Apparatus and method for intraoperative real-time tumour tissue discrimination

Abstract

Apparatus for discriminating between tumour tissue and non-tumorous tissue, the apparatus comprising: a handheld bioimpedance probe having a plurality of electrodes mounted thereon, wherein the plurality of electrodes are greater than four in number; a current source configured to generate a stimulation current; a voltage sensor; one or more multiplexers coupled between the current source, the voltage sensor and the electrodes and configured to changeably select a subset of four electrodes from the plurality of electrodes, and, for each selected subset of four electrodes, to switch between a plurality of switching configurations, wherein, in each switching configuration, a first pair of the selected subset of four electrodes are connected to the current source, and a second pair of the selected subset of four electrodes are connected to the voltage sensor, the electrodes that constitute the first pair and the electrodes that constitute the second pair changing from one switching configuration to the next; and processor-controlled circuitry configured to: control the selection of the subset of four electrodes by the one or more multiplexers; control the switching configuration of the one or more multiplexers with respect to the selected subset of four electrodes; control the generation of the stimulation current by the current source and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that, in use, the stimulation current flows through tissue between the first pair of electrodes; control the measurement, by the voltage sensor, of the voltage between the second pair of electrodes of each switching configuration, across the tissue; determine, based on the applied current and the measured voltages for each switching configuration, a measure of relative impedance of the tissue for each switching configuration; and supply such impedance measurements as output data to a data analyser to enable real-time discrimination of the tissue to be performed, based on the impedance measurements.

Description

[0001] APPARATUS AND METHOD FOR INTRAOPERATIVE REAL-TIME TUMOUR TISSUE DISCRIMINATION

[0002] Field of the Invention

[0003] This invention relates to apparatus and associated methods for intraoperative realtime tumour tissue discrimination, for example to discriminate between tumour tissue and normal (non-tumorous) tissue, or to discriminate between other types of tissues. It is particularly applicable for the identification of brain tumours, although it is by no means limited to such an application, and may instead be used to discriminate between tumour tissue and normal tissue in other parts of the human (or potentially animal) body.

[0004] Background to the Invention

[0005] For effective diagnosis and treatment of cancers in the human (or potentially animal) body, the correct identification of tumours, and accurate discrimination between tumour tissue and normal (non-tumorous) tissue, is of great importance. Although the description that follows primarily relates to the determination of brain tumours during neurosurgery (i.e. neuro-oncology), the present disclosure is also applicable to the identification of tumour tissue in other parts of the human (or animal) body.

[0006] With particular regard to brain tumours, despite several advances in their management, tumours are either associated with high mortality and morbidity (as with gliomas), or present significant challenges due to disruption of normal physiology following surgery (as with pituitary tumours). Surgical resection is a frequently-used treatment for extrinsic and intrinsic brain tumours, with key aims being to include the most malignant tissue in the resected specimen, maximal resection, and avoiding or minimising removal of functional brain tissue.

[0007] Preservation of essential brain function combined with maximal, gentle and precise resection represents the best means of providing a surgical cure. In this regard, gross total resection (GTR) (>90% of macroscopically visible tumour) is directly associated with improved outcomes (progression free survival, overall survival) and quality of life. Further, supramaximal surgical resection (i.e. removal of >90% of tumour) has been linked with overall survival but must be balanced with preservation of essential cognitive and homeostatic function. Surgically acquired functional impairments are directly linked to poorer prognoses. Therefore, a key challenge in neuro-oncology surgery is preservation of normal functioning tissue. This can be challenging given then these tumours may not be clearly distinguishable on the basis of visual inspection alone during surgery.

[0008] Tumour identification can in part be achieved through the use of pre-operative neuroimaging, e.g. high-resolution MRI, which is integrated with on-table ‘neuronavigation’ - a type of intraoperative anatomical ‘GPS’. However, a limitation of these approaches is the change in anatomical topography during surgical resection as a result of tumour removal and brain ‘shift’. As such, techniques which provide ‘real-time’ visualisation during surgical resection are currently the focus of intense investigation. Current approaches include intraoperative ultrasound-guided surgery (USS), hyperspectral imaging, confocal microscopy, and Raman spectroscopy. Intraoperative USS is the only one of these approaches in formal clinical use, but remains user dependent, and is vulnerable to difficulties in interpretation. There is therefore an outstanding need in the field of neuro-oncology for a straightforward and effective means of on-table differentiation of normal brain tissue from abnormal brain tissue.

[0009] Bioimpedance is one of the most promising means of evaluating the physiological state of different tissue types, resulting from the biological structure of tissues causing impedance differences. Bioimpedance techniques involve passing an electrical current through biological tissue while concurrently measuring the resistance to the flow of current produced by the tissue.

[0010] When existing bioimpedance measuring devices are used as an intraoperative tissue identification tool, an operator is required to probe the area of interest one point at a time, by moving the probe, to measure bioimpedance at each discrete location to form a bioimpedance map. To obtain high spatial resolution bioimpedance map, the operator will need to reduce the electrodes’ pitch sizes and increase the number of measurements, leading to increased process / operation time. In addition, the added measurement time introduce errors into measurements as bioimpedance can change due to instantaneous physiological status and environmental factors over time.

[0011] There is a need for straightforward and effective bioimpedance characterisation apparatus that can conveniently be used by a surgeon to provide real-time intraoperative differentiation of brain tumour tissue from normal tissue, thereby facilitating (a) preservation of brain function by avoiding removal of normal or predominantly normal brain tissue; and (b) increased precision and accuracy in removing abnormal brain tissue, thereby facilitating GTR of the brain tumour. Other problems addressed by the present disclosure will become apparent from the description below and the accompanying drawings.

[0012] A prior disclosure of a bioimpedance-based measuring device was made by the present inventors in WO 2022 / 079401 A1 , upon which the present work provides a number of further improvements.

[0013] Summary of the Invention

[0014] According to a first aspect of the present invention there is provided apparatus for discriminating between tumour tissue and non-tumorous tissue in real-time, as defined in Claim 1 of the appended claims. Details of certain embodiments are set out in the dependent claims.

[0015] The present work utilises a multi-electrode probe and multiplexing and a computational algorithm to enhance bioimpedance spatial resolution and reduce the overall procedure time, and which is capable of true real-time tissue identification.

[0016] More particularly, according to the first aspect of the present invention there is provided apparatus for discriminating between tumour tissue and non-tumorous tissue in real-time, the apparatus comprising: a handheld bioimpedance probe having a plurality of electrodes mounted thereon, wherein the plurality of electrodes are greater than four in number, the electrodes being arranged to be held against a region of tissue in use; a current source configured to generate a stimulation current; a voltage sensor; one or more multiplexers coupled between the current source, the voltage sensor and the electrodes and configured to changeably select a subset of four electrodes from the plurality of electrodes, and, for each selected subset of four electrodes, to switch between a plurality of switching configurations, wherein, in each switching configuration, a first pair of the selected subset of four electrodes are connected to the current source, and a second pair of the selected subset of four electrodes are connected to the voltage sensor, the electrodes that constitute the first pair and the electrodes that constitute the second pair changing from one switching configuration to the next; and processor-controlled circuitry configured to: control the selection of the subset of four electrodes by the one or more multiplexers; control the switching configuration of the one or more multiplexers with respect to the selected subset of four electrodes; control the generation of the stimulation current by the current source and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that, in use, the stimulation current flows through the tissue between the first pair of electrodes; control the measurement, by the voltage sensor, of the voltage between the second pair of electrodes of each switching configuration, across the tissue in use; determine, based on the applied current and the measured voltages for each switching configuration, a measure of relative impedance (and optionally also a measure of conductivity, resistivity, permittivity, and local electric field) of the tissue for each switching configuration; and supply such impedance measurements (or raw time-series voltage data) as output data to a data analyser to enable real-time discrimination of the tissue to be performed, based on the impedance measurements.

[0017] The term “stimulation current” should be interpreted broadly, to encompass AC waveforms or potentially a DC signal; it may in practice be any arbitrary current.

[0018] Preferably each selected subset of four electrodes forms a quadrilateral shape, such as a convex quadrilateral shape. However, other quadrilateral shapes are also possible, such as concave quadrilaterals, for example. Preferably, for each selected subset of four electrodes, the switching configurations are six in number, comprising four bioimpedance measuring configurations, in each of which the first and second pairs of electrodes are on opposite sides of the quadrilateral to one another, and two configurations in which the first and second pairs of electrodes diagonally cross one another (these two configurations advantageously providing “tissue margin estimator” or “tissue margin likelihood” values, and also enabling the symmetrical-ness of the tissue to be evaluated).

[0019] In one implementational example, each of the plurality of electrodes may be connected to four different individually-controlled bidirectional analogue multiplexers.

[0020] The plurality of electrodes may be 12 in number, for example. Other numbers of electrodes are also possible, for example ranging from 5 to 20 in number, or greater.

[0021] Preferably, in use, at least some of the selected subsets of four electrodes overlap one another.

[0022] Preferably the data analyser is configured to generate a bioimpedance value for each selected subset of four electrodes. The data analyser may also be operable to generate an indication of the symmetrical-ness of the tissue, and / or the presence of a tissue margin between different type of tissues (for example, but in no way limited to, a tissue margin between healthy and tumorous / cancerous tissues).

[0023] To facilitate visualisation of the tissue structure, preferably the data analyser is further configured to generate a bioimpedance map (e.g. on a display screen) by superimposing quadrilaterals, each quadrilateral having a shape and position corresponding to the shape and position of the respective subset of four electrodes used to generate a respective bioimpedance value, and coloured or shaded to indicate the respective bioimpedance value. The map may also indicate the symmetrical-ness of the tissue, and the presence of a tissue margin between different type of tissues. The data analyser may be further configured to generate further polygons within the bioimpedance map where quadrilaterals overlap. Preferably the further polygons are coloured or shaded by the data analyser to represent an average of the bioimpedance values of the respective overlapping quadrilaterals.

[0024] To smooth the bioimpedance map, the data analyser may be further configured to apply a low-pass image filter to the bioimpedance map, for example by convolving it with a Gaussian blur kernel.

[0025] Optionally the processor-controlled circuitry may be further configured to obtain a plurality of voltage measurements (e.g. four measurements) from each switching configuration, for example through the use of a differential voltage sensor (e.g. a differential amplifier or instrumentation amplifier) across two of the electrodes and a voltage buffer on single electrodes.

[0026] Preferably, in each switching configuration, the processor-controlled circuitry is configured to control the measurement, by the voltage sensor, of the individual voltages of the second pair of electrodes, as well as the measurement of the potential difference between the second pair of electrodes. More particularly, the voltages of individual electrodes may be measured relative to a reference electrode, such as the return-current or current sink electrode (of the stimulating electrodes).

[0027] Optionally the data analyser may be further configured to iteratively determine the path of the tissue margin across the footprint of the probe.

