SE2350149A1

SE2350149A1 - Methods and apparatuses for application of tumor treating fields based on high-frequency real-time tumor positioning

Info

Publication number: SE2350149A1
Application number: SE2350149A
Authority: SE
Inventors: Fredrik Juto
Original assignee: Force Oncology Ab
Priority date: 2023-02-15
Filing date: 2023-02-15
Publication date: 2024-08-16
Also published as: WO2024170522A1

Abstract

Abstract [0160] Methods and apparatuses are disclosed for directing TTFields at tumors based on real time tumor positioning, in particular using a model of a tumor as it moves within the body and measurements related to the respiratory cycle sampled at more than one hertz.

Description

1 / 39 METHODS AND APPARATUSES FOR APPLICATION OF TUMOR TREATING FIELDS BASED ON HIGH-FREQUENCY REAL-TIME TUMOR POSITIONING.

Technical Field

[0001] The invention relates to systems for delivering Tumor Treating Fields (TTFields) to tumors. Background

[0002] 'lTFields are typically applied as oscillating electric fields in the 100 kHz-300 kHz range to treat tumors, with a field strength of at least 1 V/cm (measured as peak voltage, or RMS). Using sensor- derived information to detect a change in the region of interest of the subject's body, in order to improve the application of 'lTFields through activating electrodes in different locations on a body is known from e.g., US-2022305277-A1 Methods and apparatuses for detecting and responding to changes in a subject. The sampling rates disclosed in that publication are not high enough to enable implementation of the invention disclosed herein, but is instead limited to tracking and responding to slower processes in the body.

Summary of the lnvention

[0003] ln some aspects, the techniques described herein relate to a method of applying 'lTFields to a tumor in a subject's body, where a pair of electrodes first target a first tumor position (or location) with an electric field; sample measurements are made of one or more sensors 195; one or more metrics are estimated based on the sampled measurements, where the metrics include a metric related to the subject's respiratory phase; a second tumor position (or location) is estimated based on the metrics; a second pair of electrodes target the second position with an electric field.

[0004] ln some embodiments, measurements are sampled at rate higher than 1 Hz.

[0005] ln some embodiments, measurements are sampled at rate higher than 10 Hz.

[0006] ln some embodiments, the metrics include a measure of respiratory volume.

[0007] ln some embodiments, when targeting a position with an electric field, the field is applied in sequence between two pairs of electrodes.

[0008] ln some embodiments, the metrics include a measure of measurement uncertainty.

[0009] ln some embodiments, the estimation of a second position considers measurement uncertainty, or any missing data during the estimation.

[0010] ln some embodiments, the metrics include respiratory period, and the electric fields are applied in bursts of certain lengths separated by certain lengths, and the lengths of the bursts and that of any separations fit an integer number of times in the respiratory period.

[0011] ln some embodiments, the burst lengths are of a pre-specified length, and the separation lengths are changed to fit the bursts and any separations an integer number of times in the respiratory period. 2/39

[0012] ln some embodiments, the metrics include an inhalation period and an exhalation period, and the bursts are adapted to synchronize with the inhalation period and exhalation period, by varying the burst lengths and/or the burst separations.

[0013] ln some embodiments, an estimated second position includes a surrounding clinical tumor volume.

[0014] ln some embodiments, initial measurements that have been made are used when estimating a second tumor position in a body.

[0015] ln some embodiments, initial measurements are made using at least one of MRI, CT, or ultrasound.

[0016] ln some embodiments, initial measurements include sampling measurements that include a signal for respiratory phase.

[0017] ln some embodiments, some or all initial measurements are made while applying electric fields. [0018] The details of one or more embodiments of the invention are set forth in the accompanying drawings and the description below. Other features and advantages of the invention will be apparent from the descriptions and drawings, and from the modes and claims.

Brief Description of the Drawings

[0019] Fig. 1 is a flowchart, in accordance with one embodiment, illustrating a first method 200 of building a tumor position model 180 and then using it with real-time measurements to update TTFields treatment targeting during treatment.

[0020] Fig. 2 is an illustration, in accordance with one embodiment, of a configuration on a subject's body with four electrodes 130 each containing a plurality of electrode elements 184.

[0021] Fig. 3 is a flowchart, in accordance with one embodiment, illustrating a second method 220 for delivering 'lTFields to a subject's body, including building a treatment pattern 183.

[0022] Fig. 4 is an illustration, in accordance with one embodiment, of a treatment pattern 183 of bursts 181 for delivery of TTFields through three different field positions (185a, 185b, 185c) as a tumor is tracked and targeted.

[0023] Fig. 5 is an illustration, in accordance with one embodiment, similar to Fig. 4 (difference highlighted in bold), but where a controller (not shown) controlling the treatment pattern 183 ensures that bursts for each position are delivered in strictly alternating directions (A and B) considering only the bursts delivered through each position 185a, 185b, or 185c.

[0024] Fig. 6 is an illustration, in accordance with one embodiment, of a tumor trajectory 187 within a lung 188 and close to a diaphragm 189, with target field positions 185a, 185b, 185c.

[0025] Fig. 7 is an illustration, in accordance with one embodiment, for a tidal volume over time for a breathing rate of about 45 breaths per minute, of a sampling rate of 10 Hz i.e., a sampling interval 191 of 0.1 seconds (shown between seconds 2 and 3). 3/39

[0026] Fig. 8 is an illustration, in accordance with one embodiment, of a predicted tumor trajectory over time 194, and a treatment pattern 183 of bursts directed through two field positions 185 (185a and 185b), that are mapped to match the predicted tumor trajectory over time 194.

[0027] Fig. 9 is an illustration, in accordance with one embodiment, of synchronizing the burst length 193 and burst spacing 186 with the estimated or predicted respiratory cycle period 192 (192a is the first respiratory cycle period, 192b is the next respiratory cycle period, etc.), such that an integer number of bursts (with associated burst spacings) fit into each respiratory cycle.

[0028] Fig. 10 is an illustration, in accordance with one embodiment, of eight electrodes 130 (130a-h) attached to the torso of a human body, and a tumor trajectory 187 inside the body.

[0029] Fig. 11 is an illustrative flowchart, in accordance with one embodiment, of a tracking 218 step, generating a treatment pattern 183 based on measurement inputs 178, from one or more sensors 195 (not shown in Fig. 11, see Fig. 15).

[0030] Fig. 12 is an illustration, in accordance with one embodiment, of a method step 205 to deliver TTFields according to a treatment pattern 183 or based on an estimated CTV (182) position.

[0031] Fig. 13 is a flowchart, in accordance with one embodiment, illustrating a method 220 of delivering 'lTFields to a patient, where initial measurements 210 are first made, and used to perform electrode location mapping 219, where suitable configurations of two or more electrodes 130 for placement on the human body are determined.

[0032] Fig. 14 is a flowchart, in accordance with one embodiment, illustrating a method 221 of delivering 'lTFields to a patient, where first electrode location mapping 219 is performed, e.g., using a default electrode mapping.

[0033] Fig. 15 is an illustration, in accordance with one embodiment, of a system 100 for delivering TTFields, containing a TTFields generator 198 and a controller 196.

[0034] Fig. 16 is an illustration, in accordance with one embodiment, of a TTFields generator 198 and a controller 196.

[0035] Fig. 17 is an illustration, in accordance with one embodiment, of a method of electrode location mapping 219.

[0036] Fig. 18 a) is an illustration, in accordance with one embodiment, of a personal computer 304 for executing a program implementing methods disclosed herein.

[0037] Fig. 18 b) is an illustration, in accordance with one embodiment, of a server 302 for executing a program implementing methods disclosed herein, interfacing with a web browser 306.

[0038] Fig. 19 is a flowchart, in accordance with one embodiment, illustrating a method 225 of creating an electrode location map. 4/ 39 Detailed description of the invention

[0039] The various embodiments presented herein aim to solve technical problems related to systems for delivering Tumor Treating Fields (TTFields) to patients with tumors. 'lTFields are typically applied as electric fields oscillating in the 100 kHz - 300 kHz range, or sometimes 50 kHz - 500 kHz, with field strengths of at least 1 V/cm, delivered through electrodes, often capacitive and electrically insulated (perhaps each comprising a single electrode element or an array of electrode elements) affixed to the body. For mesothelioma and non-small cell lung cancer (NSCLC), 150 kHz fields are typically used. Higher field strengths and power densities (e.g., at least 1 mW/cm3 or at least 2.4 mW/cm3) are generally better at interfering with dividing cancer cells and reducing or inhibiting tumor growth, potentially reducing the size of a tumor.

[0040] ln some embodiments, a 'lTFields signal 153 contains essentially only one frequency component, where the frequency component is in the 100 kHz to 300 kHz range, or in some embodiments 50 kHz - 500 kHz range. ln some embodiments, a 'lTFields signal 153 contains essentially only two frequency components, where the frequency components are in the 50 kHz to 500 kHz range, e.g., 200 kHz and 300 kHz. ln some embodiments, a TTFields signal 153 contains three or more significant frequency components in the 50 kHz to 500 kHz range. Having multiple frequency components can be beneficial, as e.g., tumor cells of different sizes can be effectively targeted, tumor motility can be reduced, increased permeability of a blood-brain-barrier, or yet other benefits. ln embodiments with multiple frequencies (e.g., two, three or more), these would superimpose to form a single waveform, or can be applied sequentially in time during overlapping or mutually exclusive time periods. ln some embodiments, only one frequency component has a field strength of at least 1 V/cm.

[0041] The electric fields that, among other things, interfere with the mitosis of fast-dividing cancer cells should ideally be targeted toward the volume of tissue where a tumor is located, in order to bring the treatment intensity to a level suitable for inhibiting growth in a tumor, or reducing its size. Especially for tumors located in the torso or abdomen, the location of a tumor can move significantly during the respiratory cycle and due to other factors. By tracking or estimating the position of a tumor in real-time during the respiratory cycle, an apparatus for delivering 'lTFields can adapt the applied treatment to consider an estimate for its position and volume.

[0042] When administrating TTFields, a Clinical Treatment Volume (CTV) 182 can be defined as a volume with a 20 mm margin around a Gross Tumor Volume (GTV). The CTV thus contains tissue that is presumed to contain tumorous cells. This is described, for example, in European Patent No. EP3368152B1.

[0043] The Operating Manual for the Optune Lua, available from Novocure Ltd., Bailiwick ofJersey, which is used for delivering TTFields to the torso, advises moving transducer arrays 2 cm when changing /39 arrays to avoid redness. This example illustrates that accuracy beneficial for effective 'lTFields treatment is on the order of 20 mm.

