DE102020118102A1 - Method for detecting the magnetic response of a distribution of magnetizable particles to a magnetic excitation field - Google Patents

Abstract

A method is described for detecting the magnetic response of a distribution of magnetizable particles to a magnetic excitation field with a predetermined fundamental frequency, the magnetic response inducing a voltage signal with components of the harmonics of the fundamental frequency in at least one detection coil and a voltage signal induced by the magnetic excitation field in at least one compensation coil Compensation signal is generated and destructively superimposed with the voltage signal, so that a compensated voltage signal is tapped at an output of the coil arrangement. The technical solution described is characterized by a special supply of the compensated voltage signal, the isolation of the leading harmonics with the first transmission element, the suppression of the leading harmonics with the second transmission element, the optional supply of the remaining signal components to a further pair of transmission elements with a vorbesti mmten bandwidth range around the next higher harmonic and repeat the previous steps, digitizing the remaining signal, with each digitizing step in the respective transmission element being preceded by signal amplification in such a way that the previously known measuring dynamics of the digitizing device are exhausted, and the reconstruction of the magnetic response of the particle distribution from the digitized signals taking into account the separate signal amplifications and previously determined calibration data.

Description

The invention relates to a measurement method for the particle response of a distribution of magnetizable particles to a magnetic excitation. The invention relates in particular to the acquisition of measurement data in magnetic particle imaging (MPI) for functionalized particles previously administered to a patient in order to determine their distribution in the patient.

At the MPI, for example, super-paramagnetic iron oxide nanoparticles (SPION) are excited by a temporally periodic excitation magnetic field. "drive field", periodically magnetized. The magnetization of the particles depends non-linearly on the magnetic field strength and itself produces a time-dependent magnetic field as the particle response. By recording and analyzing the time behavior of the particle response in relation to the characteristics of the excitation field, conclusions can be drawn about the particle concentration.

To limit the MPI measurement to small volumes, the magnetic excitation field is combined with a selection magnetic field that is usually constant over time. "selection field", overlaid. The selection magnetic field has a zero at at least one predetermined point of the measurement volume. Starting from this so-called field-free point, in short: FFP, the selection magnetic field increases rapidly in all directions, so that magnetizable nanoparticles reach magnetic saturation even at a short distance from the FFP. Nanoparticles that are far away from the FFP hardly react to the excitation magnetic field and make no significant contribution to the particle response. Rather, the MPI measurement signal comes from the local environment of the FFP and provides information about the local particle concentration there. Alternatively, instead of a single FFP, a field-free line, FFL for short, can be generated with quadrupole magnet arrangements or with a suitable configuration of magnetic field-generating coils.

It is known from the prior art to determine the position of an FFP or an FFL within a region from which MPI measurement data are to be obtained - the so-called "field of view", FOV - by overlaying the static selection field with a time-varying homogeneous magnetic field to move. In this case, FFP or FFL are moved along temporally periodic paths that sweep over or run through the FOV as completely as possible.

For example, an FFP can be guided along a two- or three-dimensional Lissajous trajectory. The detected particle response is registered with the current position of the FFP or FFL to obtain an area or volume scan of the particle distribution inside the object or a living patient.

Detection coils are commonly used as magnetic field sensors to record the particle response, with the particle response and all existing interference fields being converted into voltage signals by induction. The disturbances are then treated at the level of the electrical signals, for example by signal filtering in Fourier space. It is previously known that the particle response predominantly has the fundamental frequency of the excitation magnetic field and its higher harmonics, so that these frequency components can be measured in a targeted manner.

The problem with this approach is the limited measuring dynamics of the necessary digitization when acquiring the analog output voltage of a detection coil. The coil receives large magnetic field components with the basic frequency without significant information about the particle distribution and at the same time very much smaller, information-carrying magnetic field components - the desired particle response - with the basic frequency and its higher harmonics, which are contained in the noise of the AD converter ("analog-to-digital converter “, ADC) disappear if this should not overdrive.

