EP1351606A2

EP1351606A2 - Method of calibration of a radiology apparatus and radiology apparatus

Info

Publication number: EP1351606A2
Application number: EP01994333A
Authority: EP
Inventors: Remy Klausz; Lionel Desponds
Original assignee: GE Medical Systems Global Technology Co LLC
Current assignee: GE Medical Systems Global Technology Co LLC
Priority date: 2001-01-09
Filing date: 2001-12-27
Publication date: 2003-10-15
Also published as: FR2819368B1; WO2002058558A3; WO2002058558A2; JP2004528063A; FR2819368A1

Abstract

Method of calibration of a radiology apparatus of the type comprising a source of radiation of the X-ray beam, the X-ray beam being manifested by a flow of photons, that radiation consisting of X photons, a means of detection of the X-ray beam after it has crossed the organ to be visualized, image acquisition and processing means capable of supplying an image signal I and a means of visualization connected to the means of detection. The mean m and the variance σ2 of the image signal are calculated, the mean m and the variance σ2 being respectively equal to kN and to k2N by reason of the Poisson distribution, N being the number of photons on input of the detection device; the value of N and the value of a coefficient f are derived, f being the apparent gain of the radiology apparatus with B = fxN, B being a variable representing the intensity of the image; one then derives an index of X-ray dose having to be emitted by the radiation source in order to obtain on output an image satisfying predetermined quality thresholds.

Description

METHOD OF CALIBRATION OF A RADIOLOGY APPARATUS AND RADIOLOGY APPARATUS

CROSS REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of a priority under 35 USC 1 19 to French Patent Application No. 01 00231 filed January 9, 2001. the entire contents of which are incorporated by reference.

BACKGROUND OF THE INVENTION

The present invention concerns the field of calibration of radiology apparatuses for medical or other applications.

For all the radiological imaging systems involving a variable gain or a device for automatic adjustment of exposure, it is desirable that the intensity of the signal I constituting the image corresponds to a desired quantity of X-rays, so as to obtain satisfactory images in terms of quantum noise.

A system operating point is generally defined by simultaneous knowledge of the intensity of the signal constituting the signal, or that of a signal representing the latter, and of the image receiver input dose, as measured by a dosimeter under specified radiological conditions. The ratio of the intensity of the signal to the dose under those specified radiological conditions is called system conversion factor. The value measured by the dosimeter can be directly used for adjustments, or else supplied to an arithmetical unit in order to determine by a calibration the appropriate conversion factor or factors.

The dose is usually determined under flat field conditions, that is, by inserting in the X-ray beam a uniform part consisting of one or more materials, such as water, a plastic or copper, etc., of sufficient thickness to simulate the attenuation of the X-rays by a patient's body. BRIEF DESCRIPTION OF THE INVENTION

The present invention proposes a method of calibration of a radiology apparaius not requiring the presence of a dosimeter.

The method of calibration, according to one aspect of the invention, is a radiology apparatus of type comprising a radiation source providing X photons, a for deiection of the X-radiation after it has crossed an object, means for images of which are sought, image acquisition and processing means capable of supplying, on the one hand, an image signal I and. on the other, a signal B representing the intensity of that image, and a means for display of the image connected to the detection means. During the calibration phase, an image is acquired under specified conditions, the mean m and the variance σ- of distribution of the values 1 of the points constituting the image are calculated inside a defined zone of the image, the mean m and the variance σ being respectively equal to kN and to k N from the property of the quantum noise of the X photons of following the Poisson distribution, N being the number of photons actually contributing to formation of the image, and k being an unknown gain factor specific to the apparatus it is sought to calibrate, from which one derives the value of the coefficient k, the value of N under the conditions of calibration and the value of a factor f. with B = fxN, and one then derives an index of X-ray dose having to be emitted by the radiation source in order to obtain on output an image satisfying predetermined quality thresholds.

In an embodiment of the invention, the intensity B is obtained by processing of all or part of the image signal.

In an embodiment ofthe invention, the intensity B is equal to the mean of the points ofthe image in a predefined zone.