[0028] According to a second aspect of the present invention there is provided apparatus for discriminating between tumour tissue and non-tumorous tissue in real-time, the apparatus comprising: a handheld bioimpedance probe having four or more electrodes mounted thereon, the electrodes being arranged to be held against a region of tissue in use; a current source configured to generate a stimulation current; a voltage sensor; a multiplexer coupled between the current source, the voltage sensor and the electrodes and configured to changeably switch between a plurality of switching configurations, wherein, in each switching configuration, a first pair of the electrodes are connected to the current source, and a second pair of the electrodes are connected to the voltage sensor, the electrodes that constitute the first pair and the electrodes that constitute the second pair changing from one switching configuration to the next; and processor-controlled circuitry configured to: control the switching configuration of the multiplexer; control the generation of the stimulation current by the current source and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that, in use, the stimulation current flows through the tissue between the first pair of electrodes; control the measurement, by the voltage sensor, of the voltage between the second pair of electrodes of each switching configuration, across the tissue in use, and also the measurement of the individual voltages of the second pair of electrodes; determine, based on the applied current and the measured voltages for each switching configuration, a measure of relative impedance (and optionally also a measure of conductivity, resistivity, permittivity, and local electric field) of the tissue for each switching configuration; and supply such impedance measurements (or raw time-series voltage data) as output data to a data analyser to enable realtime discrimination of the tissue to be performed, based on the impedance measurements.

[0029] The switching configurations may be six in number, comprising four bioimpedance measuring configurations, in each of which the first and second pairs of electrodes are substantially opposite one another, and two configurations in which the first and second pairs of electrodes diagonally cross one another.

[0030] With either the first or second aspects of the invention, the multiplexer may be configured to cyclically switch between each of the plurality of switching configurations.

[0031] Optionally, in each switching configuration, the processor-controlled circuitry may be configured to control simultaneous measurement of the voltages measured. The processor-controlled circuitry may comprise a microcontroller. The data analyser may comprise a personal computer (PC), or another suitable device such as a tablet computer or smartphone. The processor-controlled circuitry and / or the data analyser may be located within, or local to, the probe itself.

[0032] According to a third aspect of the present invention there is provided a method of discriminating between tumour tissue and non-tumorous tissue in real-time, using the apparatus according to the first aspect or the second aspect of the invention.

[0033] In certain embodiments the current source may be a voltage controlled current source, and the processor-controlled circuitry may further comprise a voltage waveform generator configured to generate a voltage waveform and to supply the voltage waveform to the current source. The current source may further comprise a high pass filter configured to remove DC offset from the voltage waveform and to convert the voltage waveform to the stimulation current. Such a high pass filter may not be required if the input of the voltage controlled current source is differential or bipolar.

[0034] Alternatively, a current waveform generator may be used, to generate a current waveform to directly drive the current source.

[0035] The voltage waveform generator or current waveform generator may comprise a digital-to-analogue converter configured to receive a predefined stimulation waveform from which the voltage waveform or current waveform is generated.

[0036] Alternatively, the voltage waveform generator or current waveform generator may comprise a Direct Digital Synthesis module configured to receive a predefined stimulation waveform from which the voltage waveform or current waveform is generated.

[0037] The stimulation waveform may be stored in the processor-controlled circuitry. The voltage waveform or current waveform may comprise a mix of a plurality of different frequencies, thereby enabling the tissue impedance at the different frequencies to be rapidly extracted using Fast Fourier Transform processing (or another appropriate technique). This greatly reduces the impedance measurement time compared to a standard chirp technique that would employ frequency sweeping, and also improves measurement accuracy. Preferably the voltage waveform or current waveform has substantially equal magnitude at each of the plurality of different frequencies.

[0038] Preferably the voltage sensor comprises an amplifier, such as an instrumentation amplifier, or a differential amplifier.

[0039] Preferably the processor-controlled circuitry further comprises an analogue-to- digital converter configured to receive and sample a voltage signal from the amplifier and thereby generate digital voltage data.

[0040] For embodiments in which the voltage waveform comprises a mix of a plurality of different frequencies, the processor-controlled circuitry may further comprise a Fast Fourier Transform processor configured to receive the digital voltage data from the analogue-to-digital converter, and to convert time-domain data into frequencydomain data and thereby extract the instantaneous magnitude of the voltage data for each frequency in the stimulation waveform.

[0041] In certain embodiments the voltage controlled current source and / or the amplifier are provided on a front-end printed circuit board or integrated circuit within the probe, e.g. local to the electrodes.

[0042] The digital-to-analogue converter and / or the analogue-to-digital converter may be provided on a microcontroller unit (MCU) platform.

[0043] In certain embodiments the electrodes may be spherical or hemispherical, and / or may be spring-loaded. Alternatively the electrodes may for example be needle- shaped or pointed, or the tips of wires. In certain embodiments the distal end of the elongate body may comprise a telescopic shaft on which the electrodes are mounted. Such an arrangement enables or facilitates endonasal access of the probe to the brain.

[0044] In certain embodiments the probe may be wireless, to facilitate unencumbered manipulation of the probe by the surgeon.

[0045] In certain embodiments the probe may further comprise a pressure sensor, for measuring the contact pressure between the electrodes and the tissue in use.

[0046] In certain embodiments the probe may further comprise a blood oxygen sensor at the distal end of the elongate body, for measuring the blood oxygen level of the tissue in use.

[0047] In certain embodiments the probe may further comprise an accelerometer or inertial sensor to sense the angle of inclination of the probe.

[0048] Brief Description of the Drawings

[0049] Embodiments of the invention will now be described, by way of example only, and with reference to the drawings in which:

[0050] Figure 1 a is a perspective view of a stimulator end of a handheld tetrapolar bioimpedance measurement probe having an elongate (optionally telescopic) shaftlike body and four stimulation / sensing electrodes in a square configuration (as in WO 2022 / 079401 A1 );

[0051] Figure 1 b illustrates, in plan view, the square configuration of the four (tetrapolar) stimulation / sensing electrodes of Figure 1 a;

[0052] Figure 2 illustrates, for a set of four stimulation / sensing electrodes, a series of six multiplexed electrode configurations;

[0053] Figure 3 illustrates how tissue margin (i.e. the interface between two different types of tissue - generally healthy tissue and tumorous tissue) can be identified through a combination of the electrode configurations of Figure 2, according to the methodology of the prior art; Figure 4 illustrates the capability of the methodology of WO 2022 / 079401 A when sampling a tissue sample, showing (a) bioimpedance and (b) tissue margin estimator maps taken under a grid of 16 x 16 measurements, and (c) a photograph of the tissue sample being sampled;

[0054] Figure 5 illustrates four tetrapolar measuring setups used for tumour-tissue margin analysis simulation, namely (A) electrodes on a uniform tissue surface, (B) a tumourtissue margin across adjacent electrodes, (C) a tumour-tissue margin along a diagonal axis, and (D) a tumour-tissue margin around one electrode (i.e. tumour tissue surrounding one electrode);

[0055] Figure 6a presents an array of finite element simulation results using electric field magnitude maps, wherein each row represents a simulation setup from A (top) to D (bottom) as in Figure 5, and each column represents one of the electrode configurations from 1 (left) to 6 (right) as in Figure 2;

[0056] Figure 6b presents two tables of data, namely Table I (finite element simulation results) and Table II (biological tissue measurement results);

[0057] Figure 7a illustrates additional tissue margin forms that the present methodology can identify using tetrapolar electrodes, in addition to those of the prior art methodology as shown in Figure 3;

[0058] Figure 7b illustrate the tissue margin forms of Figure 7a together with those of Figure 3;

[0059] Figure 8a is a table to explain how tissue margins may be identified based on sensing electrodes’ voltages and potential differences between sensing electrodes; Figure 8b links the “code names” in Figure 8a to the corresponding tissue margin representations;

[0060] Figure 8c shows a photograph of a porcine tissue sample with tissue margins from Figure 8b overlayed;

[0061] Figure 9a illustrates an example of a 12-electrode probe;

[0062] Figure 9b illustrates the formation of 13 groups of four (tetrapolar) electrodes, each having a convex quadrilateral shape, as subsets of the 12 electrodes of the probe of Figure 9a;

[0063] Figure 10 shows a porcine tissue sample under the probe of Figure 9a, with the black circles representing the electrode locations; Figure 11 illustrates a limited group of convex quadrilaterals from the possibilities of Figure 9b;

[0064] Figure 12 illustrates the principles of bioimpedance and tissue margin estimator superposition;

[0065] Figures 13 and 14 respectively show complete bioimpedance and tissue margin estimator maps using the 13 tetrapolar probe formations of Figure 9b on the tissue sample shown in Figure 10;

[0066] Figure 15 shows (a) a superposition weighted bioimpedance map and (b) a Gaussian kernel filtered bioimpedance map in respect of the tissue sample shown in Figure 10, obtained with limited (13-location) multiplexing, initially plotted with a “jet” colouring scheme and then later converted into greyscale;

[0067] Figure 16 shows the superposition of 13 weighted convex quadrilaterals and Figure 17 shows the superposition of all possible, 297, convex quadrilaterals when 12 electrodes are used (in the formation of Figure 9a), indicating an increase in resolution;

[0068] Figure 18 shows a porcine tissue sample under the probe of Figure 9a, with the black circles representing the electrode placements during measurement;

[0069] Figures 19 and 20 respectively show complete bioimpedance and tissue margin estimator maps using all possible, 297, convex quadrilateral tetrapolar probe formations on the tissue sample shown in Figure 18;

[0070] Figure 21 shows (a) a superposition weighted bioimpedance map and (b) a Gaussian kernel filtered bioimpedance map in respect of the tissue sample shown in Figure 18, obtained with all possible (297-location) convex quadrilateral multiplexing when 12 electrodes are used (in the formation of Figure 9a), initially plotted with a “jet” colouring scheme and then later converted into greyscale;

[0071] Figure 22 illustrates the capability of the present methodology when sampling a tissue sample, showing (a) a photograph of the tissue sample being sampled; (b) a 16x16 grid of bioimpedance values (without superposition); (c) a comparative 32x32 grid of bioimpedance values (without superposition); (d) a 32x32 grid showing bioimpedance superposition; and (e) a Gaussian kernel blurred (low-pass) map;

[0072] Figure 23 illustrates a methodology using the detection of margin alone to iteratively determine tissue margin under a 12-electrode probe;

[0073] Figure 24 illustrates an alternative methodology to that of Figure 23; Figure 25 illustrates another alternative methodology to that of Figure 23, showing where the current source pair of electrodes can be selected when all the electrode voltages are being measured simultaneously;

[0074] Figure 26 illustrates a block diagram of a first example bioimpedance characterisation system for use with, or incorporating, the present bioimpedance measurement electrodes;

[0075] Figure 27 illustrates a block diagram of a second example bioimpedance characterisation system for use with, or incorporating, the present bioimpedance measurement electrodes;

[0076] Figure 28 illustrates a block diagram of a third example bioimpedance characterisation system for use with, or incorporating, the present bioimpedance measurement electrodes; and

[0077] Figure 29 is flow diagram illustrating one possible manner of operation of the abovedescribed bioimpedance characterisation system and measurement probe.

[0078] In the figures, like elements are indicated by like reference numerals throughout.

[0079] Detailed Description of Preferred Embodiments

[0080] The present embodiments represent the best ways known to the Applicant of putting the invention into practice. However, they are not the only ways in which this can be achieved.

[0081] The present disclosure provides apparatus and methods for discriminating between different types of tissue (e.g. tumour tissue and non-tumorous tissue) in real-time using bioimpedance measurements. The apparatus and methods may also provide an indication of the symmetrical-ness of the tissue, and / or the presence of a tissue margin between different type of tissues.