[0044] ln some embodiments, higher precision can be advantageous, e.g., within 10 mm, in order to efficiently treat cells proliferating from a tumor or CTV 182 even at the limits of the interval of precision. [0045] lt is known that a tumor in the torso can move with the respiratory cycle. For example, Yan Wang et al, Assessment of Respiration-Induced Motion and Its Impact on Treatment Outcome for Lung Cancer, Biomed Res Int 2013, found in a group of free-breathing subjects that the GTV centroid moved more in the Superior-lnferior (SI) direction, than in the Anterior-Posterior (AP) direction or the Left-Right (LR) direction, though the maximum expected motion with a 95% percentile about the centroid was about 13 mm. The closer a tumor is to a diaphragm, the more mobile the tumor typically is. Similar motion is known for abdominal and gastric tumors due to respiration, e.g., Bussels 2003 found the pancreas and liver each moved about 24 mm in the cranio-caudal (superior-inferior) direction, and the kidneys about 17 mm, and Bussels 2003 suggested these movements be taken into account when planning radiation therapy.

[0046] The precise tracking of tumors in the torso or abdomen is of particular importance when targeting tumors with various treatments, such as particle beams or other types of radiation, and various technologies have been developed for real-time tracking and positioning. ln these applications, patients are typically still and breathing calmly in a clinical setting and a high degree of accuracy and precision in tumor location tracking is required for successful treatment. ln some applications of radiation therapy, respiratory 'gating' is used, where a target is irradiated only when it moves in a predefined position in the respiratory cycle (Abbas 2014), typically at the end of exhalation as it is the most stable with limited movement. For liver and pancreatic cancers, it has been found that gating near exhale (T40-T60) reduces the range of motion with a factor of 10.

[0047] Though the diaphragm will typically move about 1.5 cm during quiet breathing, it can move as much as 10 cm during exercise, and during deep breathing, the diaphragm can move as far as 13 cm. Similarly, Suramo 1984 found 2.5 cm liver movement in normal breathing, but 5.5 cm movement in maximum breathing. Pham 2014 found eight publications showing deep breathing led to 10 to 40 mm kidney movement. These displacements can be expected to apply to some potential 'lTFields recipients at least some of the time and are of a magnitude where the effectiveness of 'lTFields therapy can potentially be affected, as a tumor might spend part of the respiratory cycle outside of the effectively targeted area. Also, depending on the frequency of respiration, and the direction-changing frequency and phase of any applied TTFields, a situation can potentially arise for a longer or shorter time where one of the applied directions is predominantly, or at least more likely to be, missed by a tumor as it moves. To illustrate, if a displacement associated with the respiratory cycle would be on the order of 100 mm, and the desired accuracy is 20 mm, then the estimation accuracy for a tumor's position (e.g., a 6/ 39 centroid of a GTV or a CTV) can be relatively coarse (e.g., +/- ca 20 %) and still deliver improved treatment benefit if used for targeting TTFields. Since it is known that abdominal organs move in unison, but with varying amplitudes, it can be expected that other organs in the torso and abdomen would show similar significant movements during the respiratory cycle to the organs mentioned above, and tumors located in them would likewise beneficially be treating using embodiments of the present invention. [0048] The lungs and e.g., several organs of the abdomen (e.g., liver, kidneys, pancreas, also in the mesothelium) can be within the subject's rib cage, which can provide a firm reference frame for any attached electrodes 130, as the rib cage has a smaller range of movement than e.g., the abdominal area, where an attached electrode 130 can be expected to move along with the tissue to a larger extent during e.g., the respiratory cycle.

[0049] ln addition to potential therapeutic improvements, targeting treatment precisely at a tumor in some embodiments has other benefits, such as e.g., reducing the system's power draw as energy is not expended in areas where no tumors are known to be located, and e.g., reducing any undesirable side effects such as the inhibition of normal cell division in non-cancerous cells.

[0050] A further potential benefit from the precise targeting of a tumor's location, is that the mapping of electrode arrays on the body can be optimized, such that electrodes are only mounted where they are needed, reducing any practical problems related to wearing them.

[0051] lf the tumor tracking is only partially effective, such that it e.g., estimates a tumor's CTV's 182 position with desired accuracy for only a part of the respiratory cycle and uses that information for targeting, or is unable to target a tumor during part of its trajectory e.g., due to the locations of electrodes 130 attached to the subject's body, but that part of the respiratory cycle is longer than the time that a tumor would happen to be in a statically targeted volume (where no attempt is made to change the treatment delivery to accommodate physiological changes associated with the respiratory cycle or other bodily processes that can be measured) in the body, then an improvement in treatment outcomes can be expected.

[0052] Some embodiments of the present invention estimate the position of a tumor (or CTV 182) by sampling one or more sensors, using the sample measures to estimate respiratory metrics (e.g., the stage or phase in the respiratory cycle of the patient and one or more metrics associated with respiratory volume (e.g., tidal volume, residual volume, end-expiratory lung volume, end-inspiratory lung volume, etc.), and use these metrics with prior knowledge about the tumor (its position and trajectory in relation to respiratory metrics, in some embodiments differentiated between inhalation and expiration, or in some embodiments its position and trajectory in relation to sensor measurements) to estimate its position. The estimated position is then used to target a volume (e.g., CTV 182) around the estimated position with TTFields. Techniques for targeting 'lTFields to a target volume are previously known, e.g., by developing maps for where electrodes should be applied prior to treatment, and 7/39 adjusting the fields delivered through the electrodes after they have been applied by selectively applying the 'lTFields signals.

[0053] To implement this approach, in some embodiments, initial measurements are first made and used to build a model of a tumor's position and movements in the torso, abdomen or other part of the body with respect to respiratory phase and volume estimates (Fig. 1) and this model is then combined with real-time measurements related to respiratory phase and volume gathered during the application of '|'|'Fields, to build a current estimate of the position of the tumor.

[0054] ln some embodiments, a default location mapping and configuration of one or more sets of electrodes 130 are applied to the body. ln such embodiments, these electrode 130 locations may in practice be determined as a first step.

[0055] ln some embodiments, the location maps for electrodes 130 on a subject's body are determined based on the measured positions and other characteristics of a tumor and/or its CTV, e.g., by magnetic resonance imaging (MRI), computer tomography (CT), X-ray imaging, ultrasound or similar, such that a full tumor trajectory 187, or a part thereof, e.g., 80 % or more, is expected to be targetable with TTFields through the various electrodes 130. ln such embodiments, the location of the electrodes 130 (e.g., a location map) can be a step subsequent to the initial measurements as described below. ln some embodiments, a part of the full tumor trajectory 187 is expected to be targetable at least 80 % of time during respiration. A location map, in some embodiments in the form of one or more printed maps on paper, in some embodiments displayed on a screen, in some embodiments generated as part of e.g., a piece of clothing containing electrodes, can contain identification of suitable locations on a subject's body where electrodes 130 are to be affixed, perhaps guided by specific anatomical reference points or previously affixed fiduciary markers, as well as information about the type and configuration of electrodes 130 that are identified as suitable for each suitable location, such that a caretaker, a subject or other person can use the location map to apply suitable electrodes to a subject's body. (A wearer of equipment in accordance with embodiments as disclosed herein would typically be a patient, or a subject, as the terms are used interchangeably herein).

[0056] Location maps for electrodes 130 can be determined and implemented in several ways. For example, U.S. Patent 7146210 BZ discloses an invention where ”the apparatus applies the fields that have maximal effect on the tumor and minimal effect on all other tissues by adjusting both the field generator output characteristics and by optimal positioning of the insulated electrodes or isolects on the patient's body." The prior art however does not consider e.g., targeting a tumor as it moves in real time with the respiratory cycle.

[0057] ln some embodiments, a set of positions along a tumor trajectory 187 are identified such that application of 'lTFields to this set of positions along a trajectory would permit effective therapy (e.g., targeted to within 20 mm distance from a tumor centroid or of a GTV, or in some embodiments, within 8/39 mm) along the whole trajectory or a part of the trajectory. Permitting effective targeting of a tumor or CTV 182 throughout a complete tumor trajectory 187 can be beneficial by increasing the probability of successful treatment, e.g., offering tumor growth control. ln some embodiments, a location map for electrodes 130 permit targeting of a tumor movement of up to 13 cm. ln some embodiments, an electrode location map can permit targeting of a more limited part of a tumor trajectory 187, e.g., up to 5 cm, or up to 10 cm, or up to 50 % of the full length of a tumor trajectory 187, as a tradeoff with improved comfort. ln some such embodiments, a patient, knowing that no physical exercise is planned, can chose to wear such a more limited embodiment.

[0058] ln some embodiments, multiple sets of possible positions along a tumor trajectory 187 are identified, and a selection between them is made based on the criteria presented herein. When comparing estimated criteria for multiple points along a trajectory, the criteria numbers for the different positions can be averaged, or a weighted average based on expected time along each part of the trajectory, can be used to determine a collective number for the set of positions such that different sets can easily be compared.

[0059] ln some embodiments, suitable electrode locations can be identified by determining, in some embodiments with the use of a 3D model of a subject's body, based on e.g., MRI, CT, X-ray imaging, ultrasound or other technique, for each position along a tumor trajectory, one or more straight lines between two electrodes on either side of a subject's body, through the tissue and through a tumor (or CTV 182) position. ln other embodiments, more advanced modelling of the electrical characteristics of tissues within a subject's body can be performed to identify one or more suitable electrode locations that can be expected to deliver optimal electrical field strength in the targeted volume. Examples of such more advanced modelling techniques include calculating an estimated field strength in a tumor given the different tissues in a model of the body and pairs simulated electrode locations, or similarly calculating estimated power density in the tumor. With these calculated estimates of field strength or power density, when given a choice between several possible pairs of electrodes to apply an electric field, the pairs with the highest field strength or highest power density, respectively, can be selected and their locations identified as the most suitable electrode positions. The above examples illustrate that there are several possible criteria for identifying suitable electrode locations, where some criteria consist of calculating a number containing a measure of quality of treatment targeting effectiveness, and then when choosing between several possible electrode locations for targeting the same tumor position, selecting the one with the highest number for the criteria used. ln some embodiments, the number for a criterion is calculated as the average of, or the minimum or maximum of, or some other combination of the numbers for two pairs of electrodes that can be used to apply roughly perpendicular fields to the same tumor position (as both pairs could then e.g., be used to treat a tumor with fields from different directions). 9/39

[0060] ln some embodiments electrode locations are identified considering practical constraints, e.g., one or more of not obstructing a patient's movements, being uncomfortable, constraints due to modular design, or other anatomica| constraints, on the placement of both of the electrodes concerned in a pair, and also in light of any additional electrode locations that are needed for targeting additional positions. ln embodiments where multiple possible electrode pairs have been identified, one for implementation can be selected based inter alia on these practical constraints, such that electrode or electrode elements are not considered for placement in the indicated undesirable areas. ln some embodiments, the practical constraints are informed by the requests of the particular patient for which the location map is getting planned.

