muscle tissue, skin tissue, kidney tissue, cornea tissue, liver tissue, abdomen tissue, head tissue, leg tissue, arm tissue, pelvis tissue, chest tissue or trunk tissue.

5. The method of claim 1, wherein the frequency of the spectrum of alternating current is between 10 kHz and 10 GHz.

6. The method of claim 1, wherein the frequency of the spectrum of alternating current is between 1 MHz and 10 GHz.

7. The method of claim 1, wherein determining the condition of the bulk tissue sample comprises detecting edema, ischemia, dehydration, extravasation, in the tissue sample, and wherein the spectrum of frequency of the alternating current is between 100 kHz to 10 GHz.

8. The method of claim 1, wherein determining the condition of the bulk tissue sample comprises detecting interperitoneal bleeding in the tissue sample, and wherein the spectrum of frequency of the alternating current is between 100 kHz to 10 GHz

9. A method of determining changes in the condition of a bulk tissue sample over time, comprising: positioning a bulk tissue sample between a pair of induction coils or antennae; passing a spectrum of alternating current or voltage through a first of the induction coils or antennae; measuring a spectrum of alternating current or voltage produced in the second of the induction coils or antennae; and comparing the phase shift between the spectrum of alternating currents or voltages in the first and second induction coils or antennae over time, thereby determining a change in the condition of the bulk tissue sample over time.

10. The method of claim 9, wherein the first and second induction coils or antennae do not contact the bulk tissue sample.

11. The method of claim 9, wherein determining the change in the condition of the bulk tissue sample over time comprises: detecting a change over time in at least one condition from the group consisting of edema, ischemia, bleeding, dehydration, water accumulation in the bulk tissue sample, extravasation, and disease.

12. The method of claim 9, wherein the bulk tissue sample is selected from the group consisting of brain tissue, lung tissue, heart tissue, muscle tissue, skin tissue, kidney tissue, cornea tissue, liver tissue, abdomen tissue, head tissue, leg tissue, arm tissue, pelvis tissue, chest tissue or trunk tissue.

13. An apparatus for determining the condition of a bulk tissue sample, comprising: a first induction coil or antenna; a second induction coil or antenna; an alternating current power supply connected to the first induction coil or antenna, the alternating current power supply configured to generate a spectrum of currents or voltages in the first induction coil or antenna; and a measurement system connected to the second induction coil or antenna, wherein the measurement system is configured to measure a phase shift difference in the spectrum of currents or voltages between the first and second induction coils or antennae when the first and second induction coils or antennae are positioned on opposite sides of a tissue sample.

14. The apparatus of claim 13, further comprising: a system to compare the phase shift between the alternating currents or voltages in the first and second induction coils or antennae to determine the condition of the bulk tissue sample.

15. The apparatus of claim 13, wherein the alternating current power supply produces a spectrum of alternating currents with a frequency between 10 kHz and 10 GHz.

16. The apparatus of claim 13, wherein the alternating current power supply produces a spectrum of alternating currents with a frequency between 1 MHz and 10 GHz.

17. The apparatus of claim 14, wherein the system to determine the condition of the bulk tissue sample comprises: a system configured to detect at least one of edema, ischemia, bleeding, dehydration, water accumulation in the bulk tissue sample, extravasation, and disease by analysis of the phase shift difference in the currents between the pair of induction coils or antennae.

18. The apparatus of claim 13, wherein the alternating current power supply comprises: a function generator configured to generate an alternating current in the first induction coil or antenna having a frequency that changes in pre-programmed steps.

19. The apparatus of claim 18, wherein the function generator supplies an excitation signal of approximately 20 mA in the range of 1 to 8.5 MHz at pre-programmed steps.

20. The apparatus of claim 13, further comprising: a first differential receiving amplifier connected to the first induction coil or antenna; and a second differential receiving amplifier connected to the second induction coil or antenna.

21. The apparatus of claim 13, further comprising: a dual-channel demodulator connected to the first and second induction coils or antennae; and an analog-digital converter connected to the dual-channel demodulator.

22. The apparatus of claim 21, wherein the dual-channel demodulator comprises: a mixer; and a narrow band pass filter.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0040] FIG. 1 is a simplified schematic of the present invention, showing a tissue sample positioned between a pair of induction coils.

[0041] FIGS. 2A, 2B and 2C show calculated bulk electrical parameters as a function of frequency for various ratios of normal tissue to edema.

[0042] FIG. 3 is a schematic of a device for electro-magnetic field generation.

[0043] FIG. 4 is a multi-frequency inductive phase shift detection spectrometer in accordance with the present invention.

[0044] FIG. 5 is a calculated graph of phase shift vs. frequency in a simulation of intraperitoneal bleeding.

[0045] FIG. 6 is a magnified view of FIG. 5 in the frequency region from 1 to 9 MHz.