[0082] More particularly, a handheld (preferably wireless) electrical ‘tumour identification’ probe is provided, which passes an electrical current through biological tissue while concurrently measuring the impedance to the flow of current produced by the tissue, i.e. bioimpedance. Away of measuring bioimpedance using four or more electrodes is provided, together with the use of an algorithm (implemented by a computer program) to reconstruct a bioimpedance map under the probe, for display on a display screen, to identify the tissue margin(s) under the probe in real-time.

[0083] Known bioimpedance measurement devices inject current through a pair of electrodes and only measure the potential differences of another pair of electrodes. Notably, the present disclosure introduces the principle of extracting individual electrode voltages (relative to a reference electrode, such as the return-current electrode of the stimulating electrodes) during bioimpedance measurements, as well as obtaining potential differences across pairs of electrodes. In this context, when we measure individual voltages, we take into account the physiological signal and electrical field at each of the electrodes, which typically would otherwise have been ignored or highly suppressed when using differential amplifiers and / or instrumentation amplifiers with high common-mode rejection ratio (CMRR).

[0084] In the present disclosure, it is also shown that, with a probe having more than four electrodes, any combination of four electrodes can be selected (as a tetrapolar subset of the total number of electrodes) to measure at least one respective bioimpedance value, and optionally an indication of the symmetrical-ness of the tissue and / or the presence of a tissue margin between different type of tissues, and this may be repeated across a plurality of, or all, the available combinations of four electrodes that likewise are subsets of the total number of electrodes. For example, for a probe having six electrodes in total, there are 15 discrete combinations of four electrodes that can be formed (using the Combination formulanCk, where n is the total number of electrodes, six in this example, and k is four). For 12 electrodes, 495 discrete combinations of four electrodes can be formed. For 13 electrodes, 715 discrete combinations of four electrodes can be formed, and so on. From each tetrapolar subset of the electrodes, the obtained bioimpedance values can then be combined to reconstruct a super-resolution bioimpedance map under the multielectrode probe through superposition, scaling, and the application of a low-pass filter to smooth out the image (based on the assumption that bioimpedance values change smoothly across adjacent tetrapolar subsets), for display on a display screen. By multiplexing within each tetrapolar subset of four electrodes, using the switching configurations 1 -6 shown in Figure 2, four bioimpedance measurements and two additional “tissue margin estimator” or “tissue margin likelihood” values, i.e. a total of six measuring configurations, can be obtained out of each four-electrode selection. As described in greater detail below, the “tissue margin estimator” or “tissue margin likelihood” values are obtained through measuring the structural symmetry of the local tissue and can advantageously be used to estimate the presence of a tissue margin bounded by the four electrodes.

[0085] To add another level of detail, a plurality of m voltage measurements can be taken from each measuring configuration. For example, m may be four.

[0086] Thus, for example, with a 12-electrode bioimpedance probe, with six measuring configurations, and taking four voltage measurements from each measuring configuration, a total of 495 x 6 x 4 = 11880 bioimpedance values (or sets of timeseries voltage raw data) can be obtained for bioimpedance super resolution and tissue identification, and optionally for tissue-margin identification at each location directly through the individual electrode voltages obtained.

[0087] Accordingly, multiple layers of multiplexing can be used to obtain bioimpedance values, where the number of obtainable bioimpedance values scales exponentially with the number of electrodes.

[0088] Review of prior art WO 2022 / 079401 A1

[0089] WO 2022 / 079401 A1 disclosed a handheld bioimpedance measurement probe. The present work provides developments over WO 2022 / 079401 A1 in respect of the processing of the data gathered by the electrodes, and also in respect of the number of stimulation / sensing electrodes used.

[0090] In the present work, when only four stimulation / sensing electrodes are used, the distal end of the probe may for example be as shown in the present Figure 1 a (which is adapted from Figure 2b of WO 2022 / 079401 A1 ). This shows the probe 10 having an elongate body 14 with the stimulation / sensing electrodes 1 a-1 d being mounted on the distal end of the elongate body 14. Additional stimulation / sensing electrodes may also be mounted on the distal end of the elongate body 14, for example in the arrangement shown in Figure 9a, that will be described later.

[0091] The rest of the elongate body 14, and the handheld probe 10 as a whole, is not shown in the present Figure 1 a, but may be substantially as shown in Figure 1 a of WO 2022 / 079401 A1 , for example. However, such a design of the overall probe is by no means essential to the present work, and alternative probe designs to support the stimulation / sensing electrodes of the present work may be used instead. The probe is preferably (but not necessarily) wireless, to facilitate unencumbered use during surgery.

[0092] In various embodiments, the electrodes 1 a-1d may be relatively long (e.g. pin-like, needle-shaped or pointed) or relatively short (e.g. stubby), and may be spherical or hemispherical in shape. In use, the tips of the electrodes 1 a-1 d are held against a region of tissue that is to be investigated for the presence of tumours, for the purpose of discriminating between tumour tissue and normal (non-tumorous) tissue. Optionally but advantageously, the electrodes 1 a-1 d may be spring-loaded so as to cushion their contact with the tissue being tested, to compensate for unevenness in the tissue surface, and to reduce the extent to which they press into the tissue.

[0093] As illustrated in the present Figure 1a, the stimulation / sensing electrodes 1 a-1 d may be mounted at the end of an optional telescopic shaft 18, that is extendable from, and retractable into, a cylindrical sleeve 16 at the distal end of the body 14. The length of the telescopic shaft 18 that protrudes from the cylindrical sleeve 16 can be lengthened or shortened and then fixed, to produce a desired case-specific or surgeon-specific length, thereby facilitating user-optimised application. An adjustment mechanism for adjusting and locking (or unlocking) the protruding length of the telescopic shaft 18 may be provided, as those skilled in the art will appreciate. The telescopic end of the probe 10 enables or facilitates endonasal access to the brain; for this the entire probe may be of a more elongated structure. In WO 2022 / 079401 A1 it was disclosed that, with at least four electrodes, using a square electrode probe arrangement as an example (as shown in the present Figures 1 a and 1 b), the presence of a margin can be determined through stimulating a first pair of electrodes with a stimulation current generated by a current source, and concurrently measuring the potential difference between a second pair of electrodes (the second pair being different from the first pair) using a voltage senor, using a sequence of multiplexer-driven electrode configurations such as is shown in the present Figure 2.

[0094] In more detail, in WO 2022 / 079401 A1 , a multiplexer is coupled between the current source, the voltage sensor and the stimulator electrodes and configured to changeably switch between a plurality of switching configurations such as is shown in the present Figure 2, wherein, in each switching configuration, a first pair of electrodes are connected to the current source, and a second pair of electrodes are connected to the voltage sensor, the electrodes that constitute the first pair and the electrodes that constitute the second pair being changed by the multiplexer from one switching configuration to the next. Processor-controlled circuitry is provided, that is configured to: control the switching configuration of the multiplexer; control the generation of the stimulation current by the current source and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that the stimulation current flows through the tissue between the first pair of electrodes; control the measurement, by the voltage sensor, of the voltage between the second pair of electrodes of each switching configuration, across the tissue; determine, based on the applied current and the measured voltage for each switching configuration, a measure of relative impedance of the tissue for each switching configuration; and supply such impedance measurements as output data to a data analyser to enable real-time discrimination of the tissue to be performed, based on the impedance measurements. The data analyser may comprise a personal computer (PC), or another suitable device such as a tablet computer or smartphone, for example. The reader is referred to WO 2022 / 079401 A1 for further details of the apparatus and methods disclosed therein. In more detail, Figure 2 shows six tetrapolar electrode configurations (configurations 1 -6) sequentially constructed by the multiplexer. In a practical implementation, the apparatus may measure the impedance in each configuration in a time frame of under 22 ms, for example.

[0095] In electrode configuration 1 , the stimulation current h is applied from electrode 1a to electrode 1 b, and the corresponding voltage Vi is measured between electrodes 1 c and 1 d.

[0096] Next, in electrode configuration 2, the stimulation current h is applied from electrode 1 d to electrode 1 a, and the corresponding voltage V2 is measured between electrodes 1 b and 1 c.

[0097] Then, in electrode configuration 3, the stimulation current I3 is applied from electrode 1 c to electrode 1 d, and the corresponding voltage V3 is measured between electrodes 1 a and 1 b.

[0098] Then, in electrode configuration 4, the stimulation current I4 is applied from electrode 1 b to electrode 1 c, and the corresponding voltage V4 is measured between electrodes 1 d and 1 a.

[0099] In electrode configuration 5, the stimulation current I5 is applied from electrode 1 a to electrode 1 c, and the corresponding voltage V5 is measured between electrodes 1 b and 1d.

[0100] Finally, in electrode configuration 6, the stimulation current h is applied from electrode 1 b to electrode 1 d, and the corresponding voltage Ve is measured between electrodes 1 a and 1 c.

[0101] The sequence of electrode configurations 1-6 may be repeated by the multiplexer in a cyclic manner. Using the principles of WO 2022 / 079401 A1 , the tissue margin can be identified through combination of the multiplexing configurations 1-6 shown in Figure 2, to determine that the tissue under the probe is one of the forms (a)-(d) shown in Figure 3. In Figure 3, the four electrodes have the same arrangement as in Figure 2. The heavy diagonal shaded areas in Figure 3(a)-(d) may represent high impedance and the lighter shaded areas may represent low impedance (or vice versa). In other words, the heavy diagonal shaded areas in Figure 3(a)-(d) may represent low conductivity and / or low permittivity, and the lighter shaded areas may represent high conductivity and / or high permittivity (or vice versa). It can be seen that, in illustration (a) of Figure 3, the tissue margin between high impedance tissue and low impedance tissue passes between electrodes 1 a & 1 b on one side, and electrodes 1 c & 1d on the other side. In illustration (b) of Figure 3, the tissue margin passes between electrodes 1 a & 1 d on one side, and electrodes 1 b & 1 c on the other side. In illustration (c) of Figure 3, all four electrodes 1 a-1 d are within high-impedance tissue, and in illustration (d) of Figure 3, all four electrodes 1a-1 d are within low- impedance tissue. That is to say, in both illustrations (c) and (d), there is no tissue margin in the vicinity of the electrodes.

[0102] Figure 4 illustrates the capability of the methodology of WO 2022 / 079401 A1 when sampling a tissue sample, showing (a) bioimpedance and (b) tissue margin estimator maps, and (c) a photograph of the tissue sample being sampled. The bioimpedance and tissue margin estimator maps were taken under a grid of 16 x 16 measurements. It can be seen that the maps resemble the tissue sample relatively coarsely. Based on bioimpedance measurements obtained using the multiplexing configurations 1-4 of Figure 2, the tissue margin can only be identified through sampling throughout the tissue sample regions. The tissue margin estimator based on multiplexing configurations 5 and 6 of Figure 2 can perform spot checks to identify areas with asymmetrical electrical tissue properties.

[0103] Part 1 of the present disclosure: Real-time Multiplexed Bioimpedance Tumour-Tissue Margin Analysis

[0104] Turning now to a first part of the present work, it will be shown how combinations of individual electrode voltages and voltages across electrodes can be used during each measuring configuration to determine the presence of tissue margin, as well as how the margin crosses the electrodes. In this respect, the present Figure 7a shows the additional tissue margin forms that the present method can identify, in addition to those of the prior art methodology as shown in Figure 3.