[0061] To implement the location maps for electrodes 130, in some embodiments, locations for multiple discrete electrodes 130 are identified, and in some embodiments, these discrete electrodes 130 are alike and of a standard configuration (e.g., each with 3-15 electrode elements 184, or nine electrode elements 184, spaced with the same geometric positions). ln some embodiments, electrode elements 184 are of uniform dimensions or of the same design. ln some embodiments, electrodes 130 can be configured to have e.g., different dimensions, number of electrode elements, etc. based on the electrode locations that have been identified as suitable. ln some embodiments, such configurable electrodes are of a modular design, such that e.g., an electrode 130 can be configured to have specific number of rows of electrode elements 184, each row of a standard width (which can e.g., be advantageous from a manufacturing point of view by limiting the number of possible configurations, or permitting the adaption of a large module by removing elements, e.g., a row, from it to get a smaller module). ln some embodiments, such a modular electrode 130 is affixed to the body such that the number of rows specified adapts its length in a superior-inferior (SI) direction.

[0062] ln some embodiments, one or more electrodes 130 is configured to use a set of electrode elements 184 in the one or more electrodes 130 to constitute a 'virtual' electrode used as part of an electrode pair. One electrode 130 can be used as part of several pairs of electrodes that are identified as suitable for targeting different position.

[0063] ln some embodiments, one or more electrodes 130 of modular design would be used with 'virtual' electrodes by using sets of electrodes elements 184 that each contain all electrode elements 184 on each of one or more rows.

[0064] ln some embodiments, a combination of some or all of the techniques mentioned above for configuring the electrodes 130 is used.

[0065] ln some embodiments, multiple trajectories (187) for different tumors are considered, and a location map for electrodes 130 determined that permits effective 'lTFields treatment to be delivered to at least two tumors at least some or all of the time. / 39

[0066] ln some embodiments, hysteresis is considered such that points targeted during inhalation and exhalation can be different.

[0067] Since wearing electrodes affixed to the body can be cumbersome to a patient, it is advantageous to affix to the body no more than the electrodes that are necessary (as few and as small electrodes as possible) for delivering effective TTFields therapy. ln some embodiments, when developing the location map, electrode locations that permit delivery to the various positions along the tumor's trajectory are chosen such that electrodes and electrode elements can be (re-)used for targeting multiple positions or with the fewest electrodes or electrode elements, where possible, in order to minimize the overall set of electrodes and electrode elements. ln some embodiments, e.g., according to Fig. 19, when creating an electrode location map, electrode locations are first selected targeting a tumor position along a tumor trajectory 187 where, when targeted, the tumor is expected to be effectively targeted by those electrodes for the longest time during a respiratory cycle, e.g., the tumor would be expected to spend the most time within 20 mm of this targeted position. ln some embodiments, this targeted position would be selected based on the expected trajectory of one of quiet breathing, normal breathing, exercise breathing, deep breathing or some blended measure of the various types of breathing. Having identified this first targeted position and suitable electrode locations, and incorporating the locations in a map, the method can then proceed to consider additional positions along the trajectory and potential electrode locations that can target them, subject to constraints such as patient comfort, constraints due to modular design, electrode element spacing and relative position requirements, and similar. Additional electrode locations can be added to the map if they meet suitability criteria. The method would repeat the steps of considering additional positions and adding electrodes targeting them to the map if the criteria are met, until no more suitable positions to consider are found and the map is complete. The map would then be one or more of presented on a display, printed, e.g., on paper, or generated as part of e.g., a piece of clothing incorporating some or all of the located electrodes.

[0068] ln some embodiments, when creating an electrode location map targeting two or more tumors, a similar procedure can be used to first consider tumor positions for a primary tumor and electrode locations suitable for targeting it, and then considering additional positions and electrode locations suitable for targeting another tumor, and assessing whether they could be beneficially added to the map, and so on.

[0069] ln some embodiments, when creating electrode location maps, for each position electrode locations would be identified and added to a map such each position can be targeted from at least two but even better three or more directions, separated by e.g., ideally 90 degrees apart, in other embodiments perhaps at least 70 degrees apart. ln some embodiments, when targeting multiple positions along a trajectory, and facing constraints as outlined herein, a map can be constructed that for some position would be able to deliver effective treatment from no more than one direction, but where 11/39 a tumor during its motion along its trajectory can be expected to be subject to electric fields applied from different directions.

[0070] ln some embodiments, the position of the tumor is determined, e.g., with MRI, computer tomography (CT), X-ray imaging, ultrasound, or other means, for different respiratory volumes and respiratory states, e.g., while the patient performs respiratory exercises to inhale maximally, exhale maximally, and in some embodiments exploring other respiratory phases, e.g., by holding breath in neutral, and ln some embodiments making determinations while breathing at different speeds or rates, e.g., fast breathing or slow breathing (respiratory rate is metric that can be measured during therapy, and thus serve as an input parameter in a positioning model). ln some embodiments, data (e.g., images) to support tumor position determination is captured using different technologies. ln some embodiments, e.g., a 4-splice scanner such as LightSpeed GX/i from GE Medical Systems can be used for respiration-correlated CT (RCCT) imaging for some of these captures, and in some embodiments e.g., a single-splice scanner such as PQ5000 from Philips Medical Systems. ln some embodiments, positions for ten or more phases, e.g., 16, of the respiratory cycle are acquired. ln some embodiments, a series of images, or other representations of the inside of a patient that contain information identifying the position of a tumor, are captured in a time sequence, e.g., with 4D CT. ln some embodiments, the capturing is performed at a rate permitting positions to be determined while breathing at a fast pace, though this might not be available with all commonly available imaging equipment if the time resolution is too low. ln some embodiments, where the time sequence is captured while the patient is breathing while at rest, the position model built as a function of respiratory rate and other parameters can model the position while a patient is breathing at a faster rate at a sufficient level of precision. ln some embodiments, a gating technique is used, where capture of images or other ways of determining a tumor's position is triggered by a threshold level for a metric, such as respiratory phase, e.g., maximum or minimum levels of measured respiratory phase of the patient. ln some embodiments, the triggering of the capture is manual while the patient assumes different respiratory states. The determination of a tumor's position from the captured images or data can be manual or automated. ln some embodiments, modelling the position or displacement of a diaphragm is used as a proxy that can be used to estimate the trajectory or instantaneous position of a tumor. Using these position determinations, and interpolating, a model can be determined for a tumor's expected position for the extremes of inhalation and exhalation, e.g., the range of the full vital capacity (VC), and at respiratory volumes in between, where the addition of measurements for additional respiratory volumes can improve the accuracy of the identified model for the tumor trajectory. ln embodiments, where different technologies are used to capture data, the different technologies are used one after another to capture data sets that can then be beneficially merged or otherwise used together to build a higher-resolution or higher fidelity model. ln some embodiments, the different techniques can be used overlapping in time, in some cases while 12/ 39 capturing at the same rate, in some cases while capturing at different rates. ln some cases the capturing with different technologies is synchronized in time.

[0071] A model can in some embodiments be |inear between points of measurement/estimation, in some embodiments just based on the two points at the extremes of the trajectory of the tumor. ln some other embodiments, other functional models can be used, such as e.g., cubic interpolation, that can generate a smoother and in some embodiments more precise trajectory model. Given the limited precision requirements of the 'lTFields application, |inear models can be sufficient in many embodiments.

[0072] ln some embodiments, hysteresis (the difference between trajectory of the tumor during inhalation and exhalation) is considered such that different trajectories are estimated for inhalation and for exhalation.

[0073] ln some embodiments, during the collection of initial measurements the position of the tumor is determined while sampling measurements are made at essentially the same time, or in some embodiments e.g., during the same session, and estimates for respiratory metrics are made with the measurements, which can serve to build a higher-fidelity model of a tumor's movement and position with regard to e.g., respiratory phase and volume. This can serve to calibrate the sensor measurements to the model better.

[0074] lt is to be understood that the multiple rounds of measurements can be made, to gather more data and to enable better estimates, as would be obvious to a person skilled in the arts.

[0075] Methods of determining suitable electrode location maps and other methods disclosed herein can in some embodiments be implemented on a personal computer 304 containing one or more processors 303 and memory 305 for program code and data, running e.g., the Windows or MacOS operating systems. ln some embodiments, a program executed on a server 302 (with one or more processors 303, storing the program in memory 305) accessible to a user, such as e.g., a medical professional, through e.g., a web browser 306 user interface, can be used to implement methods related to the determination of electrode location maps and in some embodiments implement other methods as well.

[0076] ln some embodiments, metrics, e.g., related to respiration, are estimated in real-time (meaning that measurements are used as soon as they are available with no more than a minimal delay due to processing etc.), e.g., by a controller, using measurements made with one or more sensors.

[0077] An estimated tumor position 190 and its surrounding CTV 182 can then be estimated, e.g., by a controller, by combining a tumor position model 180 as previously determined, with any of the real- time measurements, any estimated metrics or a combination thereof. ln some embodiments, using multiple input measurements and metrics, especially when derived from multiple sensors 195, can produce better estimated tumor positions 190, or predicted tumor trajectories over time 194. 13/ 39

[0078] ln some embodiments, the tumor position is estimated directly from the measurements, without explicit estimates of respiratory metrics (e.g., cycle phase or respiratory volume). This can apply both to initial measurements and to real-time measurements, and the associated tumor position models 180 and position estimates.

[0079] ln some embodiments, especially where 'lTFields targeting adjustments are not instantaneous, or there is a lag in determining the position of a tumor based on measurements, or intermittent loss of signal from sensors is expected, a predicted tumor trajectory over time 194 is estimated (which can have one, two or three spatial dimensions), which contains one or more estimated tumor positions 190 at the current time or future timepoints. The prediction timeframe in some embodiments depends on the time required to act on new tumor positionings. E.g., if treatment is delivered in bursts 181 of a certain period, and it is desirable not to interrupt or change ongoing bursts, then the prediction timeframe will ideally have to be at least about as long as that period. E.g., for bursts lasting 250 ms, the prediction will have to estimate the tumor position at least about 250 ms in the future, such that a next burst can be scheduled that accommodates the new targeting information.