1. a. Abgreifen der induzierten Spannung am Ausgang der Detektionsspule und Unterdrücken der Grundfrequenz mit einem stark dämpfenden, passiven Hochpass- oder Bandstoppfilter.
2. b. Vorsehen einer Kompensationsspulenanordnung, die das Anregungsfeld - aber nicht das Partikelsignal - genauso wie die Detektionsspule(n) erfasst und beide derart auslöschend verschaltet sind, dass in der Superposition der induzierten Spannungen die Anteile des Sendesignals bei der Grundfrequenz möglichst kompensiert, zumindest aber sehr klein sind (vgl. 1).

The publication of Grass et al. "Analog receive signal processing for magnetic particle imaging", Medical Physics, Vol. 40, no. April 4, 2013 , addresses the problem of the detection of the excitation field by the detection coil and first names the previously known solution approaches:

1. a. Tapping off the induced voltage at the output of the detection coil and suppressing the fundamental frequency with a heavily damping, passive high-pass or band-stop filter.
2. b. Provision of a compensation coil arrangement which detects the excitation field - but not the particle signal - in the same way as the detection coil(s) and both are connected in such a canceling manner that in the superposition of the induced voltages the components of the transmission signal at the fundamental frequency are compensated as far as possible, but at least are very small (see. 1 ).

1. a. Die Grundfrequenz wird so stark gedämpft, dass der Anteil der Partikelantwort bei der Grundfrequenz im Rauschen verloren geht. Sie muss je nach Rekonstruktionsverfahren für die weitere Auswertung geschätzt werden. Die Signalenergie des Partikelsignals wird verringert und der Rauschabstand verschlechtert.
2. b. Die Signalstärke des Anregungsfeldes ist um etwa sechs Größenordnungen größer als die Partikelantwort. Es ist deshalb schwierig und aufwendig, eine Kompensationsspulenanordnung so zu realisieren, dass eine vollständige Auslöschung der Signalanteile des Anregungsfeldes erfolgt. In der Praxis wird dies oft nicht erreicht, sondern es bleibt im kompensierten Spannungssignal eine Grundfrequenzkomponente zurück, die zwar den Grundfrequenzanteil der Partikelantwort umfasst, aber das Anregungsfeld dominiert typischerweise immer noch das Signal um mehr als eine Größenordnung. Digitalisiert man dieses Signal unter Ausnutzung des vollen Dynamikumfangs eines ADC, dann lassen sich weniger höhere Harmonische erfassen als mit dem obigen Filteransatz. Dafür kann der Grundfrequenzanteil der Partikelantwort mittels einer Leermessung von dem Anregungssignalanteil getrennt werden.

Both approaches have serious disadvantages:

1. a. The fundamental frequency is attenuated so much that the part of the particle response at the fundamental frequency is lost in the noise. Depending on the reconstruction method, it must be estimated for further evaluation. The signal energy of the particle signal is reduced and the signal-to-noise ratio deteriorates.
2. b. The signal strength of the excitation field is about six orders of magnitude larger than the particle response. It is therefore difficult and expensive to implement a compensation coil arrangement in such a way that the signal components of the excitation field are completely eliminated. In practice, this is often not achieved, instead leaving a fundamental frequency component in the compensated voltage signal which, while comprising the fundamental frequency component of the particle response, typically still dominates the signal by more than an order of magnitude. If this signal is digitized using the full dynamic range of an ADC, fewer higher harmonics can be captured than with the above filter approach. For this purpose, the fundamental frequency component of the particle response can be separated from the excitation signal component by means of a blank measurement.

Grass et al. combine both approaches by tapping a compensated voltage signal and filtering it with a significantly weaker attenuation. The reduced damping means that the particle response at the fundamental frequency can also remain digitized and only has to be rescaled with the known transfer function of the filter in post-processing on the computer. At the same time, with the previously known measuring dynamics of the AD converter, more higher harmonics of the signal are recorded than would be possible with a more strongly damping filter without passive compensation.