[0046] FIG. 7 is experimentally measured homogenized absolute values of the inductive phase shift as a function of frequency for various tissue volumes of saline injected into a rat abdomen. The percentage indicates the amount of saline injected relative to the total body weight of the rat. The rats used in this experiment were about 250 to 300 grin weight.

[0047] FIG. 8 is a calculated phase shift in brain tissue with various degrees of edema.

[0048] FIGS. 9A and 9B are an experimentally measured phase shift in brain tissue in vitro with various degrees of edema.

DETAILED DESCRIPTION OF THE DRAWINGS

(a) Theory:

[0049] As stated above, the present system assesses tissue condition by determining volumetric bulk properties of the water content of the tissue. Specifically, the sensing of an induction phase shift is used to assess tissue condition.

[0050] Biological tissues contain compounds with measurable electrical properties such as the intracellular and extracellular ionic solutions, the capacitative cell membrane, charged macromolecules and polar water. The combination of these compounds in terms of composition and structure affect the overall electromagnetic properties of the tissue. The particular effect of each one of these components and combinations is affected in a different way as a function of the electromagnetic field excitation frequency and magnitude applied on the tissue. Typical spectroscopic behavior of certain tissues is illustrated in the 100% blood, brain, muscle and lung curves in FIG. 2 (the other curves in the figure will be discussed later). The overall effect of excitation frequency on the electromagnetic properties is well understood since the pioneering work of Schwan (Schwan, H. P. (1957). “Electrical properties of tissue and cell suspensions.” Adv. Biol. Med. Phys. 5: 147-209.) Reviews can be can be found in many texts on the topic: S. Grimnes and O. G. Martinsen, Bioimpedance and Bioelectricity Basics. San Diego, Calif.: Academic Press, 2000, pp. 87-124; K. R. Foster and H. P. Schwan, “Dielectric properties of tissues” in Biological Effects of Electromagnetic Fields. Boca Raton, Fla.: CRC Press, 1996, pp. 27-106. Briefly, body tissues contain intra and extracellular fluids that behave as electrical conductors and cell membranes that act as electrical capacitors. At DC and low frequencies electrical current passes mainly through the extracellular fluid; at higher frequencies, however, current penetrates both intra and extracellular fluids. Therefore, body fluids and electrolytes are responsible for electrical resistance and cell membranes for reactance. At MHz frequencies the impedance of proteins becomes important and at GHz frequencies the behavior of water is sampled. In particular, up to roughly about 100 MHz the behavior should be affected by Maxwell Wagner relaxation of membranes, proteins in water and bound water whereas above about 100 MHz it should depend on the relaxation of free water, ionic conductivity and bound water. The increase in conductivity in the tissue as a function of the GHz frequency arises from the rotational relaxation of the water dipoles in tissue. In this sense, at higher frequencies the observations depend on the net water content in a volume rather than the cellular structure.

[0051] The relation between water and tissue content and properties and complex electromagnetic properties occurs throughout the frequency domain. Accordingly, the present system can be used to detect changes in tissue water properties with phase shift in a broad range of frequencies from kHz to GHz. Since the measurement of interest is phase shift in the bulk of tissue or organs there are several considerations in choosing the appropriate frequency and apparatus involving the significance of what is measured and the dimensions that can be measured. Since the interest is in phase shift DC type of measurements are not of interest for this type of measurement. Theoretically it is anticipated that in the frequency range from about 100 kHz to 1 GHz, depending on the type of tissue, the phase shift measurement will be affected by both the relative distribution between the intracellular water and the extracellular water and the relative amount of water. Therefore this range would be useful in detecting such conditions as ischemia, edema, bleeding, extravasation, dehydration. The effects of change in water content between the intracellular and extracellular would be most pronounced between 1 MHz and 400 MHz, with the most sensitive measurements between 1 and 100 MHz. Phase shift measured at the higher frequency range from about 300 MHz to 10 GHz would be most sensitive to water content and would be sensitive to detect edema, bleeding and dehydration. In particular the range of 1 GHz to 10 MHz would be most sensitive to water content. However, in addition to the sensitivity of various tissues to frequencies, which are tissue and condition dependent, since the method involves measurement in the bulk of tissue organs and the body the penetration depth of the various electromagnetic wave frequencies is of importance. For instance the penetration depth of microwave frequency energy at between 1 GHz to 3 GHz is about 2 to 10 cm for soft tissue. Therefore the particular application of the measurement would relate the integration depth with the measurement frequencies.

(b) Simplified Representation of the Invention:

[0052] The present invention deals with a method and an apparatus for detecting phase shift due to tissue water properties in the bulk of tissue. The method and the apparatus require an emitter of electromagnetic waves and a receiver to be placed in relation to the bulk tissue to be analyzed. When working in the frequency domain electromagnetic coils are used for emitters and receivers in the frequency range of up to 400 MHz. Beyond about 400 MHz to several GHz range, antennae (such as microwave antennae) may be used. In the range of frequencies from about 300 MHz to 1 GHz, both coils and antennae could be used.