[0080] Depending on the desired treatment pattern 183 characteristics (a treatment pattern can be one or more identified steps for the application of electric fields targeting certain positions from certain directions for certain durations, in some embodiments separated by silent intervals of certain durations, in some embodiments starting with the immediate next step), other prediction timeframes can be required. For example, in order to generate a treatment pattern 183 for the duration of a respiratory period, a prediction timeframe will have to be about one to six seconds long.

[0081] A prediction model can be based, e.g., on some combination of some of the phase of the respiratory cycle, the respiratory rate, other respiratory metric, or other measure related to the speed of breathing.

[0082] An example of a partially effective solution (as described earlier) is if an embodiment estimates a patient's stage in the respiratory cycle (the phase of the respiratory flow waveform), but does not have good information about a respiratory volume (either by not making measurements, or by not being able to extract meaningful information from measurements), such that the movements of a tumor along its trajectory 187 can be directionally estimated, but not the magnitude. An assumption for the magnitude can then be made, and in some embodiments, this assumed magnitude can, even if within the range of desired accuracy of a static position estimate, allow the embodiment to be at least partially more effective than static therapy delivery.

[0083] ln some embodiments, the tumor position model (180), estimated tumor position (190) and predicted tumor trajectory over time (194) are constrained to only consider a superior-inferior (SI) direction, as tumor movements are typically the largest in this direction and simplifying an otherwise 3D modelling space to one dimension (1D) in some embodiments can e.g., reduce the computational 14/ 39 requirements for implementing embodiments of the present invention. ln some embodiments, a two- dimensional (2D) model is used, e.g., where an SI direction is combined with an AP or LR direction. ln some embodiments, yet other directions, perhaps not aligned to the SI, AP or LR directions, are used to construct a 1D or 2D model.

[0084] lf an embodiment that has an estimate for the respiratory phase, but is unable to extract meaningful magnitude data for the movement along a trajectory 187, also has an additional estimate, e.g., of the respiratory rate, then in some embodiments that additional estimate can be used to provide an estimate of the magnitude of a respiratory volume-related metric (higher respiratory rates sometimes associated with higher e.g., tidal volume).

[0085] Other examples of physiological measurements (that can be included in measurement inputs 178) that can impact the movement of a tumor include blood pressure and pulse, as the cardiac beat can lead to measurable motion in a tumor. Steppenwoolde 2002 found that hysteresis and cardiac beat can have an impact of up to 5 mm and 4 mm, respectively, on lung tumors, which is within the accuracy margin for TTFields, but taking these factors into consideration can improve a model tracking a tumor and reduce uncertainty associated with its outcomes. Since the heart rate is typically higher than the respiratory rate, with a typical resting heart rate of 60 to 100 beats per minute (1 second per period and 0.6 seconds per period, respectively), and during exercise the heart rate can reach e.g., 170 beats per minute (035 seconds per period). lf TTFields are delivered as bursts 181, for the cardiac information to be relevant the bursts 181 must be short enough to permit configuration of a treatment pattern 183 to adapt to the cardiac information, e.g., by having multiple bursts fit within a cardiac period. E.g., with 50 ms bursts, that have been identified as particularly effective, seven bursts will fit in one cardiac cycle at 170 beats per minute, but with the 1 second bursts used in some 'lTFields clinical trials, the cardiac information would perhaps be more difficult to use beneficially in planning the bursts.

[0086] Yet other metrics 179 that in some embodiments can be used to improve the modelling of a tumor's position and trajectory over time include e.g., metrics based on measurement inputs 178 of the human body, such as measurements related to the patient's posture, or measurements and derived calculations that relate to the delivery of TTFields treatment, such as the applied voltage or current, the tissue resistance or impedance. ln some embodiments, tumor position models 180 that consider information from multiple sensors 195, measurements, metrics of different kinds can lead to better estimated tumor positions 190, and predicted tumor trajectories over time 194.

[0087] ln some embodiments, measures of change over time of the above-mentioned measurements and metrics can beneficially be used when building a tumor position model 180.

[0088] ln some embodiments, a tumor position model 180 can be updated or refined based on real-time measurements inputs 178 and metrics 179 gathered during use of the system. / 39

[0089] ln some embodiments, multiple tumors are detected, and their positions determined with initial measurements, and tumor position models 180 are determined that track the plurality of CTVs 182, primarily with respect to respiration but in some embodiments also considering other factors such as e.g., cardiac beat.

[0090] ln some embodiments where multiple tumors are detected, TTFields delivery will switch over time between targeting the position of one tumor or targeting the position of one or more other tumors, such that two or more tumors that have been identified for targeting are targeted some of the time, with suitable duty cycle, e.g., 50%/50%, 60%/40%, 70%/30%, 33%/33%/33%. ln some embodiments, several desired treatment plans are created, one for each tumor, and steps in two or more desired treatment plans that can be coordinated to create one common step in a unified treatment plan 183 are identified and considered in creating the unified treatment plan 183, which is then applied to the patient.

[0091] ln some embodiments, one treatment pattern 183 is determined that balances the targeting of two or more tumors, e.g., by targeting the CTV 182 of a first tumor but with a displacement, e.g., by 5 to 20 mm, in the direction of a second tumor.

[0092] ln some embodiments, a treatment pattern 183 can be determined that effectively targets a first and a second tumor, where the first tumor tends to move longer distances during the respiratory cycle, by aligning the pattern of directions in the delivered bursts 181 of 'lTFields delivered to different positions 185 as the first tumor is targeted, such that the second tumor is subjected to TTFields preferentially from a direction where it is estimated to receive the strongest therapeutic effect, e.g., the highest field strength or power density in the second tumor or second tumor's CTV 182.

[0093] The normal respiratory rate for adults on average is about 15 breaths per minute, with a normal range approximately between 12 and 20 breaths per minute. During exercise, the respiratory rate can reach about 40 - 60 breaths per minute. 12 breaths per minute translates into five seconds per breath. 60 breaths per minute means one second per breath, i.e., a 1 Hz breathing rate.

[0094] As any movement in a tumor is likely to be relatively small during calm and normal breathing, and the benefits of the present invention are particularly relevant for larger tumor movements, it is important that estimation of respiratory metrics works well for higher breathing rates, such as those associated with exercise.

[0095] To sample a full breath within the full normal range (at least 12 breaths per minute), the sampling period should be at least five seconds. Five seconds will capture five respiratory cycles at 60 breaths per minute. A longer sampling period can permit capture of a full breath even at lower respiratory rates, such as six seconds (10 breaths per minute), or eight seconds (7.5 breaths per minute).

[0096] ln order to extract respiratory metrics that will permit the respiratory volume over a respiratory cycle to be determined, such that e.g., the respiratory phase can be estimated, the sampling rate needs 16/ 39 to be significantly higher than the respiratory rate. Simple application of the Nyquist-Shannon sampling theorem, well known to any person skilled in the arts, suggests that e.g., sampling intervals (typically regularly spaced) must be less than 0.5 seconds apart in order to sample a 1 Hz waveform, but this is just a theoretical bound. Extracting therapeutically useful phase information and other metrics requires significantly shorter sampling intervals, as there is significant harmonic content in the waveform of measurements related to the respiratory cycle. For example, the times required for inhalation and exhalation are typically different, with inhalation shorter, such that the waveform does not resemble a pure sine wave. Rather, the lengths of the inhalation and exhalation periods can beneficially be estimated in order to better track and target a tumor. As can be seen in Fig. 7, if sampling at close to 1/2 the respiratory rate, it will in some situations be difficult or impossible to determine the phase of the waveform (it can be shifted 180°), making this sampling rate unsuitable for even rough approximations of the phase (if the phase is shifted 180°, this could focus the applied TTFields where a tumor is not). [0097] During the collection of measurement samples, it is understood that the samples can be stored in a buffer with a length at least as long as the sampling period, such that at any given time, the most recent measurements from a period matching the sampling period are available. ln some embodiments, for some metrics (e.g., for some autocorrelation-related metrics), a longer series of measurements in the buffer than that of a sampling period noted above can be desirable, e.g., twice as long.

[0098] When extracting useful metrics related to respiration, it should also be noted that while respiration is a continuous process, it cannot be assumed that the respiration is strictly cyclical with a constant period, especially in an ambulatory setting where the patient may be moving around, changing postures, changing level of exertion, etc. This means that the waveform cannot be determined with a lower sampling rate by accumulating data over a longer period of time (multiple respiratory cycles). [0099] A further consideration, that can be illustrated with an exemplary embodiment, is that if a TTFields signal 153 is delivered in bursts 181 of a burst length 193 of e.g., 250 ms, and these bursts have to be timed across different electrodes in order to target a tumor or a CTV 182, then the phase information should be with a resolution of at least about 45°, or a 1/8th of a second sampling period if the breathing rate is 1 Hz. More precise phase information would enable more precise targeting.

[0100] ln embodiments that consider the impact of the cardiac beat, the cardiac beat's phase must be determined based on a suitable measurement technology, e.g., ECG, or photoplethysmography, and this will require a relatively high sampling rate. E.g., Kwon et al found 250 Hz sampling rate acceptable for heart rate variability analysis, with 100 Hz sampling rate acceptable when no frequency domain analysis is needed. The required sampling frequency can depend on e.g., the precision required for a 'lTFields application. ln some embodiments, sensors 195 can include sensors that permit determination of the cardiac beat's phase. 17/ 39

[0101] Aliasing of higher harmonic content (causing noise in the sampled signal) and the time delay due to sampling (about half of the sampling period) are also factors that contribute to making sampling rates significantly higher than the respiratory rate a necessity in practice. The rule of thumb for sampled-time control systems is to choose a sampling rate equal to ten or twenty times the desired system bandwidth (Wescott 2018).

[0102] Examples of practical sampling rates known from the literature for detecting some respiratory metrics include 10 Hz (Lunschmann et al 2022), 20 Hz (Retory et al 2016) , 30 Hz (e.g., Seppenwoolde et al 2002), 200 Hz (Retory et al 2016) and 10 kHz ("to provide a temporal resolution of 0.1 ms", Mansy et al 2018) and are also known to have been used for the determination of respiratory rate and volume, or the trajectory of tumor motion.