The work of Schulz et al., "A FIELD CANCELATION SIGNAL EXTRACTION METHOD FOR MAGNETIC PARTICLE IMAGING", IEEE Trans Magn. 2015 February 1 ; 51(2 Pt 1) shows an MPI setup with specially designed compensation coils. In this work, a compensated voltage signal is tapped at the output of a serially connected pair consisting of a detection coil and a compensation coil, and amplified and digitized directly. What is remarkable is the provision of a second digitization channel, to which only the voltage across the compensation coil is fed as an input signal. It is used to determine the current that generates the excitation field and then to perform a calculation to extract the pure particle response in the compensated voltage signal. The increased measurement effort with two digitization channels is used by Schulz et al. but not the improvement of the number of measurable harmonics with an AD converter with limited dynamic range. Rather, it says there at the end of Section II B: "We assume that the analog-to-digital conversion (ADC) [...] has sufficient dynamic range to perform this correction step."

The patent specification DE 10 2018 204 311 B3 suggests, among other things, to optimize the full measurement dynamics of an AD converter at the MPI, to actively compensate the voltage signal at the output of the detection coil with the help of a signal generator that already applies a synthesized counter signal to the voltage tap. As a result of the destructive superposition, only a residual signal then remains—depending on the partly predetermined synthesis parameters and partly on control signals for controlling the signal generator determined in real time—which is to be fed to an amplifier and a digitizing device. The synthesis parameters include, in particular, how many of the first harmonics are synthesized simultaneously. The control signals alone may already enable rough MPI imaging, which can be refined if necessary by additional evaluation of the residual signal. However, active compensation is technically more complex and therefore more expensive than conventional filtering and amplification.

The invention therefore sets itself the task of proposing an improved use of the measurement dynamics of an AD converter for acquiring MPI-compatible measurement data using simpler means than in the prior art.

1. a. Zuführen des kompensierten Spannungssignals zu einem ersten Paar von Übertragungsgliedern mit einem vorbestimmten Bandbreitenbereich um die führende Harmonische umfassend ein erstes Übertragungsglied und ein zweites Übertragungsglied;
2. b. Isolieren der führenden Harmonischen mit dem ersten Übertragungsglied aufweisend eine Bandpass- oder Tiefpasscharakteristik und Digitalisieren des isolierten Signalanteils;
3. c. Unterdrücken der führenden Harmonischen mit dem zweiten Übertragungsglied aufweisend eine Bandstopp- oder Hochpasscharakteristik und Abgreifen der verbleibenden Signalanteile;
4. d. Optionales Zuführen der verbleibenden Signalanteile zu einem weiteren Paar von Übertragungsgliedern mit einem vorbestimmten Bandbreitenbereich um die nächsthöhere Harmonische und wiederholen der Schritte b und c;
5. e. Digitalisieren des verbleibenden Signals, wobei
6. f. jedem Digitalisierungsschritt im jeweiligen Übertragungsglied eine Signalverstärkung derart vorgeschaltet wird, dass die vorbekannte Messdynamik der Digitalisierungseinrichtung ausgeschöpft wird;
7. g. Rekonstruktion der magnetischen Antwort der Partikelverteilung aus den digitalisierten Signalen unter Berücksichtigung der separaten Signalverstärkungen und vorab ermittelter Kalibrierdaten.

The object is achieved by a method for detecting the magnetic response of a distribution of magnetizable particles to a magnetic excitation field with a predetermined fundamental frequency, wherein the magnetic response induces a voltage signal with components of the harmonics of the fundamental frequency in at least one detection coil and wherein in at least one compensation coil a magnetic excitation field induced compensation signal is generated and destructively superimposed with the voltage signal, so that a compensated voltage signal is tapped at an output of the coil arrangement, characterized by the steps

1. a. supplying the compensated voltage signal to a first pair of transmission links having a predetermined bandwidth range around the leading harmonic, comprising a first transmission link and a second transmission link;
2. b. isolating the leading harmonics with the first transmission link having a band-pass or low-pass characteristic and digitizing the isolated signal portion;
3. c. Suppression of the leading harmonics with the second transmission element having a bandstop or high-pass characteristic and tapping off the remaining signal components;
4. i.e. optionally supplying the remaining signal components to a further pair of transmission links with a predetermined bandwidth range around the next higher harmonic and repeating steps b and c;
5. e. digitizing the remaining signal, where
6. f. each digitization step in the respective transmission element is preceded by a signal amplification such that the previously known measurement dynamics of the digitization device are exhausted;
7. G. Reconstruction of the magnetic response of the particle distribution from the digitized signals, taking into account the separate signal amplifications and previously determined calibration data.