In an embodiment of the invention, the means of calculation of intensity B have a different slate between the calibration phase and the use phase. In an embodiment of the invention, ihe means of determination of the intensity B of the image include an optoelectronic detector receiving a light signal shunted from the optical path of the image.

In an embodiment of the invention, the means of determination of image intensity B include a dedicated detector, of optoelectronic type, for example, placed after conversion of the X-ray radiant image into an optical image and receiving a light signal shunted from the optical path of the image.

In an embodiment of the invention, the preceding stages are repeated in order to obtain the value of the factor f for different values of N.

In an embodiment of the invention, in the course of the calibration phase, the spectrum of the X-ray beam is centered on a given energy by diminishing the high voltage supply ofthe radiation source and by placing a filter significantly reducing the proportion of low-energy X-rays.

In an embodiment of the invention, the calibration operations are carried out for more than one X-ray energy band.

In an embodiment of the invention, an image subtraction is previously carried out in the course ofthe calibration phase.

In an embodiment of the invention, the calibration operations are carried out for different values of X-radiation intensity, with a view to identifying the contribution to measurement of the noise sources other than the quantum noise of the

X-rays, in order to eliminate that contribution.

In an embodiment of the invention, the stages are carried out automatically.

The invention is directed to a radiology apparatus of type comprising a radiation source, providing X photons, a means for detection of the X-radiation after it has crossed an object, images of which are sought, means for image acquisition and processing capable of supplying, on the one hand, an image signal I and, on the other, a signal B representing the intensity of that image, and a means for display of the image connected to the means for detection. The apparatus includes means for calibration capable of calculating the mean m and the variance σ of distribution of the values 1 of the points constituting the image inside a defined zone of the image. 9 the mean m and the variance σ being respectively equal to kN and to k N from the property of the quantum noise of the X photons of following the Poisson distribution, N being the number of photons actually contributing to formation ofthe image, and k being an unknown gain factor specific to the apparatus it is sought to calibrate, of deriving therefrom the value of the coefficient k, the value of N under the conditions of calibration and the value of a factor f, with B = fxN, and of then deriving therefrom an index of X-ray dose having to be emitted by a radiation source in order to obtain on output an image satisfying predetermined quality thresholds.

In an embodiment of the invention, the apparatus includes means for determination of image intensity B comprising a dedicated detector, of optoelectronic type, for example, placed after conversion of the X-ray radiant image into an optical image and receiving a light signal shunted from the optical path of the image.

The invention is also directed to a computer program comprising program code means for applying the stages of the above method, when the program is operating on a computer.

The invention is also directed to a support capable of being read by a device for reading program code means which is stored there and are suitable for application ofthe stages of the above method, when the program is operating on a computer.

BRIEF DESCRIPTION OF THE DRAWINGS

An embodiment, taken by way of nonlimitative example, is illustrated by the attached drawings, in which:

Figure 1 is a schematic view of a radiology apparatus of a first type;

Figure 2 is a schematic view of a radiology apparatus of a second type; Figure 3 describes a calibration operation according to the prior art.

DETAILED DESCRIPTION OF THE INVENTION

As can be seen in Figure 1. the radiology apparatus contains an X-ray tube 1 capable of emitting an X-ray beam 2 in the direction of an object 3 which includes an interesting detail 4 that it is desired to study, and an image receiver 5 placed in the X- ray beam 2 below the object 3. The image receiver 5 can comprise a solid state detector, hereinafter called flat panel, and as described in "Performance Characteristics of an Amorphous Silicon Flat Panel X-ray Imaging Detector" by Paul R. Granfors, in Proceedings of SPIE Volume: 3659, Medical Imaging 1999: Physics of Medical Imaging, (May 1999). The image receiver may be any other device capable of converting incident X photons into a signal usable in digital form. The signal representing the image is sent to imaging system by a link 6. In addition, the image signal is sent by a link 7 to a device 8 which extracts from it a quantity representative of the intensity of the image. The signal representing quantity B is sent by a link 9 to a control unit 1 1 which also receives an index signal 10, and determines a control signal sent by a link 12 to the high voltage generator 13 feeding the X-ray tube 1. The assembly thus constitutes a regulation loop to maintain signal B at a value equal to the index 10.