[0103] U.S. Patent Publication No. 2022/0305277 A1 describes sampling at regular intervals, ranging between every one second and every four hours (in all, ten different rates mentioned, for a span of more than four orders of magnitude) which is clearly too slow to implement the present invention. The mentioned metrics followed in that application are consistent with the slow sampling rates as e.g., some measures of ”respiratory volume” change over time at slower rates than that of the respiratory cycle, and can be estimated using the disclosed sampling intervals. There is also no suggestion of synchronization with the respiratory cycle. On the contrary, this prior art suggests that metrics or rates of changes in metrics can be averaged over a time window of 15 days.

[0104] Several techniques are known for continuously gathering respiration-related measurements in an ambulatory setting, typically using electrical or mechanical properties (shape) of some combination of the torso, abdomen and tissues within them. Examples that can be suitable for use in sensors (195) include Respiratory inductive plethysmography (RIP), Chest galvanic skin response (GSR) and Chest wall movement (WM). ln some embodiments, multiple sensors are used, in some embodiments employing different technologies, to gather respiration-related information.

[0105] RIP typically uses two bands, one around the rig cage and one around the abdomen, containing wire coils, where their inductance changes as the bands stretch and contract. RIP with one band, across the rib cage, is also known.

[0106] Simultaneous chest GSR signal can be acquired e.g., using two separate electrodes attached next to subject's right clavicle and left abdomen. ln some embodiments, WM can be measured by a piezo electric sensor attached to an elastic strap placed over the chest (Mansy 2018).

[0107] ln some embodiments, candidate locations for sensors 195 to be used for measurements are identified on a subject's body by affixing fiduciaries during or prior to initial measurements, and selection method, e.g., a genetic algorithm-based optimization method, is used to select one or more sensor locations for any sensor 195 from the candidate locations by using data gathered e.g., with a CT scanner. ln some embodiments, sensors 195 placed in the selection from candidate locations identified 18/ 39 by the fiduciaries can be accelerometers. ln some embodiments, the locations for sensors 195 are placed as customary for the type of sensor employed.

[0108] ln some embodiments, one or more sensors can be included in one or more electrodes 130, such that sensors can be applied without the practical drawbacks of having to affix yet more items to the body in addition to any treatment electrodes 130.

[0109] ln some embodiments, a sensor uses two pieces of flexible printed circuit board (PCB) material, known as a sliding strip and a reference strip, where the sliding strip can move with respect to the reference strip, and the sliding strip moves in relation to the reference strip when bending is applied. lf a probing electrode is present on one strip and a sensing electrode is present on the other, then by measuring the capacitance between the two electrodes, the degree of deformation can be estimated (e.g., by a controller). The basic principle can be extended by having a plurality of electrodes on either strip (e.g., as described in U.S. Patent Publication No. 2021/0137418 A1) which enables the shape of the sensor to be determined. ln some embodiments, one of the two strips, e.g., the reference strip, is stretchable, such that the sensor arrangement can measure the stretching of the sensor.

[0110] ln some embodiments of electrodes 130 for '|'|'Fields, flexible PCB materials are used. ln such cases, one of the two strips mentioned in a sensor above can be integrated into the flexible PCB material that is used for the 'lTFields array, saving on parts and complexity.

[0111] ln some embodiments, sensors are used to make measurements across or around the ribcage, or from a ribcage extending down to the abdomen to measure the relative motion between the rib cage and the abdomen.

[0112] Various techniques are known for calculating metrics from the measurements collected by sensors 195. E.g., autocorrelating a buffer of samples with itself shifted with some delay, and repeating this for various delays, and picking the highest correlation (peak detection) is a known technique for detecting the period of a (perhaps noisy) signal. Another approach is to use a Fourier transformation to consider the buffer data in the frequency domain, and identify the frequency with the highest signal content, and thus the period of the signal. Various filters can be applied in some embodiments to the buffer of samples, e.g., a sliding averaging window that would smooth high-frequency noise in a signal (assuming that the sampling rate is fast enough that this low-pass filter would not eliminate the respiration-related information content). Detecting peaks and troughs in a time-domain signal can be used to detect the crossing between inhalation and exhalation phases, and so on. By e.g., detecting the period of the signal and a reference point such as the time of a peak or a trough, or a maximum or minimum, in the respiratory cycle, the respiratory phase can be estimated. By e.g., detecting the magnitude of a peak and a trough, or a maximum and minimum, in the respiratory cycle, the magnitude of the tidal volume, or other respiratory volume, can be estimated. A person having ordinary skill in the art would understand that these are but examples of how metrics can be calculated, that the underlying 19/ 39 mechanics of the respiratory process and harbor additional nuances, and that other processes in some embodiments can be expected to be beneficial, e.g., by having a controller with a trained machine learning algorithm that considers input from multiple sensors, perhaps of different types, to estimate the respiration-related metrics.

[0113] Several possible sensing techniques that can be used to measure respiration-related measurements make use of electrical phenomena (e.g., capacitance, inductance). Given the significant electrical fields that can generated by TTFields equipment, delivery of TTFields could disturb the proper operation of a sensor.

[0114] These problems can be mitigated in several ways, e.g., a) by using sensors physically separated by some distance, e.g., 10 cm, from the active treatment electrodes 130, e.g., by having separate sensors not integrated into treatment electrodes 130, b) for a sensor integrated in an electrode 130 (in some embodiments in turn comprising one or more electrode elements 184), making actual measurements when that electrode is not in use, as other electrodes 130 are perhaps used to generate a field in a different direction at that time, c) making measurements when no 'lTFields are delivered, e.g., between bursts 181, d) using shielding to prevent 'lTFields from interfering with the workings of a sensor, e) eliminating any artefacts of TTFields signal interfering with the measurement by signal processing or other means.

[0115] As an example of c) above, if 50 ms bursts 181 are delivered in the treatment regime, and measurements are made between them, then up to 20 Hz sampling rate is possible, subject to the length of burst spacings 186 (or burst separations) between them. (lf the respiratory period is one second, there will be up to 20 bursts 181 in each respiratory period).

[0116] ln some embodiments, in order to facilitate example e) above, the TTFields are delivered as a pure sine wave (no other, superimposed wave components), which can simplify elimination from a measured signal of any 'lTFields artifacts, especially when the applied 'lTFields signal is stable (not ramping). ln some embodiments, example e) can be implemented with 'lTFields comprising two or more frequency components.

[0117] ln some embodiments, if the sensing of respiratory measures fails for any reason, for a longer or shorter time, e.g., due to the patient changing its posture, the system can fall back on a default or simplified treatment pattern 183 (a treatment pattern 183 can in some embodiments be indefinite in length, e.g., by repeating the same steps in a loop).

[0118] ln some embodiments, that default pattern is delivery of TTFields as if a tumor is in a static position in the torso.

[0119] ln some embodiments, the tumor position estimates or the input metrics used to perform the estimation are filtered in order to e.g., get the best possible estimates for the real-time needs of some / 39 embodiments the present invention. This can e.g., be achieved by using a Kalman filter or a non-linear filter such as an extended Kalman filter, in order to improve the position estimate. lnputs into such filters can include some or all of available measured and calculated values, e.g., any initial measurements, tumor position model 180, measurement inputs 178, metrics (e.g., respiratory rate, respiratory phase), estimated tumor positions 190 (e.g., estimated positions at previous timepoints) and current burst 180 characteristics. The estimated position can be for a point in time reflected by the measurements, but since there is a time delay associated with the measurements, the estimated position can be for a point in time subsequent to the measurements. The estimated position can in some embodiments be for a point in time for the mid-point of the next expected burst, and in some embodiments for the next expected measurements, and in some embodiments for yet other timepoints. ln embodiments using e.g., a Kalman filter, it can be beneficial to be able to calculate a measure of the uncertainty associated with the respiration-related measurements, or any other measurement inputs 178 or metrics, that are inputs to the Kalman filter, such as their variance. Since the measurements are sampled at relatively high speed, and adjustments to the applied TTFields are made in some embodiments with sub-second speed, there can be considerable uncertainties and errors in the measurements, as there is little time for filters for smoothing or averaging inputs, and there can be short-term disturbances e.g., when a patient changes posture, that can upset sensor 195 measurements. ln some embodiments, with the loss of respiration-related measurements, a Kalman filter or similar can be configured to produce a treatment pattern 183 based on the latest available information regarding the respiratory metrics and cycle, and will continue to retarget TTFields treatment to follow the estimated position and trajectory based on that information. A filter can in some embodiments be configured to account for the hysteresis in the trajectories.

[0120] ln some embodiments, if a loss of sufficiently high-quality measures extends for some duration of time, a simplified treatment pattern 183 can be deployed that assumed a static tumor position.

[0121] ln some embodiments, facing a loss of qualitative measurements, using filtering to estimate tumor positions 190 or tumor trajectories over time 194, and in turn using that to build treatment patterns 183 that are used for treatment delivery as they become available, is in some embodiments advantageous over continuing delivery according to a longer-running treatment pattern 183 based on the last available quality measurements, as e.g., a system using a Kalman-type filter which can consider measures of the uncertainty associated with input measurements or metrics in some embodiments can be expected to produce better conformity between a delivered pattern of bursts and e.g., positions of any targeted tumors.

[0122] ln some embodiments, principal component analysis (PCA) is applied to reduce a dataset of multiple sets of data, e.g., from different sensors 195 or calculated metrics, in order to determine the best contributors for the determination of a tumor's position. 21/ 39

[0123] By having multiple sets of electrodes 130, or activating different groups of electrode elements 184, 'lTFields can be directed to deliver the highest-intensity treatment to the area where a tumor is expected to be located through the respiratory cycle. ln general, the higher the electrical field strength in a targeted tumor or CTV 182, the better from a therapeutic perspective, and the lower the voltage that has to be applied across electrodes 130 the achieve a certain field strength, the better, as this will serve to reduce power draw, reducing drain on power sources such as batteries, potentially reducing the waste heat generated, while also potentially reducing any heating of a subject's skin or tissues and any associated inconveniences. ln some embodiments, TTFields are administered with 50 V placed across electrodes, to generate the desired field strength.

[0124] ln some embodiments, a first pair of electrodes 130 are used to deliver fields that alternate back- and-forth between the electrodes of the pair in a first direction, with 'lTFields field strength expected to be at least 1 V/cm in a CTV 182, and preferably stronger, in some embodiments up to 5 V/cm (e.g., a range of 1-5 V/cm), and as a tumor moves to a new position, treatment is moved to new electric field with a second pair of electrodes 130 (at least one of which is not part of the first pair) such that the TTFields field strength is expected to be at least 1 V/cm in the tumor in its new position, and preferably up to about 5 V/cm.