[0105] Moreover, instead of obtaining complete bioimpedance visibility of an area of interest, the present work enables the user to perform spot checks to locate an area with interesting properties, e.g., tumour-tissue margins, abnormal tissues, etc.

[0106] Further, in the present work, greater use is made of configurations 5 and 6 of the present Figure 2. These configurations may be used to provide “tissue margin estimator” values or “tissue margin likelihood” values, based on the symmetricalness of the electrical properties of the tissue under the probe.

[0107] Based on an initial set(s) of measurements, it is then possible to narrow down where the next set(s) of measurements should be taken. This can be performed as an iterative process, to locate the margin or produce a high-resolution image when working with complex tissue structures, for example.

[0108] - Further introduction for part 1

[0109] As outlined above, bioimpedance has been used successfully to differentiate healthy and cancerous tissue in human tissue. Low-frequency impedance provides a measurement of the extracellular matrix, while high-frequency impedance is a measure of both intra- and extra-cellular matrix. The current state-of-the-art bioimpedance tissue identification technique uses high density electrode arrays, capturing the bioimpedance differences between each electrode, similar to the rolling shutter method in a camera. This is then used to reconstruct the bioimpedance image. However, tissue impedance values vary in relation to physiological state, environmental conditions, and force applied at the tissueelectrode interface. Further, tissue impedance is likely to be different depending on the type of tissue (e.g. human versus animal). These constraints significantly increase the challenge of identifying changes in tissue properties around a tumourtissue interface using bioimpedance measurements. We now seek to reduce the number of discrete bioimpedance measurements that are required to locate the tumour-tissue margin by utilising additional information from existing tetrapolar electrodes.

[0110] In WO 2022 / 079401 A1 we demonstrated a bioimpedance spectroscopy device for performing real-time tissue analysis using a typical tetrapolar impedance measuring configuration. We showed that it can be used to generate an impedance map of the tissue sample. We now extend our previous work by multiplexing throughout the full set of available electrodes (configurations 1 -6 of Figure 2), to reduce the overall negative sensitivity of the tetrapolar bioimpedance sensing configuration, whilst also obtaining additional individual electrode voltages relative to a reference electrode (as well as obtaining potential differences across pairs of electrodes). We show that tumour tissue margin analysis can be performed locally without the need to obtain the complete impedance map. Finite element simulations are used to explain the methodology, and porcine tissue samples are used to prove that the principle can be applied to animal tissue types. We show that multiplexing throughout the full set of measuring configurations (configurations 1-6 of Figure 2) and obtaining individual electrode voltages relative to a reference electrode (as well as potential differences across pairs of electrodes) can enable more precise and more flexible delineation of the tumour-tissue margin bounded by the electrodes.

[0111] - Multiplexed electrode configurations

[0112] Bioimpedance is the measure of the tissue’s ability to trap and accumulate charge, as well as carrying ions and charges from one location to another within a specified area. As the electrode makes contact with the tissue, an ionic double-layer forms on the electrode surface, which is known as polarisation, affecting the electrode impedance. Its effect varies with contact area, excitation current density, and frequency. The tetrapolar (4-point) impedance measurements in bioimpedance measurement plays an important role in minimising systematic errors produced by this effect.

[0113] In a tetrapolar impedance measuring configuration, the pair of ‘stimulation’ (source and sink) electrodes is used to inject / extract the current, while the pair of ‘sensing’ electrodes measure the potential difference across the tissue between the sensing electrodes.

[0114] The possible measuring configurations of a multiplexed tetrapolar electrode bioimpedance probe are shown in Figure 2. Configurations 1 and 3 measure latitudinal bioimpedance relationships, and configurations 2 and 4 measure longitudinal bioimpedance relationships. We then further use configurations 5 and 6 for supplementary measurements where the ‘sensing’ electrodes are placed on an equipotential line for optimal tissue-margin analysis. Therefore, configurations 5 and 6 provide measurements which are useful when a margin crosses within the tetrapolar electrodes.

[0115] - Tumour-tissue margin analysis

[0116] The principle of detecting the margin between tissue types revolves around exploiting the discontinuity that arises in the induced electric field due to the presence of abnormal (e.g. cancerous) tissue.

[0117] Firstly, the measurement setup stimulates the probed area with an electric field distribution determined by the selected source and sink electrodes. In one example, the current source electrode may be responsible for both sourcing and sinking current, as a single-ended current source, while the sink electrode provides a current return path, being a virtual ground, ground, or biased to a reference voltage to complete the current loop (conduction path). In another example, the current source and sink electrodes may actively inject and extract the same amount of current, as a differential current source.

[0118] Next, the actual established electric field is then measured by multiplexing and measuring the voltage across all electrode combinations, by measuring potential differences between the electrodes (e.g. between a reference electrode such as the current return electrode and the three other electrodes) as well as to obtain individual electrodes’ voltages. This enables scanning of the entire probed area and the creation of a ‘voltage-based’ topographical map V(x,y). The corresponding electric field map E(x,y) can be computed from the voltage map via differentiation:

[0119] Equation (1 )

[0120] If the sampled electric field map of equation (1 ) is an accurate representation of the underlying electric field, then examining the discontinuities in this map will reveal the margin between tissue types.

[0121] A. Simulation setup

[0122] ANSYS Electronics Desktop with Maxwell AC-Conduction solver was used as the finite element simulation environment to calculate the electric field strength distribution plots present in multiple materials. Properties of the brain tissues were chosen for the simulation. The resistivity of healthy and cancerous ‘brain’ tissue (glioma) was set to 100Q. cm and 20Q. cm respectively

[0001] , and the relative permittivity of brain tissues was set to 2.25 x 104[2], The glioma was modelled as a discrete lump of tissue with the tumour-tissue interface serving as a resectable margin. The electrodes were represented by 0.5mm diameter cylinders with 2.5mm pitch on the surface. A 10 kHz excitation current with 100 pA amplitude injection current was set, and a 0V excitation voltage were used to complete the current loop.

[0123] Figure 5 shows four possible situations for the manner in which tissue / tumour interfaces may be encountered under a tetrapolar bioimpedance probe. In setup A, the electrodes are placed on a surface with healthy ‘brain’ tissue. In setup B, the bottom two electrodes (electrodes 1a & 1 b) are on tumour tissues and the top two electrodes (electrodes 1c & 1 d) are on healthy ‘brain’ tissue; a tumour-tissue margin is present between electrodes 1 a & 1 b on the one side, and electrodes 1 c & 1 d on the other side. In setup C, the tumour-tissue margin is present diagonally between electrodes 1 b & 1d, with electrodes 1 a & 1 c placed on the margin. In setup D, electrode 1 a is on the tumour tissue and the other three electrodes are on healthy ‘brain’ tissue. B. Simulation results and discussion

[0124] The electric field magnitude distribution for each setup and the measuring configurations are shown in Figure 6a, and Table I of Figure 6b shows the corresponding voltage on and the potential differences between the sensing electrodes. Relative bioimpedance is directly proportional to the potential difference between the sensing electrodes in configurations 1 to 4. This is before any scaling (such as geometric factor scaling, for compensation of the current stimulation and voltage sensing pair of electrodes not being placed at the same point in tetrapolar (4-point) impedance measurements).

[0125] In Table I (and subsequently Table II), VA denotes the voltage on electrode 1 a, VB denotes the voltage on electrode 1 b, Vc denotes the voltage on electrode 1 c, and

[0126] VD denotes the voltage on electrode 1 d. Vi denotes the potential difference between electrodes 1d and 1 c (i.e. VD - Vc), V2 denotes the potential difference between electrodes 1 c and 1 b (i.e. Vc - VB), V3 denotes the potential difference between electrodes 1 b and 1 a (i.e. VB - VA), V4 denotes the potential difference between electrodes 1a and 1d (i.e. VA - VD), V5 denotes the potential difference between electrodes 1d and 1 b (i.e. VD - VB), and 3 denotes the potential difference between electrodes 1 c and 1 a (i.e. Vc - VA).

[0127] Figure 6a presents an array of finite element simulation results using electric field magnitude maps, wherein each row represents a simulation setup from A (top) to D (bottom) as in Figure 5, and each column represents one of the electrode configurations from 1 (left) to 6 (right) as in Figure 2.

[0128] Row 1 of Figure 6a shows setup As electric field magnitude in each measuring configuration. No visible discontinuity in the normal field agrees with the absence of a tumour tissue margin in this setup. The potential difference across the sensing electrodes in configurations 1 to 4 measure approximately the same voltage (with error due to finite element meshing differences), and configurations 5 and 6 measure negligible potential differences due to the symmetry across the stimulation electrodes. A clear discontinuity in the electric field magnitude can be seen in row 2 (setup B) of Figure 6a, separating the top and bottom half of the electrodes, and the rate of change (of electrical field magnitude and voltage differences) increases in the bottom half indicating a tissue with higher conductivity. This is reflected through voltage measurements, the bioimpedance between configurations 1 & 3 and configurations 2 & 4 differ while the electrodes’ voltage measurements within configurations 1 & 2 and configurations 3 & 4 are comparable, suggesting a tumourtissue margin is likely present between the upper electrodes (1c & 1d) and the lower electrodes (1 a & 1 b). Configurations 5 & 6 further confirm the presence of a margin with a high potential difference between the sensing electrodes.

[0129] For setup C, the electric field magnitude discontinuity is clearly visible in measuring configurations 2, 3 and 6 (as shown in row 3 of Figure 6a), but the bioimpedance measurements suggest the electrodes are on a uniform tissue surface. By observing the sensing electrodes’ voltages, they can be separated into three groups of voltages, indicating the presence of a tissue margin diagonally. The sensing electrode voltages in configuration 6 are higher than those in configuration 5, confirming a margin is present diagonally, with electrodes 1a & 1 c located on the margin.

[0130] A circular shaped discontinuity can be seen in all subfigures in row 4 of Figure 6a, but the bioimpedance measurements suggest the electrodes are on a uniform tissue surface again. Observing the voltage on the sensing electrodes in each configuration confirms the observation that the margin is indeed present around electrode 1 b, with the tissue conductivity being higher directly under electrode 1 b.

[0131] Accordingly, the above simulation results confirm that obtaining voltage information in respect of specific sensing electrodes enables tumour-tissue margins to be determined with greater accuracy. - Experimental results and discussion

[0132] A. System overview and experimental setup

[0133] A custom bioimpedance probe implementing a previously established system architecture [3] was used. The bioimpedance probe featured (i) synchronised 16- bit DAC and ADC sampling at 1 MSPS; (ii) multiplexing of the measurement configurations corresponding to Fig. 2, and (iii) excitation current configured to 10 kHz sine wave with ±100pA amplitude. The bioimpedance probe was mounted onto a motorised 3-axis system for improved controlled movement and automation. Four gold-plated spring-loaded electrodes in a square configuration with a 2.54mm pitch were used.

[0134] Slices of porcine tissue samples with visible fat-muscle margins were used to demonstrate proof of concept. The tissue samples were placed on the motorised test bed, with manually implemented electrodes aligned to the tissue margin before taking each measurement.