In Figure 2, the references of the elements similar to those of Figure 1 have been increased by 100. It can be seen that the image receiver comprises an image intensifier 124, the output image of which is sent back over an optoelectronic image detector 105 by means of a tandem of two optical devices 122 and 123, for example, lenses, separated by an iris diaphragm 121 . A luminance sensor 120 placed to receive a part of the light 1 18 resulting from conversion of the X photons by the image intensifier 124 sends a signal representing the intensity B of the image through a conductor 107 to a control unit 108 which also receives an index signal through a conductor 1 10. The control unit 1 1 1 supplies control instructions to the high voltage generator 1 13 feeding the X-ray tube 101. Figure 3 describes the method of calibration of the conversion factor ordinarily used. It is represented here in relation to equipment according to Figure 1. but could just as well be represented in relation to equipment according to Figure 2.

In the course of the calibration phase carried out according to a known method, one places in the X-ray beam 2 a particular object 30, usually described as a calibration phantom, presenting a roughly uniform attenuation of the X-rays from one point to another on its surface. Therefore, the image produced of that object will present a roughly uniform intensity, and the attenuation is roughly equivalent to that of a body, the image of which is sought, under conditions of normal use of the equipment. A dose detector 31 is placed in the input plane of the image receiver 5, connected to a measuring device 32. which supplies measurements of the dose on that plane.

The usual calibration phase consists of the following stages:

1. Placing the calibration phantom in the X-ray beam;

2. Setting the radiological parameters (high voltage, current, filtration) at values representative of use ofthe equipment;

3. Starting the emission of X-rays;

4. Simultaneously measuring the value of the dose supplied by the dosimeter 32 and the value of the intensity B of the corresponding image;

5. Calculating the calibration factor, equal to the ratio of intensity B to the value ofthe dose measured.

A difficulty with that method is the need to use a dosimeter, an apparatus which in turn requires periodic calibrations, and especially the fact that it is highly sensitive to variations of position of the measuring probe 31 by reason, on the one hand, of the variation of intensity of X-radiation with the square of the distance to the source and, on the other, the sensitivity to backscattered radiation by the image receiver 5 itself. The disclosed method makes it possible to remedy those two problems simultaneously.

For a given quality of X-rays, determined by the service voltage of the tube and by the filtering undergone by the X-ray beam 2, including at the same time the effect of internal filtralions at the source 1 and of the object studied 4, the image intensity is proportional to the quantity of N photons having contributed to the formation of that image. The same is true for the dose at the inlet of the image receiver. It is therefore possible to replace calibration of the operating point establishing the relation between the image intensity and the dose on input of the image receiver with a relation of the same type between the image intensity and the number of photons on .input of the image receiver, knowing that simultaneous application of the usual method and of the method disclosed under specified conditions makes possible a passage from one to the other.

A relation of type B = fxN is established, in which N is the number of X photons and f is a conversion factor. It is therefore sought to determine the value of the number of N photons. It is known that the distribution of values in a homogeneous zone of the image is affected by a random variation obeying a Poisson distribution.

The following operations are carried out:

1. Placing the calibration phantom in the X-ray beam:

2. Setting the radiological parameters (high voltage, current, filtration) at values representative of use ofthe equipment:

3. Starting the emission of X-rays;

4. Acquiring the resultant image and simultaneously measuring the value ofthe intensity B ofthe corresponding image;

5. If necessary, correcting the image acquired by the customary methods, that is, elimination or correction of erratic image points, displacement correction of the base level (offset) for the whole image and/or point by point (correction of nonuniformity of the black level), and local correction of gain in relation to a flat reference image;

6. Calculating in a zone of the image limited by an adequate contour, for example, of rectangular or circular shape, generally designated by the term "Region

Of Interest (ROI)." the mean m of the image signal and the variance σ² of that same signal;