[0053] FIG. 1 is a simplified schematic of the present invention, as follows.

[0054] System 10 consists of two induction coils 12 and 14 with the organ or the part of the body (sample S) to be analyzed placed in a determined relation to the coils. It is to be understood that coils 12 and 14 can be replaced with antennae when operating at higher frequencies. Reference herein is made to coils, However, the present invention is not so limited as antennae can be used instead. Coil 12 is driven by an AC current power supply (not shown) while the current in the other coil 14 is produced by induction and measured. The properties of the material (sample S) between coils 12 and 14 determine the currents in the induction current coil 14. By comparing the voltages in first coil 12 to those in second coil 14, a measure of the bulk electrical properties of the tissue (sample S) there between can be made. In preferred embodiments, the present invention measures the difference in the phase between the two AC voltages, i.e. the “phase shift”. In alternate aspects of the invention, a spectrum of voltages is passed through coil 12, and a current phase shift is measured in coil 14.

(c) Mathematical Model:

[0055] Different mathematical models are used to describe the system of this invention. The various mathematical models depend on the frequency analyzed. In the frequency range of up to about 1 GHz, more rigorously up to 300 MHz a quasi-static assumption can be made. In this range a solution such as that by (Griffiths, Steward et. al., Magnetic Induction Tomography-A Measuring System for Biological Materials” Ann. NY Acad. Sci. 873:335-345) can be used for analyzing the phase shift due the water content properties of tissue. In frequencies above that range the wave propagation becomes important and the eddy currents are not constant and an analysis of the type described in the same book but on page 327 is applied.

[0056] (i) Theoretical Considerations:

[0057] The analysis here follows Griffiths and his colleagues (See: Griffiths, Steward et. al., Magnetic Induction Tomography-A Measuring System for Biological Materials” Ann. NY Acad. Sci. 873:335-345). We consider as a simple case study tissue sample S to be a circular disk of tissue of radius Rand thickness t, placed centrally and midway between a small excitation coil and a small sensing coil spaced at a distance 2a (See FIG. 1). The thickness t, was considered to be much less than 2a. A sinusoidal current, of angular frequency ω, flows in excitation coil 12 and induces a magnetic field B. The circular nonmagnetic tissue sample S has conductivity σ and relative permittivity ∈.sub.r (it is assumed that the skin depth is greater than, t, and therefore the attenuation produced by the disk is neglected).

[0058] Our bulk model of edema assumes that the edema is uniformly distributed in the tissue and that the occurrence of edema will cause the bulk electrical parameters of the combined tissue to change according to the formula:

[00001]$\begin{array}{cc}\begin{array}{c}{\sigma }_c\left(T,F\right)=\frac{.Math.\left({\sigma }_i.Math.T\right)+\left({\sigma }_l.Math.F\right).Math.}{100}\\ {\mathrm{.Math.}}_{r,1}\left(T,F\right)=\frac{\left[\left({\mathrm{.Math.}}_{r,3}.Math.T\right)+\left({\mathrm{.Math.}}_{r,5}.Math.F\right)\right]}{100}\end{array}& \left(1\right)\end{array}$

[0059] where the subscripts c, t, and f stand for the composite properties, the tissue properties and fluid properties, respectively. The symbols, T, and, F, give the percentage volume of the pure tissue or the pure fluid respectively. In accordance with experiments performed by the Applicants in testing present invention, the tissue and fluid data used were taken from Gabriel and Lau (See: “The Dielectric Properties of Biological Tissues: III. Parametric Models for the Dielectric Spectrum of Tissues” Phys. Med. Biol. 41:2271-2293, 1996) and Duck (See: “Physical Properties of Tissue”, London, Academic press, 1990, Chapter 6, 167 223.) In accordance with experiments performed by the Applicants in testing present invention, blood was considered as the edematous fluid for brain and muscle tissues; and human serum was considered for lung tissue.

[0060] FIGS. 2A, 2B and 2C show the bulk electrical parameters as a function of frequency for various ratios of normal tissue to edema calculated from Eqs. (1) and the data in Garbriel and Lau, and Duck, supra. FIG. 2A shows brain tissues, FIG. 2B shows lung tissues and FIG. 2C shows muscle tissues. As expected; three typical major dispersion regions are observed for all the ratios tissue/fluid. For the cases of brain and muscle tissues the bulk electrical conductivity of all the ratios tissue/fluids have similar values as the frequency approaches 1 GHz. It's evident that at high frequencies the electrical properties of brain, muscle and blood become essentially the same. This fact can be attributed to the y dispersion region where the dielectric properties of the tissue are dominated by the water content. In contrast; for the case of lung tissue the bulk electrical parameters of all the ratios tissue/fluids have different values in the whole bandwidth. This behavior can be explained by the different electrical properties of the human serum with respect to the lung tissue.