[0125] 'lTFields work best for an individual cancer cell if the field is properly aligned with respect to the orientation of the internal molecular structures inside the cell during mitosis. This means that switching the applied fields between different directions (using different pairs) can improve outcomes even without considering any movement in a tumor, e.g., in the torso. Typical known switching rates between different directions are 4 Hz or 1 Hz, meaning that TTFields is typically delivered in bursts lengths 193 of about 250 ms or 1 s. Burst lengths 193 as short as 10 ms or 20 ms are known.

[0126] ln some embodiments, TTFields are delivered in a treatment pattern 183 with consideration both to delivering treatment to a tumor from different directions, and to delivering treatment to a tumor at different positions as it moves within the body. This can in some embodiments mean having multiple sets of pairs of electrodes 130 (to enable delivery from two or more different directions), with additional such multiple sets of pairs to enable the delivery of fields from different directions as the tumor moves to new positions (e.g., Fig. 10). ln some embodiments, two such multiple sets of pairs are used, in some embodiments three such multiple sets of pairs, are used (e.g., 2 x 2 x 3 = 12 electrodes in total. Sometimes different pairs share an electrode, such that the total number of electrodes that needs to be deployed can be reduced).

[0127] ln some embodiments, changes in field directions or positions are not implemented by activating or deactivating electrodes 130 for 'lTFields in their entirety, but by selectively switching one or more electrode elements 184 of one or more electrodes 130 on or off (e.g., Fig. 2). ln some embodiments, a field is delivered through a selection of electrode elements 184 from more than one electrode 130 22/ 39 together, together forming a 'virtual electrode'. The other side of an electrode pair that contained such a virtual electrode can consist of an entire electrode 130, a selection of electrode elements 184 of an electrode 130 or another virtual electrode. ln some embodiments, these types of arrangements can allow smooth and 'seamless' tracking of a tumor position with 'lTFields delivery, especially if the burst length 193 is short. Unless otherwise noted, electrodes 130 as disclosed herein can refer to entire electrodes, or a selection of electrode elements 184 from one or more (i.e., virtual) electrodes.

[0128] ln some embodiments, the targeting of a tumor or a CTV 182 as it moves is performed by synchronizing the switching rate (burst lengths 193 and burst spacings 186) of delivery from different directions. lf e.g., a tumor during part of its motion in the respiratory cycle will be more exposed to an electric field applied between a first pair of electrodes 130, for one direction, than another pair for another direction, then an improved, stronger field strength can be achieved by synchronizing the directional switching such that the first pair is active (or more likely to be active, e.g., with a disproportional duty cycle) during that part of the tumor's motion, in some embodiments at the expense of directional coverage. This can happen e.g., if a tumor will tend to move along the axis between one pair of electrodes (within a field), and perpendicular to such an axis for another pair (which can make the tumor move out of a field).

[0129] Targeting a tumor by activating different sets of electrodes 130 is best performed with consideration to the burst lengths 193, as a burst that is interrupted and is partially delivered across different electrodes in effect becomes two bursts (or more), perhaps or even likely of non-ideal length. Adjusting the targeting is therefore typically best performed between bursts. ln some embodiments, it is desirable to ramp the burst up when it starts and ramp it down when it ends (with a limited slew rate, and either in a linear or a in non-linear fashion), such that EMC artifacts are minimized, as well as any negative experiences or discomfort for a patient. Switching to new electrodes mid-burst will in such embodiments preferably be associated with ramping during the switch as well.

[0130] Adjusting a treatment pattern 183 to align with the best available estimate for a tumor's position, and in some embodiments its expected future trajectory over time 194, is in some embodiments performed by adjusting burst length 193 or the burst spacing 186 between bursts or a combination of the two.

[0131] ln some embodiments, a treatment pattern 183 is synchronized with a respiratory period. [0132] ln some embodiments, the synchronization is performed by fitting an integer number of bursts 181 and burst spacings 186 within the respiratory period, but other schemes are possible, such as e.g., fitting a fractional number of bursts per respiratory cycle, e.g., an integer number of bursts for every two respiratory cycles. ln some embodiments, burst lengths 193 and/or burst spacings 186 can be adjusted based on respiratory rate, such that e.g., during fast breathing (short respiratory period), more but shorter bursts 181 are fit into a respiratory cycle, such that these bursts 181 can better follow a 23/ 39 tumor in the body. ln some embodiments, the estimated magnitude of a tumor's movement is used to adapt the burst length 193 or burst spacings 186, e.g., if the respiratory volume is low during fast breathing, limited benefit can be associated with shortening the burst lengths 193. (Since burst lengths 193 have been found in in vitro experiments on cell cultures to have suitable lengths depending on e.g., cancer cell type, not adapting the burst length 193 can be beneficial in such cases).

[0133] ln some embodiments, a treatment pattern 183 is synchronized with a respiratory phase, such that starts and stops of bursts 181 are aligned with the most advantageous part of the respiratory cycle, considering e.g., that organ movements inside the torso or abdominal areas of a body tend to be relatively small during e.g., the period of time around the point of largest exhalation, and that there can be benefits in terms of therapeutic efficacy from delivering one or more bursts 181 timed to coincide with such a part of the respiratory cycle. ln some embodiments, burst lengths 193 of one or more bursts 181 can be adjusted to match such a period of time in the respiratory cycle, to take advantage of the limited motion. ln some embodiments, where the entire tumor trajectory cannot be effectively targeted, burst spacings 186 can be timed to occur when a tumor is outside of the range that can be effectively targeted, such that potential treatment time is not lost unnecessarily.

[0134] ln some embodiments, the adjustments of burst length 193 and burst spacing 186 are determined considering some combination of pre-specified minimum and maximum burst lengths 193, and minimum and maximum burst spacings 186, to best fit the current position and projected tumor movement. ln some embodiments, there is essentially no minimum burst spacing 186 length (excepting e.g., minimum spacing required for safe switching with e.g., MOSFETs to avoid the risk of shoot-trough, which could create a potentially hazardous situation).

[0135] When a new treatment pattern 183 is available, the active treatment pattern can be updated to reflect this pattern before the previously active pattern has completed. Typically, the burst 181 currently being delivered is allowed to complete before a change to a new treatment pattern 183 is made, see e.g., Fig. 12.

[0136] As each burst is started, in some embodiments it locks the delivery for the pre-determined duration of the burst 181, e.g., 250 ms or 1 sec (this is a treatment pattern 183 of at least one step). ln some other embodiments, a burst 181 can be cut short while it is delivered, if the new burst length is at least as long as a minimally specified burst.

[0137] ln building a new treatment pattern 183, the characteristics of any currently running treatment pattern 183, or any treatment patterns 183 that were active before the current treatment pattern 183, can be considered, e.g., in order to build a suitable treatment pattern for positions along a tumor's trajectory over time.

[0138] lf the electric fields change between targeting several different positions with bursts 181 of a certain length 193, and directions change in an alternating pattern, then in some cases for some 24/ 39 positions fields will tend to be delivered disproportionally from a particular direction (Fig. 4). ln some embodiments, a controller that builds the treatment pattern 183 will also make sure that for each targeted position, the direction of field application will switch each time a new burst is applied to that position (Fig. 5). ln some embodiments, this is accomplished by storing an indicator in a memory, one for each position, where the indicator contains information suitable for determining the next application direction.

[0139] Referring again to Fig. 2, disclosing four electrodes 130 (130a, 130b, 130c, 130d) attached to the torso of a human body, where 130a and 130b form a pair for delivering electric fields in the left-right direction, and 130c and 130d similarly deliver a field in the anterior-posterior direction, each containing a plurality of electrode elements 184 (184a-c on 130c, 184d-f on 130a, 184g-i on 130b. Not shown: 184j- I on 130d), and where each electrode element 184 is connected to the other electrode elements 184 on the same horizontal row (not shown). This arrangement is suitable for changing the targeted volume by delivering bursts 181 through a selection of electrode elements 184 for each of a pair of electrodes 130, e.g., first the bursts between the row with 184d and the row with 184g, then for a different position, bursts between the row with 184e and the row with 184h, and so on. For the anterior-posterior direction, in similar fashion, the row with 184a and the row with 184j, then the row with 184b and 184k, and so on.

[0140] Referring again to Fig. 4, of a treatment pattern 183 for delivery of 'lTFields through three different field positions (185a, 185b, 185c) as a tumor is tracked and targeted, where bursts are delivered in alternating directions (through the use of electrodes 130 that have similar orientations though different positions are targeted), and where for the field position 185b, the direction of the delivered field is predominantly the same. The letters 'A' and 'B' indicate two different directions of field delivery, typically roughly orthogonal (or ideally so). The illustration shows the positive side of the voltage envelope 175 of the bursts 181, that will typically be oscillating at 100 kHz to 300 kHz within the indicated voltage envelope 175, commonly at 150 kHz, centered around zero volt (achieved e.g., by using of insulated electrodes). ln some embodiments, ramping will be employed (not shown).

[0141] Referring again to Fig. 5, the illustration shows the positive side of the voltage envelope 175 of the bursts 181, that will typically be oscillating at 100 kHz to 300 kHz within the indicated voltage envelope 175, centered around zero volt. ln some embodiments, ramping will be employed (not shown).

A controller 196 will in some embodiments, e.g., by using several counters in memory, make sure that every time an electric field is targeted at a position, it will be from a different direction from that of the last time the same position was targeted.

[0142] Referring again to Fig. 6, there are three field positions 185a, 185b and 185c, that describe different positions for a lung tumor (and its CTV 182) as it is moved by the changing shape of the lung 188 and the position of the diaphragm 189 (the movements of the lung and diaphragm are not shown), / 39 that are targeted by different configurations of electrodes 130 and in some embodiments included electrode elements 184. The letters 'A' and 'B' indicate two different directions of field delivery, typically roughly orthogonal that can be similar in direction for different field positions 185.

[0143] Referring again to Fig. 7, it is clear that a low sampling rate such as 1 Hz is insufficient to determine respiratory phase and other metrics. Considering e.g., measurements that would be made at 1 second sampling intervals 191 between seconds 2, 3 and 4 in the figure, tidal volume measurements are essentially unchanged between those second-long intervals with no indication of where in the respiratory cycle the patient is (e.g., going up or down).

[0144] Referring again to Fig. 8, in this figure, the burst length 193 is not synchronized with the respiratory cycle period, and instead the number of bursts 181 at each field position 185 is variable over time. The illustration shows the envelope around the output voltage, which will oscillate at typically 100 kHz to 300 kHz.