7. If the noise variance generated by the electronic reading, amplification and analog-digital conversion circuits is not negligible relative to the noise measured (which can be obtained by carrying out those operations at a high enough signal level), correcting the variance value measured by subtracting from it the variance of that electronic noise term. After correction, the remaining noise is exclusively attributable to the statistics of the X photons and, therefore, according to the Poisson

9 9 distributiomm = k x N and σ = k x N. One therefore derives k = σ /m; starting from that point, knowing the particular values of and σ for which that calculation has been made, the value of N is derived at the level considered, whence one can finally derive the value of the conversion factor f by f = B/N.

Then, during use of the imaging system according to an embodiment of the invention, knowing the value of factor f, it is possible to derive the index of intensity to be applied on the input 10 ofthe regulator 1 1 in order to obtain the desired quantity of N photons in the image.

The means 8 making it possible to calculate the value of B from the values of the points ofthe image can furnish a simple arithmetical mean in a predefined zone of the image, adapted in size and shape to the particular needs of the application, or else have a more improved behavior, for example, with threshold effects on the highest or lowest levels, or nonlinear transfer functions. An example of those improved treatments can be found in EP 1 ,004,875.

Those properties will be effective in use of the system outside the calibration phases. On the other hand, during the calibration phase, their effect will be minimal, since the image is of constant level. If necessary, all of part of the complex functions of the extraction circuit of quantity B may be deactivated, retaining only the gain terms compatible with the calibration phase.

The reliability of the estimate of quantum noise and. therefore, of number N can. if necessary, be improved by replacing simple measurement of the mean and of the variance of the points of the ROI by an estimate of the parameters of the Poisson distribution (or in practice of the normal law), which is best adjusted to distribution of the points measured; the influence of possible aberrant points can thus be minimized.

If necessary, a relation can be established between those exposure indices expressed in numbers of photons under specified radiological conditions and the exposure indices expressed in doses, as is cusiomary. This can be done through a consecutive application ofthe two methods (with dosimeter and following the method disclosed) all at once on conception of the system or in the course of its construction; another method is to use the known properties of X-ray spectra.

The relation between number of photons and dose can be established as follows. Let us take, as an example, a beam generated under 80 kV and crossing 20 cm of tissues. According to the tables of HPA Scientific Report 30, the mean energy of the photons on output of the object is 54.4 keV and the quantity of photons produced at 0J5 m from the source is 2.27* 10⁴ photons/(mAs.mm²). Furthermore, the same tables indicate for that high voltage value an exposure of 0J μGy.mAs. A number of photons of 3.24* 10⁴ photons (μGy.mm ) is derived therefrom. The same result can be arrived at by other means. In Physics of Radiology (Johns &

1 1 9

Cunningham) that for photons of 50 keV, 2842* 10 photons/(R.m ) are counted, that is, 3.24*10" photons/(μGy.mm²).

In fact, these calculations establishing the relation between dose and number of photons are based on the mean energy of the photons present in the beam. Results closer to the conditions of the calculation will be obtained by restricting the energy range of the spectrum. For that purpose, for choice of the calibration conditions, a material of atomic number high enough to reduce the quantity of low-energy X phoions will advantageously be chosen for the phantom, while reducing the high- voltage value, in order to reduce the quantity of high-energy photons X contributing to the image. For example, a copper plate and a voltage of less than approximately 75 kV will be used. Such a beam can be generated, for example, by using 60 to 70 kV and 2-mm-thick copper. It is to be noted that if applications of the apparatus necessitate significantly different radiation conditions, such as low voltage/low filtration for the bone X-ray and high voltage/high filtration for the chest X-ray, it may be of interest to perform calibration according to the invention in each of the configurations, with possibly a specific calibration phantom, thus benefiting from the simplification of operations.

