CN101779956A - Early tumor detection method - Google Patents

Abstract

The invention provides an early tumor detection method. The method combines traditional medical detection equipment to obtain tissue partition information of an interest region, distributes the range of electromagnetic property parameters to different tissue structure partitions by referring to the electromagnetic property parameter distribution database of all tissues of the human body, and then works out the electromagnetic property parameter distribution database of the tissue of the interest region through electromagnetic inverse problem solution, and effectively detect early tumors and the tissue lesions without clear boundary. The method combines the tissue partition information with the electromagnetic property parameter distribution range information of the tissue during solving electromagnetic inverse problem, thereby greatly reducing the computation. The electromagnetic inverse problem solving method of the invention which combines with the traditional medical detection equipment can effectively detect the early tumors and the tissue lesions without clear boundary.

Description

A kind of early tumor detection method

Technical field

The present invention relates to a kind of medical detection method that tumor exists of surveying, especially can survey infantile tumour and not have the medical detection method of obvious border lesion tissue.

Background technology

At present, main tumor detection method is to utilize existing equipments such as CT, nuclear magnetic resonance, NMR, B ultrasonic to survey.These medical detection equipment all are the energy waves of launching a definite form to human body, receive then from human body transmission or reflection and the signal that comes finally generates image by imaging algorithm.Although the imaging of medical technology has had significant progress and development, but in clinical use, still there is a lot of problems, mainly comprises: 1. survey infantile tumour and do not have the lesion tissue difficulty on obvious border big 2. not high low-density tissue and the big regional detection of discontinuity accuracy.

Except above-described traditional tumour detection method, recent two decades a kind of formation method that grows up based on the anti-parameter of tissue resistance, this method mainly is divided into two big classes: microwave imaging and electrical impedance imaging.These two kinds based on electromagnetic imaging technique process years of researches, no matter be imaging system or imaging algorithm, all obtained very big progress, but simultaneously these two kinds of magnetography technology all are the imagings that organism integral body or extensive area are carried out, for containing multilayer tissue and every layer of all very thin position, these present two kinds of imaging systems also are not enough to the resolution that provides enough, only can provide local on a large scale in the equivalent conductivity of biological tissue.And organism is the uneven texture of a complexity, and equivalent conductivity can not accurately reflect the change in electrical characteristics of the concrete tissue of organism.

Summary of the invention

In order to overcome the deficiency of existing tumor detection method, the invention provides a kind of early tumor detection method, this method can organize electromagnetic property parameter distribution data base that different organizational structure subregions is distributed electromagnetic property parameter value scope with reference to human body is various in conjunction with having the area-of-interest organizational structure partition information that medical detection equipment obtains now.Find the solution by the electromagnetism inverse problem then and obtain region of interest tract tissue electromagnetic property parameter distribution, effectively survey infantile tumour and do not have the lesion tissue on obvious border.

The technical solution adopted for the present invention to solve the technical problems is:

The first step, electromagnetism inverse problem constraints is obtained

Electromagnetism inverse problem constraints is obtained and is comprised: test the signal of telecommunication boundary value that obtains area-of-interest by the signal of telecommunication, utilize existing medical detection equipment to obtain the organizational structure partition information of area-of-interest.

In second step, the electromagnetism inverse problem is found the solution

With reference to the various electromagnetic property parameter distribution of organizing of human body, set the sectional electromagnetic property parameter range of each organizational structure according to the organizational structure subregion of area-of-interest.Signal of telecommunication boundary value with signal of telecommunication test gained is a condition, uses the improved Monte Carlo method of inversion to find the solution the electromagnetism inverse problem, and the electromagnetic property parameter of organizing of area-of-interest is carried out backstepping, and what obtain area-of-interest organizes electromagnetic property parameter distribution information.

In the 3rd step, judge lesion region

The zone that the area-of-interest electromagnetic property parameter distribution of trying to achieve according to the electromagnetism inverse problem is searched the electromagnetic property abnormal parameters.

Widely different (dielectric parameter difference reaches 1: 5 because normal structure and tumor tissues are in the electromagnetic property parameter of microwave radio frequency range, pcrmeability parameter difference reaches 1: 10), can obtain abundant pathological information by the electromagnetic property parameter distribution figure that obtains tested tissue.But the electromagnetism inverse problem is a kind of ill-conditioning problem, and is very responsive to the border disturbance, and amount of calculation greatly and what obtain is not unique solution.So organizational structure partition information of the area-of-interest that our consideration collects in conjunction with existing medical detection equipment such as CT, nuclear magnetic resonance, NMR and B ultrasonic when finding the solution the electromagnetism inverse problem, improve electromagnetism inverse problem algorithm, the various electromagnetic property parameter distribution range information of organizing that comprise in using-system structural division information and the area-of-interest are as priori conditions, making detected object no longer is one " black matrix ", and become one " grey body ", improved inverting efficient and accuracy to a great extent.