[0135] B. Biological tissue measurement results and discussion

[0136] Table II of Figure 6b shows the voltage on and the potential differences between the sensing electrodes when applied to porcine tissue. The bioimpedance probe was located in a region with visible muscle-only content in setup A(1 ) and fat-only content in setup A(2). In setup B, electrodes 1 a & 1 b were on muscle and electrodes 1 c & 1 d were on fat content. Finally, in setup D, the bioimpedance probe was placed on an area with one electrode on muscle tissue and the other three electrodes on fat.

[0137] In Table II, the measurements of setup A(1 ) show a typical uniform surface impedance, with configurations 1 to 4 measuring a similar voltage and configurations 5 & 6 measuring negligible potential differences; this aligns to the setup A simulation results of Table I.

[0138] In setup A(2), configurations 1 & 3 measure a lower voltage than configurations 2 & 4, indicating a potential margin - however, this was not indicated by configurations 5 & 6. This could be explained by underlying tissue properties having a directional dependency, as has previously been observed in other tissue types [4],

[0139] In setup B, the bioimpedance between configurations 1 & 3 and configurations 2 & 4 differ while the electrodes’ voltage measurements within configurations 1 & 2 and configurations 3 & 4 are reasonably comparable. In addition, the measurements of configurations 5 & 6 were higher compared to the measurements in setup A(1 ) and A(2), indicating there was a margin bounded by the electrodes.

[0140] The measurements performed on porcine tissue show that the presence of a margin bounded by the tetrapolar bioimpedance probe can be interpreted through the electrodes’ voltages and potential differences across the sensing electrodes. The measurement-deduced margin patterns also align with the observed tissue margin.

[0141] - Conclusion

[0142] In part 1 of the present work, a bioimpedance tumour-tissue margin analysis methodology based on a multiplexed tetrapolar bioimpedance probe has been presented. Finite element simulation based on healthy and cancerous brain tissue properties and porcine tissue sample testing have shown promising results produced by multiplexing between different measurement configurations while concurrently monitoring the voltages of the individual sensing electrodes. We have demonstrated a robust proof-of-concept, utilising a multiplexed bioimpedance methodology superimposed on a tetrapolar electrode architecture to enable identification of the existence and orientation of bioinspired tumour tissue margins.

[0143] Examples of the tissue margins that can be identified through this approach are shown in Figure 7a, to supplement those of Figure 3. Specially, images (a) to (d) of Figure 7a each show a single tissue margin surrounding a single electrode, images (e) and (f) both show diagonal tissue margins spanning three electrodes, and images (g) and (h) both show two tissue margins, each surrounding a diagonally- opposing electrode. - Further implementational details

[0144] Figure 8a illustrates how tissue margins may be identified based on sensing electrodes’ voltages and potential differences between sensing electrodes. We start with a set of pre-defined tissue electrical properties data and simulate the corresponding electrode voltages and potential differences between electrodes when the electrodes are on a surface with tissue-margin patterns similar to the ones shown in Figure 7b.

[0145] The simulated results are used to generate a table, as in Figure 8a, detailing if the measured data follows the pattern shown on Figure 8a, the margin would likely be one of the tissue-margin patterns shown in Figure 8b (which correspond to the first column of tissue margin naming).

[0146] More particularly, in the table of Figure 8a: MagZ denote the measuring configuration 1 -4 voltages (as shown in Figure 2), MagS denote the measuring configuration 5 & 6 voltages (as shown in Figure 2), PosE & NegE denotes the positive and negative sensing electrode voltages.

[0147] The naming of “L”, “M”, “H” represents three coarse level of voltages / values, correspond to low, medium and high, respectively.

[0148] The results of Figure 8a are simulated results using tissues without a directional dependency property. Hence, differences might be observed between these and the porcine tissue sample measurements given in Table II of Figure 6b.

[0149] Figure 8c shows a porcine tissue sample with the corresponding estimated tissuemargin overlay (black lines) based on Figure 8a data features patterns.

[0150] This process allows spot checking of the tissue under test, identifying the presence of tissue margins without first needing to determine the bioimpedance of the surrounding tissue. Instead of using binary or fixed levels of L, M, H patterns, these patterns can be numerical. Machine learning may be used to recognise these patterns and feedback into the system for tissue-margin analysis.

[0151] Part 2 of the present disclosure: Multielectrode Multiplexing for Bioimpedance Surface Topography Mapping

[0152] In part 2 of the present work, we extend the four-electrode probe configuration of e.g. Figures 1 a and 1 b to include more electrodes. The resulting plurality of electrodes (more than four in number) are then used to form different tetrapolar configurations, as subsets of the overall set of electrodes, to increase the resolution of the bioimpedance measurements made. The concepts of part 2 of the present work may be employed in combination with those of part 1. Alternatively, the concepts of either part 1 or part 2 may be used separately.

[0153] - Further introduction for part 2

[0154] Bioimpedance at different frequencies provides a means of tissue identification. Through the injection of electrical current into tissues, the measured voltage is a representation of the macroscopic (mm-to-cm) and mesoscopic (sub-mm to subcm) structure of the tissue.

[0155] When bioimpedance is used as an intraoperative tissue identification tool, conventionally an operator is required to probe the area of interest one point at a time, by moving the probe, to measure bioimpedance at each discrete location to form a bioimpedance map. The spatial resolution of the resulting bioimpedance map is typically based on the pitch of the electrodes and the number of discrete measurements made, but this adds extra time to the process. Furthermore, since bioimpedance can change due to instantaneous physiological status and environmental factors over time, bioimpedance measurements within the area of interest should be taken in the least amount of time, where possible. The increased procedure time for higher bioimpedance spatial resolution has been one of the limiting factors towards real-time tissue identification with bioimpedance. In WO 2022 / 079401 A1 we demonstrated a bioimpedance spectroscopy device for discrete location bioimpedance measurement. We have leveraged the same data converters and analogue front-end electronics architecture for bioimpedance measurements and used it as the system demonstrator foundation in the present work. We extend our work here, using a computational bioimpedance technique with an analogue multiplexing circuit and a multi-electrode probe architecture. We therefore demonstrate the ability to reconstruct a high spatial resolution ‘foveated’ local bioimpedance map, which exceeds the spatial resolution previously limited by the pitch size of the electrodes.

[0156] - Methodology

[0157] A. Bioimpedance electrode multiplexing

[0158] In traditional bioimpedance measurement, a tetrapolar measuring configuration is preferred, to minimise electrode polarisation effects [5], Figure 2 shows all possible measuring configurations that can be formed with a tetrapolar electrode probe. Configurations 1 to 4 measure the bioimpedance and configurations 5 & 6 measure the structural symmetry of the local tissue, to estimate the presence of a tissue margin bounded by the four electrodes - we may refer to this as the “tissue margin estimator”. These six measuring configurations form the basic tetrapolar probe element for the internal multiplexing of the present multi-electrode probes.

[0159] Here, we now develop and scale this fundamental concept by implementing a probe geometry having a plurality of electrodes, greater than four in number. By way of example, Figure 9a illustrates a 12-electrode probe geometry (that, in practice, may be between 15 mm and 20 mm in diameter, for example, or smaller). Each electrode may be connected to four different individually-controlled bidirectional analogue multiplexers (MUX), where each MUX common terminal connects to either the current source sink / source outputs or the input terminals of instrumentation amplifiers (or differential amplifiers, or differential amplifier circuits formed with generic operational amplifiers (op-amps), for example). With any four separated electrodes selected, a bioimpedance value can be measured. Optionally the bioimpedance values may then be scaled, e.g. with a geometry factor, for example. Figure 9b shows a basic example of this multi-electrode multiplexing methodology, with 13 groups of tetrapolar electrodes being formed (represented by different styles of lines), as subsets of the overall set of 12 electrodes.

[0160] With any 12-electrode probe, a total of 495 unique subsets of four electrodes can be selected to form a quadrilateral tetrapolar probe for bioimpedance measurements.

[0161] More generally, the number of quadrilateral tetrapolar probes that can be formed, as subsets of the overall set of electrodes, is given by the Combination formulanCk, namely: where k = 4, and n is equal to the total number of electrodes (e.g. 12 in the above example).

[0162] The number of discrete bioimpedance measurements that can be measured within a multi-electrode probe grows exponentially as the number of electrodes increases. To simplify the generation and visualisation of the bioimpedance map, only convex quadrilateral groups of electrodes may be selected as the tetrapolar subsets. From the 12-electrode geometry of Figure 9a, a total of 297 unique convex quadrilaterals when 12 electrodes are used (in the formation of Figure 9a), can thereby be formed for bioimpedance measurement purposes, with each convex quadrilateral acting as a basic tetrapolar probe element. Some examples of such convex quadrilateral subsets of electrodes are illustrated in Figure 9b. It will be appreciated that many of the convex quadrilateral subsets overlap one another.

[0163] For each tetrapolar subset of four electrodes that is formed, the multiplexer switching configurations 1-6 shown in Figure 2 can then be applied, to obtain four bioimpedance measurements (from configurations 1 -4) and two additional “tissue margin estimator” or “tissue margin likelihood” values (from configurations 5 & 6), as discussed above. The four bioimpedance measurements may be averaged to form a representative bioimpedance value for the quadrilateral defined by the four electrodes. The “tissue margin estimator” or “tissue margin likelihood” values, obtained through measuring the structural symmetry of the local tissue, can advantageously be used to estimate the presence of a tissue margin bounded by the four electrodes.

[0164] To add another level of detail, a plurality of m voltage measurements can be taken from each measuring configuration 1 -6. For example, m may be four.

[0165] For completeness, though, it should be noted that the present methodology is not limited to only the use of convex quadrilaterals as the tetrapolar subsets. Both convex and concave quadrilaterals can be used as the tetrapolar subsets, to measure bioimpedance, conductivity, tissue margin estimator, etc. The measured number obtained through electrode configurations 1 -4 or 5 & 6 of Figure 2 can be scaled or normalised based on the area bounded by the four electrodes, combinations of individual or combined distances between each electrode, a geometry factor based on Ohm’s law and Laplace’s equation, etc.

[0166] B. Computational bioimpedance

[0167] As each convex (or other shaped) quadrilateral has a corresponding bioimpedance value and tissue margin estimator value, where each convex (or other shaped) quadrilateral might have different sizes and shapes that also partially overlap with another, this creates a set of basis functions from which a filled image can be made from a single tetrapolar probe position, the image having a shape and position corresponding to the respective quadrilateral (i.e. corresponding to the shape and position of the respective subset of four electrodes used to generate the respective bioimpedance value), and having a colour or shade (e.g. greyscale shade) that represents the corresponding bioimpedance value. A bioimpedance image with straight edges can then be formed by superimposing all the convex (or other shaped) quadrilaterals, whilst retaining their relative positions within the overall set of electrodes. To overcome sharp boundaries caused by the overlap of coarse-scale basis functions, we may assume that the bioimpedance values vary smoothly across the boundaries of the quadrilaterals. This is achieved by applying a low-pass filter to the image by convolving it with a Gaussian blur kernel. As there are limited numbers of overlaps across all quadrilateral regions, the standard deviation for the Gaussian kernel for the 13-location reconstructed bioimpedance image was set to a value 3.5 times higher than the one for the 297-location reconstructed bioimpedance image.