[0145] Referring again to Fig. 9, in the illustration, the respiratory rate is 60 breaths per minute, for a 1 second respiratory cycle. ln the illustration, ramping up and down of each burst (with some limited slew rate) is shown. Delivering the bursts in different directions and/or to different positions is not shown in the illustration for brevity, but is understood to potentially be taking place. Note that in many embodiments the oscillating period of 'lTFields (shown only for the first burst) is orders of magnitude shorter than the burst length; in the illustration the frequency of 'lTFields is shown lower for clarity. The bursts are shown illustratively with the positive side of their voltage envelopes 175.

[0146] Referring again to Fig. 10, 130a and 130b form a pair for delivering electric fields in a superior position (cranial direction) in the left-right direction (B), and 130c and 130d similarly deliver a field in the anterior-posterior direction (A). The pairs of 130e and 130f, and 130g and 130h, respectively, are used to generate similar directions in inferior positions (caudal direction) on the torso. An illustration of a tumor moving along a trajectory 187 inside the torso during the respiratory cycle is shown. But switching between using the superior and the inferior sets of electrodes, a tumor can be targeted as it moves.

[0147] Referring again to Fig. 15, where a system 100 contains a TTFields generator 198 and a controller 196, connected to at least one pair of electrodes 130, where the TTFields generator 198 is controlled by a controller 196 which in turn is connected to at least one sensor 195, and where the controller has a connected memory 197 for storing e.g., a treatment pattern 183, or a log or buffer of measurement samples, or a combination thereof. The exemplary embodiment in Fig. 15 can be used to implement one or more of the methods disclosed herein. ln some embodiments, the memory 197 is part of the controller 196.

[0148] Referring again to Fig. 16, where the generator 198 contains a voltage regulator 161 (that in some embodiments can provide an input voltage envelope 175 for bursts 181 of TTFields), a power 26/ 39 circuit 160 which switches the input voltage such that a bipolar pulse train with peaks at the input voltage is created, an insulated transformer 167 which can perform one or more of insulating the output, changing the output amplitude from the input amplitude, and filtering the pulse train such that a sinusoidal output, or output that contains controlled frequency content, is created (or at least filtering the signal such that it can be advantageously filtered in a downstream output filter 168), and an optional output filter 168 which can be present to filter out any remaining unwanted waveform components or characteristics. ln some embodiments, the isolating transformer 167 is left out. ln embodiments where 'lTFields are applied from different directions (through different electrodes 130 or electrode elements 184), direction switching circuitry 199 are included that connects any output TTFields signals with selected electrodes 130 or electrode elements 184 (ln some embodiments, the direction switching circuitry 199 can be eliminated). ln a typical exemplary embodiment, the voltage regulator 161, the power circuit 160 and the direction switching circuitry 199 are controlled by a controller 196. ln some alternative embodiments, the power circuit 160 can switch to create a pulse suitable for creating a 'lTFields signal of a certain frequency without having its switching action under control of the same controller 196 that controls the burst envelopes and the directional switching. (E.g., by having a separate controller). Likewise, the directional switching 199 can be under the control of a separate controller, e.g., by having the separate controller detect the start and stop of any bursts and performing direction switching between them, in some embodiments saving on e.g., cabling and communications needs.

[0149] Referring to Fig. 1, a first method 200 for targeting 'lTFields based on real-time measurements is disclosed. According to this process, first, initial measurements 210 are made, by performing a tumor position determination 201 where multiple determinations for a tumor's position in the body are made, e.g., using MRI, CT, X-ray imaging, ultrasound, or other technique, for different respiratory states (fully inhaled, fully exhaled, etc.). Using these determinations, as a next step 202, a Tumor Position Model 180 is made that can be used to estimate the position of a tumor together with measurements made later, in real-time. Sensor measurements are made 204 while a patient is receiving 'lTFields treatment, and these measurements are used together with the tumor position model 180 previously built to estimate a current CTV 182 position. Using the estimated tumor position, 'lTFields are delivered 205 to the targeted position.

[0150] ln a second method 220, shown in Fig. 3, first initial measurements with respiratory sensing 211 are made, where tumor position determination 201 is performed while also performing sensor measurement 206 and estimation of respiratory metrics 208. ln some embodiments, the sensor measurement 206 is performed at the same time as the tumor position determination 201, e.g., while the patient is in an MRI, CT or X-ray machine making measurements. ln some embodiments, tumor position determination 201 and sensor measurement 206 are performed as separate steps, in some 27/ 39 embodiments on the same occasion, where the same respiratory states are measured for both steps (fully exhaled, fully inhaled, etc.). The tumor position data and the respiratory data thus accumulated are used to build a tumor position model 180 as a step 202. Next, during treatment of a patient, sensor measurements 204 are made, used to estimate respiratory metrics 208, and these estimated respiratory metrics are then used together with the previously built tumor position model 180 to estimate a tumor trajectory 212 (in some embodiments, this can be equivalent to estimating a CTV position 203). Using the estimated CTV position/trajectory, a treatment pattern 183 is constructed 209. Then, TTFields are delivered 205 based on the updated, new treatment pattern 183. Next, the process makes new sensor measurements 204 and the cycle continues.

[0151] Referring again to Fig. 3, in a first step measurements and estimated respiratory metrics are used when building a tumor position model 180; and then during treatment real-time measurements are used to estimate respiratory metrics 208, which are then used to estimate a clinical tumor volume (CTV 182)'s trajectory 209, and using the tumor position model 180 and the estimated trajectory 194, an updated treatment pattern 183 is built and the 'lTFields delivery 205 updated to reflect this new pattern. The sensor measurement 204 step in some embodiments can be triggered by any timers, availability of new sensory data, or other.

[0152] Referring again to Fig. 11, the measurement inputs include a measure of respiratory volume sampled at high frequency (higher than 1 Hz), and can also include other measures such as e.g., blood pressure which, also sampled at high frequency, can be used to calculate e.g., cardiac beats. The measurement inputs in some embodiments include estimates and their variance. ln some embodiments, the measurements have been filtered previously, to e.g., eliminate noise or confounding factors. ln a step 208, a number of measurement inputs related to respiration 178 (e.g., containing information related to respiratory parameters such as volume, etc.), as well as yet other metrics in some embodiments, are fed into a filter, where the filter in some embodiments is a type of Kalman filter, and where the filter, based on the input and any of its internal feedback loops (not shown), outputs one or more metrics 179, e.g., related to respiration, including phase, and in some embodiments respiratory volume, and in some embodiments respiratory rate or period (the respiratory rate and period being reciprocals of each other). lt is understood that the key desired output from steps 204 and 208 is to identify the respiratory phase, and that measures other than respiratory volume can be used to determine the phase. ln a step 212, a predicted tumor trajectory over time 194 is estimated based on a tumor position model 180 and the respiratory metrics. ln some embodiments the estimates consider hysteresis to create different trajectory paths during inhalation and expiration, in some embodiments hysteresis is not considered and an average or similar weighted measure balancing inspiration and expiration is used. The predicted tumor trajectory over time is then in a step 209 used as input when building a treatment pattern 183, consisting of planned bursts 181 though two different sets of 28/ 39 electrodes 130, 185a and 185b, respectively. ln the shown embodiment, the burst lengths 193 are configured such that an integer number of bursts fit in the expected respiratory period. ln some embodiments, the predicted trajectory over time 194 consists of a single estimated position, and the treatment pattern 183 consists of a single next burst 181. ln the illustrative embodiment in Fig. 11, the respiratory period is about 1.3 seconds. The desired burst length 193 is 250 ms and the minimum burst spacing 186 is 20 ms. The number of bursts 181 in one respiratory cycle is calculated as 1.3 seconds/ (250 + 20 ms) = 4.81, which is rounded down to four. ln some embodiments under these conditions, the burst length 193 in the treatment pattern 183 is 250 ms and the burst spacing about 62 ms. The treatment pattern 183 is created by a controller 196 (see Fig. 15). ln the illustration in Fig. 11, only the positive side of the bursts in the treatment pattern 183 is shown. The illustration shows a voltage envelope for the actual waveform, which will oscillate typically at 100 kHz to 300 kHz.

[0153] Referring again to Fig. 12, where in a first step 214, delivery of a burst 181 according to the treatment pattern 183 is started, in a following step 215 the completion of the burst 181 is waited for, in a next step 216 any electrodes 130 or electrode elements 184 that are not needed for delivery of 'lTFields according the next step in the treatment pattern 183 (or to treat an estimated CTV position) are deactivated, and then in step 217 any electrodes 130 or electrode elements 184 needed for the next TTFields electric field application are activated. The method 205 then loops back to start over at step 214. The selection of electrodes 130 or electrode elements 184 for the application of 'lTFields can be performed according to embodiments disclosed herein. The 'lTFields electric fields applied through the use of the method 205 e.g., can relate to the targeted position, the direction of the applied fields, or a combination thereof, as disclosed herein.

[0154] Referring again to Fig. 13, where a method 220 for delivering TTFields to a patient, where initial measurements 210 are first made; then electrode location mapping 219; then TTFields treatment is started 213, then one or more tumors are tracked 218; then TTFields delivery 205 is performed. ln some embodiments, the step 210 is replaced with a step 211, where initial measurements with respiratory sensing are used. The initial measurements made are then used for at least one of tracking 218 and TTFields delivery 205.

[0155] Referring again to Fig. 14, where a method 221 for delivering 'lTFields to a patient, where the electrode location mapping 219 is first performed, and can be based on default electrode mapping or based on previous analysis of the condition of the subject. ln a next step, initial measurements 210 are first made. ln some embodiments, the step 210 is replaced with a step 211, where initial measurements with respiratory sensing are used. The initial measurements made are then used for at least one of tracking 218 and TTFields delivery 205.

[0156] Referring again to Fig. 17, where a method 219 for determining electrode location mapping contains the steps of generating a 3D model of a subject's body 222, generating a tumor trajectory 29/ 39 model 223, identifying a plurality of positions 224 along the trajectory, and determining a location map 225 indicating the locations and configurations of electrodes to be placed on a subject's body for the delivery of 'lTFields according to embodiments of the present invention.

[0157] Referring again to Fig. 19, where a method 225 of determining an electrode location map contains the steps of identifying first electrode locations (for targeting a first tumor position) 226, adding the electrode locations to a map 227; if no more potential electrode locations remain, then outputting an electrode location map 228, e.g., by printing it, displaying it or generating it; if more potential electrode locations remain, they are assessed by specified criteria, and if they meet the criteria, the they are added to the electrode location map 227 and the method continues from there, if they do not meet the criteria, it is again considered if more potential electrode locations remain. /39 "Modes" of the invention

[0158] The following is a numbered list of non-limiting illustrative embodiments of the invention in several different modes: 1. . 11.