[0061] (ii) Phase Shift in Sensing Coil 14:

[0062] Considering the thin “tissue” disk model described above, a sinusoidal current of angular frequency ω, flows in excitation coil 12 and induces a magnetic field B in sensing coil 14. According to Griffiths, supra, the current induced in the “tissue” disk (sample S) placed between the excitation coil 12 and the sensing coil 14 causes a perturbation AB in the field of the sensing coil given by:

[00002]$\begin{array}{cc}\frac{\Delta .Math..Math.B}{B}=\left(\omega .Math..Math.{\mathrm{.Math.}}_0.Math.{\mathrm{.Math.}}_r-j.Math..Math.\sigma \right).Math.\left(\frac{{\mathrm{ta}}^3.Math.{\mathrm{\omega \mu }}_a}{2}\right).Math.\left\{\frac{1}{a^2}-\frac{a^2+2.Math.R^2}{{\left(a^2+R^2\right)}^2}\right\}.Math.& \left(2\right)\end{array}$

[0063] where ∈.sub.0 and μ.sub.0 are the permittivity and permeability of free space, respectively. The total magnetic field B+A B in sensing coil 14 is shifted relative to the primary magnetic field B by an angle θ. The magnetic field and its perturbation can be obtained from the voltages induced in the sensing coil, V.sub.i and ΔV.sub.i. ΔB/B can be defined in terms of the induced voltage in sensing coil 14, by:

[00003]$\begin{array}{cc}\frac{\Delta .Math..Math.B}{B}=\frac{\Delta .Math..Math.V_i}{V_i}& \left(3\right)\end{array}$

[0064] We define a constant k:

[00004]$\begin{array}{cc}k=\left(\frac{{\mathrm{ta}}^3.Math.{\mu }_o}{2}\right).Math.\left\{\frac{1}{a^2}-\frac{a^2+2.Math.R^2}{{\left(a^2+R^2\right)}^2}\right\}& \left(4\right)\end{array}$

[0065] Substituting (3) and (4) into (2), the phase of the total induced voltage θ(V.sub.ind) in sensing coil 14 with respect to the induced voltage by the primary magnetic field in coil 14 could be expressed as a function of frequency and electrical parameters in the “tissue” disk between the coils [16], by:

[00005]$\begin{array}{cc}\theta \left(V_{\mathrm{ind}}\right)=\mathrm{arc}.Math..Math.\mathrm{tg}\left(\frac{k.Math..Math.\omega .Math..Math.\sigma }{k.Math..Math.{\omega }^2.Math.{\mathrm{.Math.}}_0.Math.{\mathrm{.Math.}}_r+1}\right)& \left(5\right)\end{array}$

[0066] (ii) Phase Shift in Excitation Coil 12:

[0067] The magnetic field in the present invention can be generated in the device as shown in FIG. 3. Specifically, an oscillator supplies an excitation signal (V.sub.exc) through an output impedance, Z.sub.out. The reference voltage (V.sub.ref) measured in the excitation coil is given by expression (6) where Z.sub.L is the impedance of a coil composite made of the resistance R.sub.L and the inductance X.sub.L, in series.

[00006]$\begin{array}{cc}{V}_{\mathrm{ref}}=V_{\mathrm{exc}}\left(\frac{Z_L}{Z_{\mathrm{out}}+Z_L}\right)& \left(6\right)\end{array}$

[0068] According to Hart L. et al. (See: A noninvasive electromagnetic conductivity sensor for biomedical applications” IEEE Trans Biomed Eng 32(12): 1011-1022, 1988), the presence of a conductive sample (the “tissue” disk between the coils) causes a change in the impedance of the excitation coil given by ΔZ.sub.L=ΔR.sub.L+ΔX.sub.L, where: ΔR.sub.L is the increase in the coil resistance and ΔX.sub.L is the increase in the coil inductance. The expressions for ΔR.sub.L and ΔX.sub.L were derived in Hart et. al., supra as:

ΔR=32π.sup.3*10.sup.−14N.sup.2f.sup.2R′.sup.3I′Δσ  (7)

ΔX=64π.sup.4*10.sup.−14N.sup.2f.sup.2R′.sup.3I′∈.sub.0Δ∈.sub.r  (8)

[0069] where: f=ω/2π is the frequency of the excitation signal, N is the number of coil turns, R′ is the coil radius, ∈.sub.0 is the permittivity of free space, and ∈.sub.r and a are the relative permittivity and electrical conductivity of the “tissue” disk sample respectively. The term I′ is a positive definite constant determined for a specific geometry and several approximations are given in Hart et. al., supra. In this study substitutions of σ.sub.c.fwdarw.Δσ and ∈.sub.r,c.fwdarw.Δ∈.sub.r were made for the expressions (7) and (8) because changes in electrical conductivity and relative permittivity of the “tissue” sample are considered.