[0168] Figure 11 shows a limited group of four adjacent convex quadrilaterals. Figure 12 then illustrates the basics of bioimpedance and tissue margin estimator superposition. Figure 13, and later Figure 15, show the superposition of the bioimpedance measurements over the 13 locations shown in Figure 9b, and Figure 14 shows the corresponding superposition of the tissue margin estimator measurements.

[0169] More particularly, in Figure 12 the top row of sub-figures ((a) to (d)) shows how individual bioimpedance values can form a filled image at each of the individual tetrapolar probe positions shown in Figure 11. Two neighbouring filled images can be superimposed on each other and the overlapped polygonal area (which may be triangular, for example) can be represented by an average of the two filled images (average of the two bioimpedance I tissue margin estimator values), as shown in the middle row of sub-figures ((e) and (f)). Alternatively, instead of just using averaging for the overlapped area, a combination of averaging with a scaling parameter (e.g. an area-dependent factor) can be used. And finally, further superposition of the already superposition filled images can be combined together in the same fashion (averaging) to form the final superposition filled image at the bottom (sub-figure (g)) that combines four bioimpedance I tissue margin estimator values together. As there are overlapped areas, this essentially increase the resolution of the bioimpedance and tissue margin estimator map.

[0170] The measured results can be fed back into a suitable processor, to create a computational model representing the measured results. Through an iterative process, the computational model can be used to give out the graphical representation of the tissue under the probe or over the measured areas.

[0171] - System demonstrator

[0172] A system demonstrator was produced, consisting of a 12-electrode probe with four 16-to-1 MUX connecting to the bioimpedance probe analogue front-end electronics and data converters. The bioimpedance probe featured (i) synchronised 16-bit DAC and ADC sampling at 1 MSPS; (ii) 3-bit programmable gain amplifier (up to 64x gain); and (iii) voltage-controlled current source with a maximum of ±100pA amplitude. A 12-electrode probe with gold-plated spring-loaded electrodes was used, based on the geometry of Figure 9a, during the measurements of Figures 13, 14, 15, 19, 20 and 21.

[0173] A multi-frequency waveform with 1 kHz, 20 kHz and 100 kHz frequency components was used, where 8000 samples were saved from each measurement. Sampling was performed with limited multiplexing (using the 13 tetrapolar probe formations shown in Figure 9b) on the tissue sample of Figure 10, and with all possible convex quadrilateral electrode multiplexing (297 tetrapolar probe formations) on the tissue sample of Figure 18. The total recorded times for the 13-location and 297-location samples were 624 ms and 14.3 s, respectively. Fast Fourier transforms were applied on the measured voltage to extract the magnitude and phase at each frequency; only the magnitude information was used to calculate the bioimpedance and form the reconstructed images.

[0174] - Resuits and discussion

[0175] The complete bioimpedance and tissue margin estimator maps based on the 13 single tetrapolar probe locations on the tissue sample shown in Figure 10 are shown in Figures 13(a) and 14(a), and the corresponding Gaussian kernel blurred (low- passed) maps are shown in Figures 13(b) and 14(b).

[0176] Figure 15(a) shows the bioimpedance of Figure 10 initially plot with a “jet” colouring scheme, then later converted into grey scale, and again as a Gaussian kernel blurred (low-passed) map in Figure 15(b). In the reconstructed bioimpedance images shown in Figures 15(a) and 15(b), regions with values greater than 150 indicate fat tissues and regions with values less than 100 indicate muscle tissues. Thus, the centre darker shade with high bioimpedance values highlights an area of fat; this aligns well with the ex-vivo sample image shown in Figure 10. (In passing, we note that the outer edge of the Gaussian kernel blurred image has some artifacts and are not representative of the measured tissue.) The emergence of superresolution, i.e. increased spatial resolution, using the presented methodologies and constrained number of electrodes is evidenced by the high degree of concordance between the location of the muscle and fat in Figure 10 as compared to the recreated computational image of Figure 15(b).

[0177] Figure 16 shows the superposition of 13 equally weighted convex quadrilaterals and Figure 17 shows all possible, 297, equally weighted convex quadrilaterals, indicating an increase in resolution. More particularly, Figures 16(a) and 16(b) show Figure 9b’s 13 groups of tetrapolar probe positions’ superposition of equally weighted quadrilaterals and the Gaussian kernel interpolated bioimpedance image, respectively. Figures 17(a) and 17(b) show all the possible equally-weighted convex quadrilateral superpositions and the reconstructed bioimpedance image, respectively.

[0178] Figure 18 shows the sampled porcine tissue sample on which all possible convex quadrilateral electrode multiplexing (297 tetrapolar probe formations) was performed, Figures 19(a) and 20(a) show the respective bioimpedance and tissue margin estimator maps, and Figures 19(b) and 20(b) show the corresponding Gaussian kernel blurred (low-passed) maps.

[0179] Figure 21 (a) shows the bioimpedance of Figure 18 initially plot with a “jet” colouring scheme, then later converted into grey scale, and again as a Gaussian kernel blurred (low-passed) map in Figure 21 (b). The reconstructed bioimpedance image from Figures 21 (a) and 21 (b) shows that the south-west to north-east sectors have a lower bioimpedance, the east sector has a higher bioimpedance, and a mixture of bioimpedance is measured near the south sector. Accordingly, it can be seen that significantly higher spatial resolution bioimpedance imaging was achieved using a complex, comprehensive multiplexing approach (as in Figure 21 ) as compared with the constrained multiplexing approach of Figure 15.

[0180] - Conclusion

[0181] In part 2 of the present work, we have demonstrated a bioimpedance sensing modality whereby multiplexing is achieved throughout all available electrodes within a multi-electrode bioimpedance probe, and sampling is obtained as a set of weighted quadrilateral basis functions. Superimposing the bioimpedance values of many independently sampled and partially overlapped quadrilateral areas leads to a significant increase in the bioimpedance spatial resolution as compared to the traditional single-area bioimpedance measurements. This methodology was tested on multiple ex-vivo porcine tissues, with the bioimpedance and tissue margin estimator images correlating well with tissue structure images obtained with a digital camera and providing mesoscopic tissue structure information. Accordingly, this work provides a platform for forming high-resolution local bioimpedance images for real-time tissue identification.

[0182] To demonstrate this further, Figure 22 illustrates the capability of the present methodology when sampling a tissue sample, showing (a) a photograph of the tissue sample being sampled; (b) a 16x16 grid of bioimpedance values (without superposition); (c) a comparative 32x32 grid of bioimpedance values (without superposition); (d) a 32x32 grid showing bioimpedance superposition; and (e) a Gaussian kernel blurred (low-pass) map.

[0183] Iterative determination of tissue margin under the footprint of a multielectrode probe

[0184] Figure 23 illustrates a methodology (which may be implemented by means of a computer program or software algorithm executed by the data analyser) using the detection of margin alone to iteratively determine the path of tissue margin under a multi-electrode probe having more than four electrodes. To illustrate this methodology, a 12-electrode probe will be used, by way of example. In each subfigure, the filled circles represent the selected four electrodes (as tetrapolar subsets of the overall number of electrodes), the dotted line represents the general direction of the tissue margin, and the lack of dotted line indicates there is a lack of tissue margin within the four electrodes. This methodology takes an iterative approach in the selection of the tetrapolar subsets of electrodes, to iteratively determine the path of the tissue margin across the footprint of the multi-electrode probe.

[0185] Referring to the sub-figures from left to right and top to bottom (a - h):

[0186] 1. Starting based on user input or default setting to select the centre four electrodes, a horizontal tissue margin was identified through measurements.

[0187] 2. As the direction of the tissue margin is on the left- and right-hand side of the initial four electrodes, the next selected subset of four electrodes would include the right two electrodes (notionally electrodes b and c in the insert in the bottom right corner of Figure 23) or the left two electrodes (electrodes a and d).

[0188] 3. Moving on to sub-figure b, it is measured that there is a margin surrounding electrode a (bottom-left) within the next-selected subset of four electrodes.

[0189] 4. Moving on to sub-figure c, following the direction of the margin, the four selected electrodes also have two common electrodes to the last set of selected electrodes. After measurements, it is estimated that there is a margin surrounding electrode c (top-right).

[0190] 5. Moving on to sub-figure d, as the last measurement margin has reached the outer edge of all electrodes, the next group of electrodes will include the left two electrodes from the initial electrode group as shown in sub-figure a. After measurements, it is estimated that there is a tissue margin separating the top and bottom two electrodes within the selected group.

[0191] 6. As a tissue margin has been estimated reaching from one side of the probe to the other side, the next step would be to measure the unknown shaded area as shown in sub-figure e.

[0192] 7. In sub-figure f, the selected group of electrodes are selected to split this particular shaded region into three groups, and the group of electrodes in sub-figure g are selected at random. After both these measurements, it is determined that there is no tissue margin present within each of these groups of electrodes.

[0193] 8. Sub-figure h shows the result of this set of measurements, indicating the likely tissue margin under the overall footprint of the probe. In an alterative approach to the above, in step 1 the initial four electrodes could be outermost, and the system could progressively work inwards with its selection of the tetrapolar subsets of electrodes, to iteratively home-in on the tissue margin.

[0194] Figure 24 illustrates an alternative methodology to that of Figure 23, and follows a broadly similar logic. The sub-figures again run from left to right and top to bottom.

[0195] In Figure 24 we introduce zig-zag and diagonal stripe shadings, representing lower and higher impedance regions respectively. In other words, the zig-zag shaded areas in Figure 24 may represent high conductivity and the diagonal stripe shaded areas may represent low conductivity (or vice versa). These regions can be determined based on the corresponding electrode voltage or potential difference between this corresponding electrode and a reference electrode. The size of the shading is not to any scale, and is shown here only to indicate the likely tissue conductivity under each electrode with a shading patch.

[0196] Once the tissue margin within the probe area has been identified, more sophisticated bioimpedance I tissue margin estimator measurements can be made around the margins, using quadrilateral tetrapolar electrode configurations as described above. Then, through superpositioning these quadrilaterals, a higher resolution region can be formed at and / or around the tissue margin.

[0197] Figure 25 illustrates another alternative methodology to that of Figure 23, and again follows a broadly similar logic. The sub-figures again run from left to right and top to bottom.

[0198] Here, it is noted that, in a multi-electrode probe (having more than four electrodes), such as the 12-electrode probe formation of Figure 9a, any two of the electrodes could be selected for current injection and extraction multiple times throughout scanning a plurality of quadrilateral locations. A more time-efficient way of doing this is to enable simultaneous measurement of all the electrodes’ voltages and only multiplex the current injection and extraction electrode pair. Thus, instead of having one set (pair) of a voltage-controlled current source and an instrumentation or differential amplifier, all of the electrodes’ voltages can be measured simultaneously. Every electrode can have their own voltage buffer, instrumentation or differential amplifier referencing the ground or reference voltage, then sampled with a simultaneous analogue-to-digital converter (ADC). With this electronics setup and using the 12-electrode probe example (Figure 9a), instead of having to multiplex to 495 locations (include concave quadrilaterals), the voltage controlled current source pair of electrodes can be multiplexed to only 66 locations to obtain the same amount of information, reducing the overall measurement time.