The present invention organizes partition information and organize electromagnetic property parameter distribution range information carrying out combining when the electromagnetism inverse problem is found the solution, and has reduced amount of calculation to a great extent, can form a high efficiency electromagnetism inverse problem derivation algorithm.The present invention can calculate larger problem under same hardware condition, for example can improve resolution to mesh refinement, embodies tiny organizational structure from the result.

The electromagnetism inverse problem method for solving that this and existing medical detection equipment combines can effectively detect infantile tumour and not have the lesion tissue on obvious border.For patient strives for the valuable previous tretament time.

Description of drawings

Fig. 1 is an early tumor detection method schematic flow sheet of the present invention;

Fig. 2 is acquisition methods one sketch map of signal of telecommunication boundary value;

Fig. 3 is acquisition methods two sketch maps of signal of telecommunication boundary value;

Fig. 4 is the schematic flow sheet that the second step electromagnetism inverse problem of the present invention is found the solution;

The specific embodiment

Below in conjunction with accompanying drawing embodiments of the present invention are elaborated.

Figure 1 shows that early tumor detection method schematic flow sheet of the present invention, comprise that electromagnetism inverse problem constraints is obtained, the electromagnetism inverse problem is found the solution and judge three steps of lesion region.The first step wherein, electromagnetism inverse problem constraints is obtained and is comprised obtaining of signal of telecommunication boundary value and obtaining of organizational structure partition information.In second step, it is that the improved Monte Carlo of the utilization method of inversion is carried out backstepping to the electromagnetic property parameter distribution of organizing of area-of-interest that the electromagnetism inverse problem is found the solution.The 3rd step, judge lesion region, be the electromagnetic property parameter distribution of the area-of-interest of trying to achieve according to the electromagnetism inverse problem, the zone that utilizes existing medical knowledge manually to search the electromagnetic property abnormal parameters.Thereby reach the purpose of surveying infantile tumour and not having obvious border lesion tissue.

In the first step, the test of the utilization signal of telecommunication is surveyed area-of-interest, obtains the signal of telecommunication boundary value C=[c of area-of-interest ₁, c ₂..., c _n], can select Fig. 2 or any method shown in Figure 3 for use.The acquisition methods of signal of telecommunication boundary value shown in Figure 2 is by exciting electrode the biological tissue 3 of 1 pair of area-of-interest to be carried out signal of telecommunication excitation, obtains biological tissue's signal of telecommunication boundary values from being distributed in the borderline electrode battle array 2 of biological tissue.The acquisition methods of signal of telecommunication boundary value shown in Figure 4 is the biological tissue's 3 emission wideband pulse signal by 4 pairs of area-of-interests of transmitting antenna, and the scattered signal of receiving antenna array 5 receiving areas 6 promptly can be considered biological tissue's signal of telecommunication boundary value of area-of-interest.

In the first step, obtaining of organizational structure partition information, be that traditional medical detecting devices such as utilization CT, nuclear magnetic resonance, NMR and B ultrasonic are surveyed area-of-interest and imaging obtains gray-scale map, after the gray-scale map of gained carried out high-pass filtering, extract the HFS among the figure, obtain the structural division positional information of the various tissues that area-of-interest comprises, one of constraints of finding the solution as the electromagnetism inverse problem.

Fig. 4 is the schematic flow sheet that the present invention's described electromagnetism inverse problem of second step is found the solution, and specifically comprises the steps:

Step 1 produces the initial value of electromagnetic property parameter distribution at random

Area-of-interest is divided into some grids according to practical situation.

Utilize the organizational structure partition information of the area-of-interest that the first step obtains, organize electromagnetic property parameter distribution data base with reference to human body is various, according to the different characteristics of its electromagnetic property parameter value scope of different organizational structuries, distribute the initial value G of electromagnetic property parameter for each grid ₀=[g _0,1,, g _0,2..., g _{0, n}] ^T, wherein n is the number of dividing grid in the area-of-interest.If current electromagnetic property parameter distribution G _NOW=G ₀

Step 2 is calculated the fitness value under the current electromagnetic property parameter distribution

The fitness value of the current electromagnetic property parameter distribution of utilization numerical calculations.Fitness function is to propose according to the requirement of organism electromagnetic property parameter distribution and constraints, and these require and constraints comprises the pumping signal information that priori, organizational structure partition information and the utilization signal of telecommunication are tested.There is m if find the solution the requirement and the constraints of this problem, then have