[0070] The phase of the reference voltage θ(V.sub.ref) with respect to the excitation signal in the presence of a “tissue” sample can be estimated from the following expression:

[00007]$\begin{array}{cc}\theta \left(V_{\mathrm{ref}}\right)=\mathrm{arc}.Math..Math.\mathrm{ig}\left[\mathrm{Im}\left[\frac{Z_2+\Delta .Math..Math.Z_L}{Z_{\mathrm{out}}+Z_1+\Delta .Math..Math.Z_{\mathrm{.Math.}}}\right]/\mathrm{Re}\left[\frac{Z_L+\Delta .Math..Math.Z_L}{Z_{\omega }+Z_{\mathrm{.Math.}}+\Delta .Math..Math.Z_L}\right]\right]& \left(9\right)\end{array}$

[0071] Later in this study, the analysis for estimation of phase shift with edema was performed by using tissue properties from the experimental data in Gabriel and Duck, supra, and from the solution of equation (5) and (9) with the bulk properties from equation (1). The total change in phase shift (Δθ) between the reference and induced voltages in the excitation and sensing coil respectively is given by the expression:

Δθ=θ(V.sub.ind)−θ(V.sub.ref)  (10)

(d) Experimental System:

[0072] FIG. 4 illustrates a multi-frequency inductive spectrometer system 10 as designed, constructed and operated by the Applicants. This system is preferably used for a frequency of up to 400 MHz. System 10 comprises four modules: function generator 20, transceiver 30, dual-channel demodulator 40 and analog digital converter 50. A personal computer 60 with a Pentium 2 GHz processor (model 4400, Dell Inc. Round Rock, Tex.) controls the system and processes the data.

[0073] Function generator 20 uses two identical programmable synthesizers 22 and 24 (NI 5401, National Instruments Inc, Austin, Tex.) as oscillators. Oscillator 22 supplies an excitation signal I cos(ω.sub.et) of approximately 20 mA in the range of 1 to 8 MHz at pre-programmed steps. A modulation signal I cos(ω.sub.mt) is generated by second oscillator 24. The difference ω.sub.e−ω.sub.m=ω.sub.o=100(2π) rad/sec is maintained constant in the whole bandwidth in order to produce a narrow band measured voltage signal on a constant low intermediate frequency for processing and demodulation, as proposed by Ristic, B. et al. (See: “Development of an impredance spectrometer for tissue monitoring: application of synchronous sampling principle” Proc 21st IEEE Annual Northeast Conference, 22-23 May 1995, pages 74-75).

[0074] The excitation and modulation signals are connected to transceiver 30 and dual-channel demodulator 40 modules respectively. Transceiver 30 consists of an excitation coil 12 and a sensing coil 14 coaxially centered at a distance d=10 cm and two differential receiver amplifiers 32 and 34 (AD8130, Analog Devices Inc. Norwood, Mass.). Both coils 12 and 14 were built with magnet wire AWG32 rolled on a cylindrical plastic former with radius r=2 cm, five turns. The coil inductance, as calculated on the basis of Faraday's law, is approximately 40 pH. The excitation coil 12 generates a primary oscillating magnetic field. The sensing coil 14 detects the primary magnetic field and its perturbation through a proximal conductive tissue sample S. To avoid inductive pickup the leads of the coils are twisted. The amplifiers 32 and 34 were connected as conventional operational amplifiers and collect the reference voltage (V.sub.ref) and the induced voltage (V.sub.ind) in the excitation 12 and sensing 14 coils respectively. The gain of amplifiers 32 and 34 was adjusted in order to obtain a dynamic range of ±5V throughout the whole bandwidth.

[0075] Dual-channel demodulator 40 uses a pair of mixers 42 and narrow band pass filters 44 to transfer the information of any excitation and sensing signal of a variable frequency to a constant low frequency (ω.sub.o). A multiplier (AD835, Analog Devices Inc. Norwood, Mass.) was used as mixer 42. Narrow band pass filter 44 was designed on the basis of operational amplifier 32 (AD844, Analog Devices Inc. Norwood, Mass.). This module used two identical channels for parallel demodulation.

[0076] To avoid additional inductance and stray capacitance in the circuit, amplifiers 32 and 34 and dual channel-demodulator circuits 40 were shielded by a metallic box and connected to coils 12 and 14 with short coaxial cables (length less than 0.8 m). The current passes through the shield to minimize any inductance mutual between the circuit and the coils.