[0199] Figure 25 shows some possible electrode pair selections. With simultaneous electrodes’ voltages sampling, more than one electrode can be used to inject and / or extract the current. The voltage on the current injecting or extracting electrode can be buffered and supplied to another electrode as a bias point to alter the current travel path for a more sophisticated bioimpedance and tissue margin estimator measurement method.

[0200] Bioimpedance characterisation system electronics

[0201] With reference initially to Figure 26, the present characterisation system comprises (in addition to a probe 10 as described above, having at least four stimulator electrodes 1 a-1d at its distal end): a current source (e.g. voltage controlled current source, VCCS) 37 configured to generate a stimulation current; a voltage sensor (e.g. instrumentation amplifier, IA) 36; and a multiplexer (MUX) 39 coupled between the current source 37, the voltage sensor 36 and the stimulator electrodes and configured to changeably (e.g. cyclically) switch between a plurality of switching configurations, such as configurations 1-6 described above with reference to Figure 2, for the (or for any selected) tetrapolar group of four electrodes.

[0202] If more than four electrodes are present (e.g. as with the 12-electrode probe of Figure 9a), the multiplexer 39 is also operable to changeably select which subset of electrodes form the active tetrapolar group at any moment in time, e.g. by switching through a plurality of quadrilateral subsets such as those shown in Figure 9b. For each instantaneously-selected tetrapolar group of electrodes, the multiplexer 39 is further operable to changeably (e.g. cyclically) switch between the plurality of switching configurations, such as configurations 1 -6 described above with reference to Figure 2.

[0203] In each of the illustrated switching configurations, a first pair of the electrodes are connected to the current source 37, and a second pair of the electrodes are connected to the voltage sensor 36, the electrodes that constitute the first pair and the electrodes that constitute the second pair rapidly changing from one switching configuration to the next, e.g. in the manner shown in Figure 2. Optionally, the electrodes may initially be floating (electrically not connected to the system), then later connected to the corresponding nodes in the system.

[0204] Processor-controlled circuitry (e.g. microcontroller unit, MCU, or processor) 31 is also provided, configured to: control the switching configuration of the multiplexer 39 (by application of a control signal to the multiplexer); control the generation of the stimulation current by the current source 37 and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that, in use, the stimulation current flows through the tissue between the first pair of electrodes; control the measurement, by the voltage sensor 36, of the voltage between the second pair of electrodes of each switching configuration, across the tissue in use, and preferably the individual electrode voltage measurements as well; determine, based on the applied current and the measured voltage for each switching configuration, a measure of relative impedance of the tissue for each switching configuration; and supply such impedance measurements (or raw time-series voltage data) as output data to a data analyser 30 (e.g. PC or another suitable device) to enable realtime discrimination of the tissue to be performed, based on the impedance measurements. Considering the impedance of normal and cancerous brain tissues, the voltage level received at the sensing electrodes is typically in the range of a few millivolts, depending on the stimulation current used and the spacing of the electrodes.

[0205] To enable real-time intraoperative impedance characterisation, the processing time should be minimised. Therefore the control electronics (e.g. microcontroller unit (MCU) 31 ) should be capable of converting the analogue signal if an on-board analogue-to-digital converter 32 is used, and extracting frequency domain signals by using Fast Fourier Transform (FFT) or envelope detection. As saline solution has to be poured over the cortex of the brain frequently to avoid drying, an analogue multiplexer 39 is preferably used to switch between different measuring configurations to ensure rapid measurement time and consistent measurements recorded over each configuration.

[0206] - System architecture

[0207] In more detail, a block diagram of the electronics system of one example implementation is shown in Figure 26. It consists of four main parts:

[0208] 1 ) A compact handheld probe (probe 10 described above) with multi-pole electrodes 1 a-1 d etc. to perform versatile impedance characterisation configuration.

[0209] 2) A front-end printed circuit board (PCB) 35, or integrated circuit, which may be incorporated within the probe 10, or provided in a separate interface unit to which the probe 10 (specifically the electrodes thereof) is connected. In this example the front-end board 35 includes a voltage controlled current source 37 to generate the stimulation current; an instrumentation amplifier 36 to sense the voltage signal; and a multiplexer array 39 for switching the current source 37 and the instrumentation amplifier 36 between electrodes. Optionally a Direct Digital Synthesis (DDS) module may also be provided. In this example the front-end PCB 35 includes a digital-to-analogue converter (DAC) 33 which is configured to generate a voltage waveform and to supply the voltage waveform to the voltage controlled current source 37, and an analogue-to-digital converter (ADC) 32 configured to receive and sample a voltage signal from the instrumentation amplifier 36 and thereby generate digital voltage data. In other embodiments the current source may directly generate a stimulation current for driving the electrodes, with the current source being driven by a current waveform generator (e.g. a current DAC).

[0210] 3) An MCU platform 31 to control the ADC 32 and the DAC 33, and pre-process the data. This may also be incorporated within the probe 10, or provided in a separate interface unit to which the probe 10 (specifically the electrodes thereof) is connected. As illustrated, the MCU platform 31 may also include a general purpose input / output (GPIO) 40, a serial peripheral interface (SPI) 41 , a universal asynchronous receiver / transmitter (UART) 42, and an SD card 43. With suitable processors, the MCU platform 31 may include a GPIO 40, SPI 41 , universal flash storage (UFS), embedded multimedia card (eMMC), double data rata (DDR) RAM, Peripheral Component Interconnect Express (PCIe), field programmable gate array (FPGA), universal serial bus (USB), 5G cellular / private network connection, etc.

[0211] 4) A back-end host computer 30 (or other processing / visualisation device, such as, but not limited to, a tablet computer or smartphone) to reconstruct the data for visual display. In alternative embodiments this need not provide visualisation, and could just provide an indicator (e.g. an audible notification) instead.

[0212] A predefined stimulation waveform is stored in the MCU 31 and sent to the DAC 33 or the DDS waveform generator, to cause the voltage waveform to be generated. The generated voltage waveform is then converted to current source with required slew rate and bandwidth. This differential current is directed to the target electrode pair of the desired tetrapolar electrodes of the probe 10 by the multiplexer 39. The other pair of electrodes of the desired tetrapolar electrodes of the probe 10 are set to sense the voltage induced by the impedance of the tissue. This voltage signal is amplified by the instrumentation amplifier 36 and converted into the digital domain by the ADC 32. Then the measured waveform is processed using FFT to obtain the impedance measurement at each frequency of interest. In one example implementation we have made, the overall processing time is under 73 ms for 4096 data points in each configuration. Alternative block diagram examples of the electronics system are given in Figures 27 and 28. In Figure 27, only the current source (e.g. voltage controlled current source, VCCS) 37 is connected to the probe 10 via the multiplexer 39. Simultaneous measurement of all the electrodes’ voltages, in the manner described above with reference to Figure 25, is enabled via a voltage buffer / operational amplifier (BUF) incorporated in (or coupled to) the voltage sensor (instrumentation amplifier) 36.

[0213] Meanwhile, Figure 28 illustrates a system that does not use a multiplexer at all. Instead, the current source (VCCS) 37 and the voltage buffer / operational amplifier (BUF) of the voltage sensor 36 are connected to each of the electrodes 1 a-1 d etc., and switching is provided to disable the unused VCCS and BUF circuitries at each moment in time.

[0214] - Waveform characteristics

[0215] Advantageously, in the presently-preferred embodiments, the voltage waveform that is generated and supplied to the voltage controlled current source 37 (i.e. by the DAC 33 or DDS) may comprise a mix of a plurality of different frequencies.

[0216] As mentioned above, the predefined voltage waveform is converted to the stimulus current by the VCCS 37. The waveform may be designed to have substantially equal magnitude at the different frequencies of 1 kHz, 3kHz, 5kHz, 10kHz, 16kHz, 24kHz, 32kHz, 48kHz, 60kHz, and 80kHz, sampled at a frequency of 192kHz. The number of sinusoidal waveforms and their frequencies is not fixed, but the maximum measurable sinusoidal frequency should be half of the sampling frequency (Nyquist sampling criterion). Advantageously a multi-frequency waveform may be used such that multiple frequencies’ impedance can be rapidly extracted using FFT, greatly reducing the measurement time compared to a standard chirp technique. This is because a standard chirp technique would require frequency sweeping to take place. However, the present technique can employ multiple frequencies simultaneously, enabling the impedance result to be rapidly obtained, after only a small number (e.g. a few) of electrode switching cycles. The above frequencies have been chosen to keep the overall amplitude of the waveform low to maximise the dynamic range of the system and with a low slew rate to ease requirements for electronic components. The chosen frequency range may be limited due to component limitations. In addition, the phase of an individual sinusoidal waveform can be changed to further reduce the overall amplitude of the waveform.

[0217] The predefined waveform may be repeated multiple times in the measurement period, such that the calculated impedance would be an averaged overall repetition of the waveform, resulting in lower noise measurement.

[0218] - Operating method

[0219] Figure 29 provides a flow diagram illustrating one possible manner of operation of the above-described bioimpedance characterisation system and measurement probe, for a multi-electrode probe having more than four electrodes - for example, an array of 12 electrodes as shown in Figure 2, operated for example with 13 tetrapolar multiplexed subsets of electrodes as shown in Figure 9b. The constituent steps S1 -S13 are as follows:

[0220] Step SI : The probe 10 (more particularly, the electrodes 1 a-1 d etc. thereof) are placed in an area of interest on the tissue under examination.

[0221] Step S2: The user provides user input (via a user interface) that is received by the processor-controlled circuitry, to select an area of interest (or to select a respective tetrapolar subset of electrodes) within the probed area, with which to begin bioimpedance measurements.

[0222] Step S3: The processor-controlled circuitry determines whether valid user input was received in step S2.

[0223] Step S4: If valid used input was received in step S2, the processor- controlled circuitry identifies the respective tetrapolar subset of electrodes to be used to perform initial bioimpedance measurement.

[0224] Step S5: If valid used input was not received in step S2, the processor- controlled circuitry selects a tetrapolar subset of electrodes closest to the centre of the probe to be used to perform initial bioimpedance measurement.

[0225] Step S6: The processor-controlled circuitry configures the multiplexers to select the tetrapolar subset of electrodes identified in step S4 or step S5 (e.g. the centremost tetrapolar subset of electrodes shown in Figure 9b) with which to perform bioimpedance measurement.

[0226] Step S7: The processor-controlled circuity operates the multiplexers to performing a switching routine, to sequentially apply a series of switching configurations (e.g. configurations 1 -6 of Figure 2) to the tetrapolar subset of electrodes selected in step S6. Bioimpedance values and additional “tissue margin estimator” or “tissue margin likelihood” values may be measured, and individual electrode voltages may also be obtained.

[0227] Step S8: Data processing may then be performed by the data analyser 30, e.g. to analyse tissue margins, identify tissue types, and to estimate the physiological state of the tissue.

[0228] Step S9: Intermediate measurement results / maps are output to the user.

[0229] Step S10: The user is presented with a query as to whether they are satisfied with the results so far (e.g. bioimpedance values, and / or tissue margin estimator map content, and / or its resolution).

[0230] Step S11 : If, in step S10, the user is not satisfied with the results so far, they process continues to step S11 , to perform measuring at all possible measuring locations under the footprint of the probe (or a subset of possible measuring locations, based for example on the previously identified location and by selecting the neighbouring ones to continue, as described with reference to Figures 23 and 24 above), using different tetrapolar subsets of electrodes from within the overall set of electrodes of the probe.