$\left\{\begin{array}{c}{s}_{11}{\mathrm{<figure-callout\; id="1"\; label="g"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">g</figure-callout>}}_1+s_{12}{\mathrm{<figure-callout\; id="2"\; label="g"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">g</figure-callout>}}_2+\&CenterDot;\&CenterDot;\&CenterDot;+s_{1n}{g}_n=c_1\\ s_{21}{\mathrm{<figure-callout\; id="1"\; label="g"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">g</figure-callout>}}_1+s_{22}{\mathrm{<figure-callout\; id="2"\; label="g"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">g</figure-callout>}}_2+\&CenterDot;\&CenterDot;\&CenterDot;+s_{2n}{g}_n={\mathrm{<figure-callout\; id="2"\; label="c"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">c</figure-callout>}}_2\\ \&CenterDot;\&CenterDot;\&CenterDot;\\ s_{m1}{\mathrm{<figure-callout\; id="1"\; label="g"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">g</figure-callout>}}_1+s_{m2}{\mathrm{<figure-callout\; id="2"\; label="g"\; filenames="HSA00000020231200012.png"\; state="\{\{state\}\}">g</figure-callout>}}_2+\&CenterDot;\&CenterDot;\&CenterDot;+s_{\mathrm{mn}}{g}_n=c_m\end{array}\right.$

Wherein

Be known constraint matrix, G=[g ₁, g ₂..., g _n] be the true electromagnetic property parameter vector of area-of-interest, C=[c ₁, c ₂..., c _m] ^TBe the signal of telecommunication boundary value of the area-of-interest that obtains, electrode battle array 2 promptly shown in Figure 2 obtains the scattered signal that biological tissue's signal of telecommunication boundary values or receiving antenna array 5 shown in Figure 3 obtain.At problem involved in the present invention, m＜n is often arranged.This fitness function equation group is write as matrix form, be expressed as

SG＝C

Calculate the fitness value ‖ Δ C ‖=‖ C-C under the current electromagnetic property parameter distribution _NOW‖, wherein C _NOW=SG _NOW

Step 3 judges whether current electromagnetic property parameter distribution meets the demands or iterations is full.If the fitness value of current electromagnetic property parameter distribution meets the demands or iterations is full, then withdraw from circulation, and output fitness value hour corresponding electromagnetic property parameter distribution G _BESTAs result of calculation; Otherwise, execution in step four.

Step 4, in the electromagnetic property parameter value scope of being distributed to area-of-interest in the electromagnetic property parameter of each several part adjust, to obtain the electromagnetic property parameter distribution G of next round iterative computation _NOW, return step 2.

Claims (2)

1. an early tumor detection method specifically comprises the steps:

The first step, electromagnetism inverse problem constraints is obtained

Test the signal of telecommunication boundary value that obtains area-of-interest by the signal of telecommunication, utilize existing medical detection equipment to obtain the organizational structure partition information of area-of-interest;

In second step, the electromagnetism inverse problem is found the solution

With reference to the various electromagnetic property parameter distribution of organizing of human body, set the sectional electromagnetic property parameter range of each organizational structure according to the organizational structure subregion of area-of-interest; Signal of telecommunication boundary value with signal of telecommunication test gained is a condition, uses the improved Monte Carlo method of inversion to find the solution the electromagnetism inverse problem, and what obtain area-of-interest organizes electromagnetic property parameter distribution information;

In the 3rd step, judge lesion region

The zone that the area-of-interest electromagnetic property parameter distribution of trying to achieve according to the electromagnetism inverse problem is searched the electromagnetic property abnormal parameters.

2. early tumor detection method according to claim 1 is characterized in that, described electromagnetism inverse problem was found the solution specifically and comprised the steps: second step

Step 1 produces the initial value of electromagnetic property parameter distribution at random

Area-of-interest is divided into some grids according to practical situation; The initial value G that distributes the electromagnetic property parameter for each grid ₀=[g _0,1,, g _0,2..., g _{0, n}] ^TIf current electromagnetic property parameter distribution G _NOW=G ₀

Step 2 is calculated the fitness value under the current electromagnetic property parameter distribution

Calculate the fitness value under the current electromagnetic property parameter distribution || Δ C||=||C-C _NOW||, C wherein _NOW=SG _NOW, S is known constraint matrix, C=[c ₁, c ₂..., c _m] ^TSignal of telecommunication boundary value for the area-of-interest that obtains;

Step 3 judges whether current electromagnetic property parameter distribution meets the demands or iterations is full

If the fitness value of current electromagnetic property parameter distribution meets the demands or iterations is full, then withdraw from circulation, and output fitness value hour corresponding electromagnetic property parameter distribution G _BESTAs result of calculation; Otherwise, execution in step four;

Step 4, in the electromagnetic property parameter value scope of being distributed to area-of-interest in the electromagnetic property parameter of each several part adjust, to obtain the electromagnetic property parameter distribution G of next round iterative computation _NOW, return step 2.

Images