[0077] Analog-digital conversion module 50 digitized the reference and induced voltage signals on the constant low frequency. A data acquisition card NI 6071E (National Instruments Inc, Austin, Tex.) with a sample rate of 1.25 MSamples/seg and a resolution of 12 bits was used as analog-digital converter 50. The phase of the reference and induced voltages were calculated in software over approximately five cycles by an extract single tone function available in LAB VIEW V 6.1 (National Instruments Inc, Austin, Tex.). This function was programmed to find the highest amplitude at 100(2π) rad/sec and return the phase. The phase shift between the reference and induced voltage was estimated as Δθ=θ(V.sub.ind)−θ(V.sub.ref). The ratio signal to noise (SNR) for phase shift measurement was improved by averaging over twenty spectra (39 dB at 1 MHz).

[0078] For use up to higher frequencies including GHz range the present apparatus may comprise of a source of electromagnetic energy such as an RF Signal generator (Agilent 8648D9 KHz-4 GHz. The source is connected to an emitter which is a single frequency commercial antenna for microwave or radiofrequency placed in relation to the analyzed tissue and another similar receiving antenna. The receiving antennae is connected to an amplifier such as (Low Noise Amplifier, Agilent 11909A, 9 KHz-1 GHz), or (Microwave system amplifier Agilent 83006A10 MHz-26.5 GHz). The signal and the phase shift can be detected with Agilent 4396B RF Network/Spectrum/Impedance Analyzer, 100 kHz to 1.8 GHz).

(e) Experimental Results—Detection of Interperitoneal Bleeding in Rats:

[0079] Intraperitoneal bleeding in the abdomen of a rat was simulated by infusion of various volumes of physiological saline into the abdominal cavity in rats. Specifically, experiments were performed identically on each of five rats. The experiments started with anesthetization of the animal via intraperitoneal injection of Nembutal solution (50 mg/ml sodium pentobarbital, Abbot Labs, North Chicago, Ill.) for a total of 100 mg sodium pentobarbital per kg of rat. To simulate intraperitoneal bleeding and accumulation of fluids in the abdomen we injected various volumes of physiological saline (0.9% w/w NaCl) into the abdominal cavity through a short intravenous catheter (Venflow). The catheter remained in place throughout the experiment. We injected increasing volumes of saline and the measurements were done for four volumes of: 1, 2.5, 5 and 7.5% of (weight of saline)/(weight of the tested rat). The physiological saline was maintained at approximately 36.5° C. prior to injection. In all the experiments the baseline reference measurement was for the experimental subject prior to the intraperitoneal physiological saline injection. In all the experiments the coils were placed around the abdomen of the rat in such a way that the abdominal cavity was centered between the excitation and sensing coils. The geometrical position was carefully maintained as similar as possible for all the subjects.

[0080] The Applicants studied the phase shift due to four different volumes of saline in the frequency range from 1 MHz to 8.5 MHz with an induction system for measuring bulk phase shift. As will be shown below, the test results show that inductive bulk measurements of phase shift are sensitive to the relative volume of saline at frequencies higher than approximately 1 MHz, which is qualitatively consistent with our theoretical predications. In addition, the phase shift detected increases as a function of frequency and the fluid volume also qualitatively consistent with the theoretical predictions. As such, the results indicate that bulk induction measurement of the phase shift has the potential for becoming a robust means for non-contact detection of intraperitoneal bleeding.

[0081] FIG. 5 is a calculated graph of phase shift vs. frequency. FIG. 5 was obtained from our theoretical calculation for gut tissue and shows the absolute homogenized values of the inductive phase shift as a function of frequency for various volumes of physiological saline into the abdominal cavity, simulating various degrees of bleeding. The results are shown in a homogenized form with respect to the values without saline. As can be seen, an increase in the volume of injected saline causes an increase in inductive phase shift. Specifically, the relative phase shift caused by internal bleeding begins at about 1 MHz. The phase shift difference due to internal bleeding has a characteristic inverse U type shape with a maximal at about 1000 MHz. The behavior shown in the figure beyond 1 GHz is highly approximate.

[0082] As can be seen, the values of the phase shift obtained from the analytical calculations provide an excellent qualitative indication of the effect of internal bleeding. Thus, the present system operates even if the bulk properties of the tissue in the abdomen are substantially different from the values that we used, since we did not consider fat, muscle, food and many other components in the abdomen. Nevertheless, the results suggest that the resolution of the measurement is greater for certain optimal values. In optional clinical applications, the phase shift can be scanned over a wide range of values to determine the best frequency for the highest signal to noise measurement.

[0083] The results in FIG. 5 indicate that the phase shift due to internal bleeding should be detectable from about 1 MHz. FIG. 6 is a magnified view of the phase shift in FIG. 5 in the frequency region from 1 to 9 MHz, which confirms this. Here it is important to notice that in this range the phase shift relative to baseline increases with an increase in measurement frequency and amount of simulated internal bleeding. We have chosen this range of frequencies for our experimental studies because it is the onset of the phenomenon of phase shift due to internal edema.