The dependent claims specify advantageous refinements of the method.

All technical devices that are suitable for transforming an electrical signal according to a frequency-dependent transfer function - here also: having a frequency characteristic - are referred to as a transmission element here and that are arranged in the signal path between the compensated receiving coil and the digitizing device. Various options are known to those skilled in the art for this purpose, such as passive oscillating circuits made up of coils and capacitors or active filter circuits based on operational amplifiers.

A pair of transmission elements is understood here to mean two transmission elements which are matched to one another and which transform the signal in the same predetermined bandwidth range around the same predetermined frequency.

The harmonic of the lowest order that has not yet been suppressed in the signal is always referred to here as the leading harmonic. In the case of multidimensional excitation with commensurable excitation frequencies, this also applies to the mixed frequencies.

According to the invention, the compensated voltage signal is tapped at the output of the coil arrangement and then fed to two or more different digitization channels. The signal is broken down according to frequency ranges by means of pairs of transmission elements, with the frequency components suppressed by one transmission element also being recorded and digitized. All parts of the particle response are recorded digitally. In each individual separation step, the leading harmonic is separated from the rest of the voltage signal in a defined bandwidth range using a pair of transmission elements and fed to separate amplification and digitization.

After isolating and suppressing the fundamental frequency in steps b and c, the 2nd harmonic is usually leading in the remaining signal components. If steps b and c are repeated with this same signal, the 2nd harmonic is also isolated and suppressed by a filter pair adapted to the frequency of the 2nd harmonic, and the 3rd harmonic then leads in the remaining signal components. In principle, this can be continued.

The main advantages of the method according to the invention include the fact that the addition of the portion of the particle signal to the fundamental frequency significantly improves the signal-to-noise ratio of the overall signal and that fewer assumptions have to be made for the reconstruction of the particle distribution. The pairs of transmission elements can also only be implemented with passive components and thus only add thermal noise to the reception chain. If the signal is tapped parallel to the receiving compensation coil unit with a unit transmission element, the filter for isolating the fundamental frequency can also be designed to be active, depending on the signal level. With each additional separate digitization channel of a leading harmonic, the remaining signal components can be amplified more, namely always in such a way that the previously known measuring dynamics of the digitization device are exhausted. This can be used to digitize a larger number of higher harmonics than was previously possible in the prior art.

• 1 Die Kompensationsmethode zur passiven Entkopplung von Sende- und Empfangskette. Das Sendefeld wird dabei von der Spule L_tx erzeugt und sowohl in der Empfangsspule L_rx als auch der Kompensationsspule L_c detektiert. Die Ausgangspannung u_e ist das durch auslöschende Verschaltung der Spulen L_rx und L_c kompensierte Spannungssignal enthaltend die Partikelantwort der SPIONs (Stand der Technik).
• 2 Ein aus zwei Übertragungsgliedern gebildetes Paar zur Erzeugung separater Signalpfade. Das kompensierte Spannungssignal U_e wird durch das Paar von Übertragungsgliedern aufgespalten.
• 3 Die bevorzugte Ausführungsform für ein Paar von Übertragungsgliedern. Das am Ausgang der Spulenanordnung abgegriffene, kompensierte Spannungssignal wird als Eingangsspannung u_e zugeführt.

Three figures serve to illustrate the invention. It shows:

• 1 The compensation method for passive decoupling of the transmission and reception chain. The transmission field is _{generated by the coil L tx} and detected both in the receiving coil L _rx and the compensation _{coil L c.} The output voltage u _e is the voltage signal compensated by the canceling connection of the coils L _rx and L _c and contains the particle response of the SPIONs (prior art).
• 2 A pair formed from two transmission links to create separate signal paths. The compensated voltage signal U _e is split by the pair of transmission elements.
• 3 The preferred embodiment for a pair of transfer links. The compensated voltage signal tapped off at the output of the coil arrangement is supplied as input voltage u _e .