[0231] Step S12: If, in step S10, the user is satisfied with the results so far, or once, in step S11 , measuring has been completed at all possible measuring locations under the footprint of the probe, further data processing can then be performed to produce the final measurement results or maps.

[0232] Step S13: The final measurement results or maps can then be output to the user.

[0233] The process may be repeated as and when the probe is moved to a new region of tissue to be examined. Modifications and alternatives

[0234] Detailed embodiments and some possible alternatives have been described above. As those skilled in the art will appreciate, a number of modifications and further alternatives can be made to the above embodiments whilst still benefiting from the inventions embodied therein.

[0235] Notably, in the above embodiments, the current source is primarily described as being a voltage controlled current source, driven by a voltage waveform (generated by a voltage waveform generator). However, in alternative embodiments the current source may directly generate a stimulation current for driving the electrodes, with the current source being driven by a current waveform generator. The current waveform generator may for example comprise a digital-to-analogue converter configured to receive a predefined stimulation waveform from which the current waveform is generated. As in the case of the voltage waveform, such a current waveform may comprise a mix of a plurality of different frequencies, and be of substantially equal magnitude at each of the plurality of different frequencies.

[0236] Optionally the above-described probe 10 may be wireless, facilitating unencumbered manipulation of the probe by the surgeon. However, it may alternatively be connected to the control and analysis system by a cable.

[0237] The present disclosure is not limited to any specific tissue type. That is to say, it is tissue type agnostic.

[0238] Optionally the above-described probe 10 may further comprise a pressure sensor, for measuring the contact pressure between the electrodes 1 a-1 d etc. and the tissue being tested. Contact pressure has been found to have an effect on the impedance measurements, and consequently measurements of the contact pressure obtained using the pressure sensor may be used to normalise the impedance measurements, thereby improving the accuracy of the impedance measurements.

[0239] Optionally the above-described probe 10 may further comprise, at the distal end of the elongate body, a blood oxygen sensor (e.g. an optical sensor to measure peripheral oxygen saturation (SpC )), for measuring the blood oxygen level of the tissue being tested. This may be used to notify the surgeon that a blood vessel is present within or near the tissue being tested. Accordingly, the surgeon may take suitable measures before cutting out a tumour from the tissue in question, for example.

[0240] Optionally the above-described probe 10 may be incorporate an accelerometer or inertial sensor to sense the angle of inclination (i.e. tilt) of the probe. This may be used to alert the surgeon if they inadvertently change the angle of inclination of the probe to such an extent that that it risks causing injury to the patient, for example.

[0241] Optionally the surface of the probe 10 on which the electrodes are mounted can be continuous in a loop, e.g. around a needle or lead. In use, the needle or lead can be inserted into tissue while mapping the surrounding tissue regions in contact with the needle or lead.

[0242] The present disclosure is not limited to the end of the electrode probe 10 having a circular form. The array of electrodes can be in any shape, being placed on the surface or inserted into the tissue.

[0243] Optionally the electrodes can be inserted into a surface mounted on an arm, where the movement of this arm is monitored. The results can then be transformed into volumetric or surface layer scan results.

[0244] In an alternative setup, tissue identification may be performed through measuring the tissue response to surrounding electrical (current / voltage), mechanical (force), and chemical (electrode material) stimulation. The non-electrical stimulating electrodes could be landed on the tissue surface first, then later the electrical stimulating electrodes can be applied. Alternatively the non-electrical stimulating electrodes and the virtual ground or ground electrode (current return or sink electrode) could be placed on the tissue surface first, then later the current injecting (source) electrode can be applied. The voltage measuring electrodes can then capture the tissue response to the mechanical touchdown of the electrodes, and chemical effects (e.g., additional electrolyte forming at the tissue surface(s) touching the stimulating electrode(s), interaction between tissue and electrode material, etc.).

[0245] References

[0246]

[0001] A. Bs, A. V. Gk, S. Rao, M. Beniwal, and H. J. Pandya, “Electrical phenotyping of human brain tissues: An automated system for tumor delineation,” IEEE Access, vol. 10, pp. 17 908-17 919, 2022.

[0247] [2] C. Gabriel, “Compilation of the dielectric properties of body tissues at RF and microwave frequencies,” Occupational and environmental health directorate, Radiofrequency Radiation Division, Brooks Air Force Base, Tech. Rep., Jan 1996. [Online], https: / / itis.swiss / virtual-population / tissue-properties / database / dielectric- properties /

[0248] [3] S. S. Wong, J. Ekanayake, Y. Liu, and T. G. Constandinou, “An impedance probing system for real-time intraoperative brain tumour tissue discrimination,” in 2019 IEEE Biomedical Circuits and Systems Conference (BioCAS), 2019, pp. 1-4.

[0249] [4] S. B. Rutkove, “Electrical impedance myography: Background, current state, and future directions,” Muscle nerve, vol. 40, no. 6, pp. 936-946, 2009. [Online], https: / / doi.Org / 10.1002 / mus.21362

[0250] [5] S. Grimnes and 0rjan G Martinsen, “Sources of error in tetrapolar impedance measurements on biomaterials and other ionic conductors,” Journal of Physics D: Applied Physics, vol. 40, no. 1 , p. 9, Dec 2006. [Online], https: / / dx.doi.Org / 10.1088 / 0022-3727 / 40 / 1 / S02

Claims

CLAIMS1 . Apparatus for discriminating between tumour tissue and non-tumorous tissue in real-time, the apparatus comprising: a handheld bioimpedance probe having a plurality of electrodes mounted thereon, wherein the plurality of electrodes are greater than four in number, the electrodes being arranged to be held against a region of tissue in use; a current source configured to generate a stimulation current; a voltage sensor; one or more multiplexers coupled between the current source, the voltage sensor and the electrodes and configured to changeably select a subset of four electrodes from the plurality of electrodes, and, for each selected subset of four electrodes, to switch between a plurality of switching configurations, wherein, in each switching configuration, a first pair of the selected subset of four electrodes are connected to the current source, and a second pair of the selected subset of four electrodes are connected to the voltage sensor, the electrodes that constitute the first pair and the electrodes that constitute the second pair changing from one switching configuration to the next; and processor-controlled circuitry configured to: control the selection of the subset of four electrodes by the one or more multiplexers; control the switching configuration of the one or more multiplexers with respect to the selected subset of four electrodes; control the generation of the stimulation current by the current source and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that, in use, the stimulation current flows through the tissue between the first pair of electrodes; control the measurement, by the voltage sensor, of the voltage between the second pair of electrodes of each switching configuration, across the tissue in use; determine, based on the applied current and the measured voltages for each switching configuration, a measure of relative impedance of the tissue for each switching configuration; andsupply such impedance measurements as output data to a data analyser to enable real-time discrimination of the tissue to be performed, based on the impedance measurements.

2. The apparatus according to claim 1 , wherein each selected subset of four electrodes forms a quadrilateral shape.

3. The apparatus according to claim 2, wherein each selected subset of four electrodes forms a convex quadrilateral shape.

4. The apparatus according to claim 2 or claim 3 wherein, for each selected subset of four electrodes, the switching configurations are six in number, comprising four bioimpedance measuring configurations, in each of which the first and second pairs of electrodes are on opposite sides of the quadrilateral to one another, and two configurations in which the first and second pairs of electrodes diagonally cross one another.

5. The apparatus according to any preceding claim, wherein each of the plurality of electrodes is connected to four different individually-controlled bidirectional analogue multiplexers.

6. The apparatus according to any preceding claim, wherein the plurality of electrodes are 12 in number.

7. The apparatus according to any preceding claim, wherein at least some of the selected subsets of four electrodes overlap one another.

8. The apparatus according to any preceding claim, wherein the data analyser is configured to generate a bioimpedance value for each selected subset of four electrodes.

9. The apparatus according to claim 8 when dependent on claim 2, wherein the data analyser is further configured to generate a bioimpedance map bysuperimposing quadrilaterals, each quadrilateral having a shape and position corresponding to the shape and position of the respective subset of four electrodes used to generate a respective bioimpedance value, and coloured or shaded to indicate the respective bioimpedance value.

10. The apparatus according to claim 9, wherein the data analyser is further configured to generate further polygons within the bioimpedance map where quadrilaterals overlap.

11. The apparatus according to claim 10, wherein the further polygons are coloured or shaded by the data analyser to represent an average of the bioimpedance values of the respective overlapping quadrilaterals.

12. The apparatus according to any of claims 9 to 11 , wherein the data analyser is further configured to apply a low-pass filter to the bioimpedance map, for example by convolving it with a Gaussian blur kernel.

13. The apparatus according to any preceding claim, wherein the processor- controlled circuity is further configured to obtain a plurality of voltage measurements from each switching configuration.

14. The apparatus according to any preceding claim, wherein, in each switching configuration, the processor-controlled circuitry is configured to control the measurement, by the voltage sensor, of the individual voltages of the second pair of electrodes, as well as the measurement of the voltage between the second pair of electrodes.

15. The apparatus according to any preceding claim, wherein the data analyser is further configured to iteratively determine the path of the tissue margin across the footprint of the probe.

16. Apparatus for discriminating between tumour tissue and non-tumorous tissue in real-time, the apparatus comprising:a handheld bioimpedance probe having four or more electrodes mounted thereon, the electrodes being arranged to be held against a region of tissue in use; a current source configured to generate a stimulation current; a voltage sensor; a multiplexer coupled between the current source, the voltage sensor and the electrodes and configured to changeably switch between a plurality of switching configurations, wherein, in each switching configuration, a first pair of the electrodes are connected to the current source, and a second pair of the electrodes are connected to the voltage sensor, the electrodes that constitute the first pair and the electrodes that constitute the second pair changing from one switching configuration to the next; and processor-controlled circuitry configured to: control the switching configuration of the multiplexer; control the generation of the stimulation current by the current source and the application of the stimulation current to the first pair of electrodes of each switching configuration, such that, in use, the stimulation current flows through the tissue between the first pair of electrodes; control the measurement, by the voltage sensor, of the voltage between the second pair of electrodes of each switching configuration, across the tissue in use, and also the measurement of the individual voltages of the second pair of electrodes; determine, based on the applied current and the measured voltages for each switching configuration, a measure of relative impedance of the tissue for each switching configuration; and supply such impedance measurements as output data to a data analyser to enable real-time discrimination of the tissue to be performed, based on the impedance measurements.

17. The apparatus according to claim 16 wherein the switching configurations are six in number, comprising four bioimpedance measuring configurations, in each of which the first and second pairs of electrodes are substantially opposite one another, and two configurations in which the first and second pairs of electrodes diagonally cross one another.

18. The apparatus according to any preceding claim, wherein the multiplexer is configured to cyclically switch between each of the plurality of switching configurations.

19. The apparatus according to any preceding claim, wherein, in each switching configuration, the processor-controlled circuitry is configured to control simultaneous measurement of the voltages measured.

20. The apparatus according to any preceding claim, wherein the processor- controlled circuitry comprises a microcontroller.21 . The apparatus according to any preceding claim, wherein the data analyser comprises a personal computer, a tablet computer or a smartphone.

22. A method of discriminating between tumour tissue and non-tumorous tissue in real-time, using the apparatus according to any preceding claim.

Images