[0084] FIG. 7 shows the experimentally measured homogenized absolute values of the inductive phase shift as a function of frequency for various volumes of physiological saline into the abdominal cavity, simulating various degrees of bleeding. The results are shown in a homogenized form with respect to the control values, the baseline measurements. In this mode of presentation, the experimental subject without water produces zero phase shift at all frequencies and the injection of the physiological saline solution produces the departure from zero. Another advantage of presenting the results homogenized with respect to the control values is to overcome possible bias in the electronic circuitry. The frequency is given in a logarithmic scale from 1 to 8.5 MHz. The value of one standard error is also shown in the figure. The beating of the heart, breathing as well as abdominal motion might change the bulk electrical properties of the composite body under measurement. These factors may affect the magnitude and phase of the induced voltage at the sensing coil in the whole bandwidth. To remove natural artifacts due to physiological activity, 20 measurements were taken at each frequency. Averaging over these measurements has the effect of a robust filter to physiological activity artifacts. The figure shows that the change in phase shift due to simulated internal bleeding begins at about 1 MHz and increases with frequency and amount of internal bleeding. Qualitatively, the experimental results are very similar to the theoretical calculations.

[0085] In summary: these experimental results confirm that measuring the relative spectroscopic distribution of induction phase shift in the bulk of the abdominal cavity can be used for non-contact detection of intraperitoneal bleeding. Thus, in clinical practice, induction phase shift can be measured as a function of time and frequency in patients who are in danger of internal bleeding.

(f) Experimental Results—Detection of Brain Edema:

[0086] As will be shown herein, our results verified that bulk measurement of inductive phase shift can be used for non-contact detection of the content of water in brain, lung and muscle tissue. The analytical results of FIG. 2 showed an increase of the phase shift proportional to the water content starting at frequencies as low as 10 MHz. In addition, our results show that the phase shift changes with frequency, in a rather complex manner.

[0087] The results show that the phase shift is sensitive to the relative volume of edema at frequencies higher than approximately 10 MHz. The effect of edema on brain, lung and muscle tissues is tissue type specific. Increasing the volume of tissue has the effect of lowering the frequency at which the phase shift becomes sensitive to the volume of edema. The results indicate that bulk induction measurement of the phase shift has the potential for becoming a simple means for non-contact detection of formation of edema in brain, lung and muscle tissues.

[0088] As a first order model of edema in the brain, ex-vivo brain tissue of pig (used approximately 8 hours after the animal sacrifice) was processed through a mixer and combined with various volumes of a physiological saline solution (0.9% w/w NaCl) to form a homogeneous paste. In accordance with experiments performed by the Applicants in testing present invention, the brain conductivity data used in the relevant calculations were taken from Gabriel's experimental report for excised bovine brain tissue supra. The data for excised bovine brain tissue was obtained two hours postmortem and at body temperature. The fluid was taken as a solution of 0.9% w/w of NaCl, with a constant electrical conductivity σ=1.3 S/m and a relative permittivity ∈.sub.r=80.

[0089] In our preparation, the changes in the electrical properties of brain tissue with the increase in water content may be explained as the dilution of a mixture of water and the proteins present in dried tissue. In the analyzed frequency range the cellular membranes have low impedance, and the tissue may be treated as a suspension of proteins in water. The significance of this is that, in the brain, at frequencies at the high end of the beta dispersion and above our experimental model will be comparable to that in viable tissue. At frequencies at the low end of the beta dispersion a greater difference between live tissue and edematous fluid is seen, and therefore, in that range, the effect of edema is more pronounced than determined in our experiment. Therefore, the present experimental model could be considered to provide a lower limit in the sensitivity of edema detection with our techniques and we can anticipate that the detection will be even better in a living organism.

[0090] Three different volume ratios between the volume of brain tissue and of saline were evaluated: 10, 20 and 30%. The paste was placed in a cylindrical and nonmagnetic recipient made of Teflon with a radius R=7.5 cm and height t=8 cm. This volume was chosen because it is on the order of a typical adult brain volume. In addition we studied samples of pure brain tissue that was also homogenized and pure physiological saline.

[0091] A calculation of the penetration depth (δ) as a function of frequency for saline and brain tissue was done according to the expression:

δ=(2/ωμ.sub.0σ).sup.1/2

[0092] where μ.sub.0 is the permeability of free space. We used the electrical conductivity data reported in Gabriel, supra for excised bovine brain tissue. The data for excised bovine brain tissue was obtained two hours post-mortem and at body temperature. The fluid was taken as a solution of 0.9% w/w of NaCl, with a constant electrical conductivity σ=1.3 S/m and a relative permittivity ∈.sub.r=80. The result shows that at 10 MHz; the skin depth is around 14 and 30 cm for pure saline and brain respectively. These values are larger than the thickness of the sample (8 cm).

[0093] All the samples were geometrical and vertically centered between excitation coil 12 and sensing coil 14. The geometrical position was carefully maintained as similar as possible for all the samples.