In 2 shows schematically that the invention feeds the compensated voltage signal to at least two different signal paths simultaneously.

mit G_R,1 und G_R,2 als Übertragungsfunktionen zur Abgrenzung der zusammengeführten Signale.The voltage signal U _a,2 describes the signal transformed in the second transfer element with the transfer function G ₂ , in which the bandwidth range of the leading harmonics was suppressed by a band-stop or high-pass characteristic. The voltage signal U _a,1 in turn describes the signal transformed by the transfer function G ₁ in the first transmission element, in which the bandwidth range of the leading harmonics was changed by a band-pass or low-pass characteristic. In theory, the input signal U _e could be calculated from the respective output signals by inverting either G ₁ or G _{2 .} As already described, this is not possible due to the limited dynamic range of real AD converters. However, the reconstruction of U _e is still possible in the invention described here by introducing a combination of the signals that is limited to certain band ranges in the sense of $u_e=G_{R\mathrm{,1}}{\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="DE102020118102A1\_0004.png"\; state="\{\{state\}\}">G</figure-callout>}}_1^{-1}{u}_{a\mathrm{,1}}+G_{R\mathrm{,2}}{\mathrm{<figure-callout\; id="2"\; label="G"\; filenames="DE102020118102A1\_0003.png,DE102020118102A1\_0004.png"\; state="\{\{state\}\}">G</figure-callout>}}_2^{-1}{u}_{a\mathrm{,2}}$

with G _R,1 and G _R,2 as transfer functions to delimit the combined signals.

the 3 outlines a preferred embodiment of a pair of transmission elements, which is formed solely from passive components such as inductors and capacitors. In the conventional treatment of the voltage signal, eg according to Gräser et al., only the output voltage u _{a,2 is} fed to a low-noise amplifier (LNA) and then recorded digitally. Those skilled in the art will immediately recognize that u _{a,2 is} the band-stop filtered signal u _e , with a frequency band predetermined by the choice of L ₁ , C ₁ , L ₂ and C _{2 being suppressed.} Conventionally, the fundamental frequency - typically around 25 kHz with a certain frequency environment - is filtered out. However, the part of the signal that is removed is in 3 can also be _{tapped off as u a,1} and can also, if necessary, be fed to an amplifier and a digitizer. In fact, U _{a,1 is} the output voltage of a bandpass filter for u _e at the input. It can also be advantageous _{to isolate the signal U a,1} at least for the fundamental frequency with an active filter.

In addition to considering a band stop filter as a parallel circuit consisting of a low-pass filter and a high-pass filter, you can alternatively see it as a frequency-selective voltage divider in which a series connection of an LC parallel oscillating circuit and an LC series oscillating circuit has a voltage tap across each of the oscillating circuits. The same circuit thus generates two different output signals simultaneously from one input signal and in this respect forms two different, mutually corresponding filters - a filter pair.

The passive components L ₁ , C ₁ , L ₂ and C ₂ are preferably determined such that a filter pair for the Nth harmonic consists of LC resonant circuits each with the resonant frequency at N times the fundamental frequency, where N is a natural number .

The output signal u _a,2 can optionally follow the input of further pairs of transmission elements, for example further filters according to a circuit 3 , are supplied, with the next pair now being designed for the 2nd harmonic and its frequency environment, if necessary the subsequent pair for the 3rd harmonic and its frequency environment and so on. In each stage of this cascade, the respective leading harmonic is isolated and separately amplified and digitized - in each case as an output signal U _a,1 -, while the signal u _a,2 generated in each stage is either passed on to the next pair of transmission elements or finally amplified and as Whole thing is digitized.