Another aspect should likewise be taken into account. The doses customarily indicated in adjustment procedures or the object of regulation are by necessity measured in a plane different from that of the image receiver (not attainable), which generally necessitates the application of corrective factors and before the mechanical protection elements ofthe image receiver, which attenuate the beam. According to an embodiment of the disclosed method, only photons actually reaching the plane of the image receiver are measured, which is preferable from the standpoint of image quality, but can raise problems from the standpoint of radiation protection. If the point of operation obtained in terms compatible with traditional procedures has to be specified, it will then be necessary to resort either to calculation of a correction factor or to perform a calibration proper to the equipment, but not needing to be remeasured on all the installations.

Finally, a last correction will have to be applied. In fact, the noise term measured can present increases due to other effects (Swank effect, Lubberts effect), or inversely be reduced by the fact that the transmission of spatial modulations in the image is not 100% for all frequencies. That behavior is described by the modulation transfer function (MTF). The apparent noise reduction factor associated with the MTF can be calculated independent of the latter or again determined by an initial calibration. The MTF can be measured by using the known methods, such as measurement of amplitude on a sinusoidal target, or calculation from measurement of the effective value of modulation in a rectangular periodic target image (Droege's method), or derived from the response to the image of an edge.

A detailed description of the theory and of the associated calculations will be found in "DQE(f) of an Amorphous Silicon Flat Panel X-ray Detector: Detector. Parameter Influences and Measurement Methodology" by Paul R. Granfors and

Richard Aufrichtig, Proceedings of SPIE. volume 3977, Medical Imaging 2000: Physics of Medical Imaging, (April 2000). It can be observed that measurement of the image noise serving for calculation of the number of photons and the MTF measurement can advantageously be carried out from a single image, for example, by using a test object and the associated methods described in U.S. 5,841 ,835.

For that specific point, the systems using a flat panel detector will be favored, since the MTF is essentially stable, resulting from the construction ofthe detector.

On the other hand, in the case of image intensifier systems, as described in

Figure 2, it will be necessary to make an MTF measurement before any dose calibration operation according to the method proposed, for the MTF can more easily vary with the variations of adjustment of the intensifier 124 or of focusing of the optical devices 121 and 123.

It will be deduced from this whether the initial conditions are still present or whether the MTF has changed, necessitating a readjustment or application of a correction factor. The phenomenon would be still worse with an image detector 105 of the scanning tube type, which is subject to a regular variation of spatial resolution with wear.

These calibration calculations may be made by one of the processors of the radiology apparatus, normally assigned to other tasks.

As a variant, those same operations can be provided for after a subtraction of images making it possible to retain only the terms which vary from one image to another and corresponding to the fluctuation from one measurement to another due to the amplification noise and to the quantum noise. The amplification noise can be eliminated by calculation, using, for example, two measurements made at appreciably different image levels, obtained by adjusting the supply current of the X-ray tube or the duration of exposure.

By applying the method and the corrections as described, an exact relation is thus established between the number of photons and the dose on input of the image receiver respectively.

As a result, calibration of the radiology apparatus can be carried out without a dedicated dosimeter by using the image receiver as photon counter. A simplification of the radiology apparatus is thus accomplished.

That method makes possible a simplified calibration, reducing the labor necessary to put the radiology apparatus into service, resulting in a cost saving.

The method of calibration is based on an absolute measurement and does not require the presence of external measuring instruments, which eliminates the need for external calibration of those measuring instruments. The method of calibration can be applied in any type of digital radiology apparatus. The digital X-ray detector is used as a relative photon counter during the operating phase in order to determine the appropriate calibration factor to be used during the operational use phases of the apparatus.

The quantum nature of the X-rays makes the noise of an image, as measured by the variance of levels of the points constituting that image, proportional to the X- ray signal, when the random variations of level in the image (also known and described as noise) are exclusively due to the quantum nature of the X photons

(quantum noise).

In case other quantum noise sources are likely to be present in the image, such as, for example, the electrons in a CCD type optoelectronic image detector, the respective statistics of the X photons and other particles will have to be differentiated by varying in a manner known a priori the gain between the two stages

(image intensifier, on the one hand, and optoelectronic image detector, on the other). for example, by varying the numerical aperture o the optics ensuring formation ofthe image gathered by the radiological image intensifier on the CCD detector.