[0094] For every sample, twenty spectra of phase shift were obtained in the range of from 1 to 8 MHz. The data were averaged over twenty separate measurements, for each frequency. To overcome the bias in the phase shift due to the system electronics, the data were homogenized with respect to the values for brain 100%. In this way the changes observed in phase will depend essentially only on the electrical properties of the sample. The measurements were made at the room temperature (approximately 24° C.).

[0095] FIGS. 8 and 9 show the difference between the calculated (FIG. 8) or measured (FIG. 9) phase shift in brain tissue with various degrees of edema and the calculated or measured phase shift in the case with 100% brain tissue, as a function of frequency. In this mode of presentation brain tissue without edema produces zero phase shift at all frequencies and the addition of saline produces the departure from zero. FIG. 8 shows the calculated inductive phase shift as a function of frequency for various ratios of normal brain tissue to physiological saline, simulating various degrees of edema. The calculations were made by solving Eq. (10) and inserting in Eq. (1) brain tissue properties taken from Gabriel, supra. The edematous fluid was taken as saline (NaCl, 0.9% w/w) with a constant electrical conductivity σ=1.3 S/m and a relative permittivity ∈.sub.r=80. FIG. 8 shows the calculated phase shift as a function of frequency in the range of from 1 MHz to 1000 MHz. The analytical study shows that the phase shift is changing with frequency in a U shaped form and appears to have a maximum at about 100 MHz. Evidently, the phase shift increases with saline content, at any frequency. FIG. 8 shows that the phase shift can be used to measure edema in a wide range of frequencies and that there may be some optimal values of frequency that produce the highest signal.

[0096] The outcome of our experiments is shown in FIG. 9A, which shows the phase shift of various compositions of brain tissue and saline as a function of frequency. Specifically, FIG. 9A shows the phase shift from three different volume ratios between the volumes of saline to brain tissue: 10, 20 and 30%. Data for 100% saline is also shown. FIG. 9B shows the part of the curve developed in FIG. 8 in the range of the experimental measurements, to 8 MHz. A comparison of FIGS. 9A and 9B shows that the experimental results are quantitatively and qualitatively similar to the theoretical predictions. The phase shift increases with frequency and with water content. The frequency is given in a logarithmic scale from 1 to 8 MHz. The value of one standard error is also shown in the figure. The error in our experimental apparatus increases with an increase in frequency for all the volume ratios of saline to brain tissue.

[0097] The data in FIGS. 8 and 9 are presented by homogenizing the measured phase shift with respect to the phase shift in the case with 100% brain tissue. Therefore, the difference between the calculated or measured phase shift in brain tissue with various degrees of edema and the calculated or measured phase shift in the case with 100% brain tissue is shown as a function of frequency.

[0098] In this mode of presentation, brain tissue without edema produces zero phase shift at all frequencies and the addition of saline produces the departure from zero. With this mode of presentation, the sensitivity of our method for detecting edema by measuring bulk phase shift becomes clear, as does the effect of measurement frequency. Another advantage of presenting the results homogenized with respect to 100% brain tissue phase shift is to overcome possible bias in the electronic circuitry. Furthermore, this homogenized mode of presentation removes any systemic errors that could be caused by the electronics circuitry producing a bias in the experimentally measured phase shift.

[0099] A further advantage of the present system is that in the analyzed frequency domain phase shift is a measure that is strongly dependent on water content in relation to organic molecular cellular contents and not on cell structure.

[0100] It is clear from FIGS. 8 and 9 that phase shift due to changes in water content is substantial and detectable. In FIGS. 8 and 9, the departure from zero is the indication for change in water content. The change in phase shift increases with frequency and with water content. The experimental results suggest that the capability of the measurement system to detect water content improves at high frequencies. For example, the phase shift value detected at 8 MHz is clearly larger for all the tested samples.

[0101] Our results also show that measurable differences in phase shift are noted between 3 MHz to 4 MHz with higher volumes of saline producing measurable phase differences at lower frequencies. Our results demonstrate that valuable information for detection of phase shift with edema can be obtained at frequencies that are three orders of magnitude lower than the microwave frequencies and in a broad range of frequencies. The curve of saline alone provides the upper limit of the expected phase shift measurement relative to pure brain tissue.

[0102] The present invention can thus be used to continuously monitor phase shift to detect worsening conditions of increase in edematous fluid in the brain. Specifically, continuously measuring the relative changes in phase shift with time would produce curves as shown in FIG. 9 in the case of formation of edema. Thus, detecting changes in phase shift could point to worsening conditions of the patient.

[0103] Extending the present study over a wider range of frequencies may also hold much information since our analytical studies predicts a non-linear behavior throughout the range from MHz to GHz.

[0104] In summary, the results of this theoretical and in vitro study provide substantive preliminary information, which suggests that measuring the relative spectroscopic distribution of induction phase shift can produce a robust means for noncontact detection of occurrence of edema in the brain.