In the practical implementation with passive filters as transmission elements, each stage of the filter _{cascade introduces additional attenuation into the signal u a,2} , so that a cascade with a few stages, for example up to four, is to be regarded as preferred. However, one step is advantageously sufficient to improve the signal-to-noise ratio (SNR) of the measured particle response.

As is known per se from the prior art, the proportion of the excitation field in the fundamental frequency can be estimated by means of a calibration measurement and later corrected in the computer. It is advantageous if the previously determined calibration data include at least the signal damping and phase angles of the output signals U _a,1 and u _a,2 digitized after filtering and are determined by means of a one-off calibration measurement without particles in the measurement volume depending on the excitation frequency - the fundamental frequency . The different amplifications of the individual digitization channels according to the invention are known to the user and must be taken into account accordingly in the numerical reconstruction of the particle signal.

The example circuit 3 is just a particularly simple variant for easier understanding. Other circuits, for example symmetrical circuits, in which both outputs of the coil arrangement are grounded sen, and circuits with active components, are also usable.

The central idea of the invention consists in an easy-to-implement splitting of the MPI coil output signal into at least two digitization channels, ordered according to the harmonics of the excitation frequency. At least a portion of the particle response is supplied to each of these channels. The individual channels can each have their own amplifications appropriate to the maximum amplitude of the particle signal components supplied to the channel—possibly isolated there. In this way, each digitization step can be preceded by a separate signal amplification in such a way that the previously known measurement dynamics of the digitization device are fully utilized. This avoids overloading of the AD converter and improves the SNR for all frequency components of the particle response.

QUOTES INCLUDED IN DESCRIPTION

This list of documents cited by the applicant was generated automatically and is included solely for the better information of the reader. The list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions.

Patent Literature Cited

• DE 102018204311 B3 [0012]

Non-patent Literature Cited

• Grass et al. "Analog receive signal processing for magnetic particle imaging", Medical Physics, Vol. 40, no. April 4, 2013 [0008]
• Schulz et al., "A FIELD CANCELATION SIGNAL EXTRACTION METHOD FOR MAGNETIC PARTICLE IMAGING", IEEE Trans Magn. 2015 February 1 [0011]

Claims (5)

Method for detecting the magnetic response of a distribution of magnetizable particles to a magnetic excitation field with a predetermined fundamental frequency, wherein the magnetic response induces a voltage signal with components of the harmonics of the fundamental frequency in at least one detection coil and wherein a compensation signal induced by the magnetic excitation field is generated in at least one compensation coil and is destructively superimposed with the voltage signal, so that a compensated voltage signal is tapped at an output of the coil arrangement, characterized by the steps a. supplying the compensated voltage signal to a first pair of transmission links having a predetermined bandwidth range around the leading harmonic, comprising a first transmission link and a second transmission link; b. isolating the leading harmonics with the first transmission link having a band-pass or low-pass characteristic and digitizing the isolated signal portion; c. Suppression of the leading harmonics with the second transmission element having a bandstop or high-pass characteristic and tapping off the remaining signal components; i.e. optionally supplying the remaining signal components to a further pair of transmission links with a predetermined bandwidth range around the next higher harmonic and repeating steps b and c; e. Digitizing the remaining signal, f. each digitization step being preceded by a signal amplification in the respective transmission element in such a way that the previously known measuring dynamics of the digitizing device are exhausted; G. Reconstruction of the magnetic response of the particle distribution from the digitized signals, taking into account the separate signal amplifications and previously determined calibration data. procedure after claim 1 , characterized in that the pairs of transmission elements each have a series connection of an LC parallel resonant circuit and an LC series resonant circuit with a voltage tap across each of the resonant circuits. procedure after claim 2 , characterized in that the pair of Nth harmonic transfer elements comprise resonant circuits having resonant frequencies at N times the fundamental frequency, where N is a natural number. procedure after claim 1 , characterized in that the first transmission link comprises an active filter and at least the fundamental frequency is actively isolated. Method according to one of the preceding claims, characterized in that the previously determined calibration data include at least the signal damping and phase angles of the digitized output signals and are determined by means of a one-time calibration measurement without particles in the measurement volume as a function of the excitation frequency.

Images