The disclosed method is advantageous to the development of flat detectors, which render it much sturdier.

Various modifications in structure and/or steps and/or function may be made by one skilled in the art without departing from the scope and extent of the invention as recited in the claims.

Claims

WHAT IS CLAIMED IS:

1. A method of calibration of a radiology apparatus of type comprising a radiation source providing X photons;

means for detection of the X-radiation after it has crossed the object, images of which are sough;

means for image acquisition and processing capable of supplying, on the one hand, an image signal I and, on the other, a signal B representing the intensity of that image; and

means for display of the image connected to the detection means which method comprising the steps during the calibration phase of:

acquiring an image under specified conditions, the mean m and the variance σ- of distribution of the values I of the points constituting the image are calculated inside a defined zone ofthe image, the mean m and the variance σ being respectively equal to kN and to k N from the property of the quantum noise of the X photons of following the Poisson distribution, N being the number of photons actually contributing to formation of the image, and k being an unknown gain factor specific to the apparatus it is sought to calibrate, from which one derives the value of the coefficient k, the value of N under the conditions of calibration and the value of a factor f, with B = fxN; and

deriving an index of X-ray dose having to be emitted by the radiation source in order to obtain on output an image satisfying predetermined quality thresholds.

2. The method according to claim 1. in which the intensity B is obtained by treatment of all or part ofthe image signal.

3. The method according to claim 1 , in which the intensity B is equal to the mean ofthe points ofthe image in a predefined zone.

4. The method according to claim 1. in which the means of calculation of intensity B have a different state between the calibration phase and the use phase.

5. The method according to claim 1 , in which the preceding stages are repeated in order to obtain the value of factor f for different values of N.

6. The method according to any one of the foregoing claims, in which, in the course of the calibration phase, the spectrum of the X-ray beam is centered on a given energy by diminishing the high voltage supply of the radiation source and by placing a filter significantly reducing the proportion of low-energy X-rays.

7. The method according to any one of the foregoing claims, in which the calibration operations are carried out for more than one X-ray energy band.

8. The method according to any one of the foregoing claims, in which an image subtraction is previously carried out in the course of the calibration phase.

9. The method according to any one of the foregoing claims, in which the calibration operations are carried out for different values of X-radiation intensity to identify the contribution to measurement of the noise sources other than the quantum noise ofthe X-rays, in order to eliminate that contribution.

10. The method according to any one of the foregoing claims, in which the stages are carried out automatically.

1 1. Radiology apparatus comprising:

a radiation source (1 ), providing X photons;

means of detection (5) of the X-radiation after it has crossed an object (3), images of which are sought;

means for image acquisition and processing capable of supplying, on the one hand, an image signal 1 and, on the other, a signal B representing the intensity of that image; means for display of the image connected to the means for detection; and

means for calibration capable of calculating the mean m and the variance σ² of distribution of the values I of the points constituting the image inside a defined zone of the image, the mean m and the variance σ² being respectively equal to kN and to k²N from the property of the quantum noise of the X photons of following the Poisson distribution, N being the number of photons actually contributing to formation of the image, and k being an unknown gain factor specific to the apparatus it is sought to calibrate, of deriving therefrom the value of the coefficient k, the value of N under the conditions of calibration and the value of a factor f, with B = fxN, and of then deriving therefrom an index of X-ray dose having to be emitted by a radiation source in order to obtain on output an image satisfying predetermined quality thresholds.

12. The apparatus according to claim 1 1 , including means for determination of image intensity B comprising a dedicated detector placed after conversion of the X-ray radiant image into an optical image and receiving a light signal shunted from the optical path ofthe image.

13. The apparatus according to claim 12 wherein the detector is optoelectronic.

14. Computer program comprising program code means for applying the steps of the method according to any one of claims 1 to 10, when the program is operating on a computer.

15. Support capable of being read by a device for reading program code means which are stored there and are suitable for application of the steps of the method according to any one of claims 1 to 10. when the program is operating on a computer.