A method for applying tumor treating fields to a tumor of a subject's body, the method comprising: applying an electric field between a first pair of electrodes targeting a first location of a subject's body, sampling measurements from one or more sensors (195), estimating one or more metrics including a metric related to the subject's respiratory phase based on the sampled measurements, estimating a second location of the tumor based on the metrics, and applying an electric field between a second pair of electrodes targeting the second location of a subject's body.

The method of mode 1, wherein the measurements are sampled at a rate higher than one hertz. The method of mode 1, wherein the measurements are sampled at a rate higher than ten hertz. The method of any of the previous modes, wherein the metrics comprise a measure of respiratory volume.

The method of mode 1, wherein during application of an electric field targeting a location, the field is applied in a sequence between two pairs of electrodes targeting the location.

The method of any of the previous modes, wherein one or more of the measurements or metrics include a measure of measurement uncertainty.

The method of mode 1, wherein the estimation of a second location of a tumor is configured to consider measurement uncertainty or missing data during the estimation.

The method of mode 1, wherein the metrics comprise respiratory period, and the electric fields are delivered in bursts of a certain lengths separated by separation of certain lengths, and the lengths of bursts and lengths of separations are configured to fit an integer number of bursts and separations into a respiratory period.

The method of mode 8, wherein the burst lengths are of a pre-specified length, and the separation lengths are configured to fit the bursts and separations into the respiratory period. The method of mode 8, wherein the metrics comprise the inhalation period and the exhalation period, and the bursts are configured to be synchronized with the inhalation and exhalation periods by varying one or more of the burst lengths and the burst separations.

The method of any of the preceding modes, wherein the estimated second location of a tumor includes a surrounding clinical tumor volume. 12. 13. 14. . 16. 17. 18. 19. . 21. 22. 23. 24. . 26. 27. 28. 31/ 39 The method of mode 1, wherein the estimation of a second tumor location considers initial measurements that are made to determine the location of a tumor in a subject's body.

The method of mode 12, wherein the initial measurements are made using at least one of magnetic resonance imaging, computer tomography or ultrasound.

The method of mode 12, wherein the initial measurements include sampling measurements that comprise a signal for respiratory phase.

The method of mode 14, wherein some or all of the initial measurements are made while applying electric fields.

The method of any of modes 1 to 15, wherein an electric field comprises an alternating electric field oscillating at between 50 kHz and 500 kHz.

The method of any of modes 1 to 15, wherein an electric field comprises an alternating electric field oscillating at between 100 kHz and 300 kHz.

The method of any of modes 16 or 17, wherein the electric field comprises a frequency of 150 kHz.

The method of any of modes 16 to 18, wherein the electric field contains essentially a single frequency component.

The method of any of modes 1 or 16 to 19, wherein a field strength is at least 1 V/cm.

The method of any of modes 16 to 20, wherein a field strength is between 1 to 5 V/cm.

The method of any of modes 1 to 21, wherein electrodes used to target a tumor are insulated. The method of mode 1 or 5, wherein one or more of the electrodes comprises several electrode elements such that when an electric field is applied through an electrode, some or all of the electrode elements are used.

The method of mode 23, wherein electrode elements from two electrodes are used together to form a virtual electrode that can be used to apply an electric field.

The method of either mode 1 or mode 5, wherein at least two pairs share a common electrode. The method of mode 1, wherein the estimation a second location includes a plurality of future locations forming a trajectory, and wherein a treatment pattern is configured to applying electric fields over time based on the estimated trajectory.

The methods of mode 5, wherein when applying electric fields to a first location using a first pair and a second pair of electrodes, during a course of treatment wherein electrical fields are applied to several locations, the fields when applied to the first location are applied in an strictly alternating sequence between the first and second pairs.

An electrode (130) for applying an electric field to a subject's body, wherein the electrode comprises a sensor (195) wherein the sensor can make measurements that can be used to estimate a respiratory phase. 29. . 31. 32. 33. 34. . 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 32/ 39 An electrode (130) for applying an electric field to a subject's body, comprising a flexible printed circuit board material for connecting a plurality of electrode elements (184), wherein the same printed circuit board material is used as strip as part of a sensor (195).

The method of mode 11, wherein the respiratory phase can be estimated for respiratory rates between 10 and 60 breaths per minute.

The method of mode 1, wherein the measurements are sampled for a period of at least 6 seconds.

The method of mode 1, wherein measurements include a measure of cardiac activity.

The method of mode 11, wherein the surrounding clinical tumor volume extends 20 mm around a tumor volume.

The method of mode 1, wherein the tumor is located in one of the lungs, kidneys, liver, pancreas, or mesothelium.

The method of mode 1, wherein the precision of the second estimated location is +/- 20 mm. The method of mode 1, wherein the second location is estimated only along a superior-inferior axis.

The method of mode 1, wherein the method is repeated in a loop.

A method for creating a map of electrode placement locations on a subject's body, for a plurality of electrodes (130), wherein the map is planned to enable application of electric fields at plurality of positions along a tumor's trajectory during a respiratory cycle.

The method of mode 12, wherein the subject performs respiratory exercises, such that tumor positions for different respiratory volumes can be determined.

The method of mode 39, wherein the respiratory exercises include maximum inhalation.

The method of mode 39, wherein the respiratory exercises include maximum exhalation.

The method of any one of modes 12 to 15, wherein the initial measures are used to build a tumor position model (180).

The method of mode 42, wherein a tumor position model (180) intrapolates between a plurality of points of estimation for a tumor along a tumor trajectory (187) to build a trajectory model with respect to respiratory metrics.

The method of mode 42 or 43, wherein multiple rounds of initial measurements are made and the combined to build a tumor position model (180).

The method of mode 43, wherein the tumor trajectories are different for inhalation and exhalation.

The method of mode 1, wherein the sampling of measurements and the estimation of a second location of the tumor are performed in real-time.

The method of mode 1, wherein a second location of a second tumor is estimated also. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 33/ 39 The method of mode 47, wherein an electric field is applied targeting a location that is displaced in the direction of the estimated second location of a second tumor.

The method of mode 48, wherein the displacement is between 5 mm and 20 mm.

The method of mode 47, wherein the application of electric fields alternates between targeting a second location of the tumor of mode 1 and the second location of a second tumor.

The method of mode 47, wherein the tumor of mode 1 tends to move a longer distance during the respiratory cycle than the second tumor, and wherein the directions from which application of electric fields are applied is configured such that the second tumor is subject to electric fields with the higher field strength of the directions for more than half the time.

The method of mode 1, wherein a sensor (195) comprises one of respiratory inductive plethysmography, chest galvanic skin response and chest wall movement.

The method of mode 1, wherein the method is performed in an ambulatory setting.

The method of mode 1, wherein the second location of the tumor is estimated along the superior-inferior axis of the subject's body only.

The method of mode 8, wherein the burst length is between 10 ms and 1 second.

The method of mode 55, wherein the burst length is 250 ms.

The method of mode 1, wherein the metrics comprise respiratory period, and the electric fields are configured to be delivered in bursts of a certain lengths separated by separation of certain lengths, and the electrodes on the subject's body do not permit targeting a tumor along its full expected trajectory, wherein burst separations are configured to occur while a tumor cannot be effectively targeted.

A method for applying alternating electric fields to a target area of a subject's body, the method comprising: applying an electric field between a first pair of two sets of electrode elements targeting a first location of a subject's body, sampling measurements from one or more sensors (195) at a rate higher than one measurement per second, estimating a second location based on the sampled measurements, and applying an electric field between a second pair of two sets of electrode elements targeting the second location of a subject's body.

The method of mode 58, wherein the measurements are sampled at between 1 Hz and 10 kHz. The method of mode 58, wherein the measurements are sampled at between 10 Hz and 10 kHz. The method of mode 58, wherein measurements are made between bursts when no electric fields are applied. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 34/ 39 The method of any of modes 58 to 61, wherein an electric field contains a frequency between 50 kHz to 500 kHz, with a field strength of at least 1 V/cm. he method of mode 62, wherein the electric field contains a frequency between 100 kHz and 300 kHz, with a field strength of at least 1 V/cm.

The method of mode 62, wherein the electric field has a frequency at 150 kHz.

The method of any of modes 62 to 64, wherein the electric field contains essentially a single frequency component.

The method of any of modes 62 to 64, wherein the electric field comprises a further AC electric signal with a frequency of between 50 kHz and 500 kHz.

The method of mode 66, wherein the electric field comprises frequencies at 200 kHz and 300 kHz.

The method of any of modes 66 or 67, wherein the electric field contains essentially two frequency components.

The method of any of modes 62 to 64, wherein the electric field contains three or more frequency components.

A method of creating a map of electrode locations and configurations, wherein the electrodes are configured to impose an alternating electric field in targeted tissue, comprising the steps of: generating a model of a tumor trajectory (187) for a tumor's movement within the subject's body, identifying a plurality of tumor positions along the trajectory (187), and determining a location map and configuration for a set of electrodes (130) that permit effective electric fields to be targeted to each identified tumor position by activating sets of electrode elements in the electrodes.

The method of mode 70, wherein a 3D model of a part of a subject's body is constructed.

The method of mode 71 to 70, wherein a 3D model is based on one or more images gathered using one or more of the magnetic resonance imaging, computer tomography or ultrasound techniques.

The method of any of modes 71 to 72, wherein the tumor trajectory is estimated using the 3D model.

The method of mode 70 or 73, wherein a model of a tumor trajectory (187) comprises positions of a tumor during at different stages of the subject's respiratory cycle.

The method of mode 74, wherein a model of a tumor trajectory (187) comprises tumor positions when the subject fully inhales and fully exhales. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. / 39 The method of any one of modes 70 to 75, wherein an effective tumor treating fields therapy comprises an electrical fields strength of between 1 and 5 V/cm within 20 mm of a centroid of a tumor, or within 20 mm of a clinical tumor volume (182).

The method of any one of modes 70 to 76, wherein the plurality of identified tumor positions comprise positions that together permit the entire trajectory (187) to be effectively targeted. The method of any one of modes 70 to 77, wherein the plurality of identified tumor positions can effectively target at least 80 % of the length of a target trajectory (187).

The method of any one of modes 70 to 77, wherein the plurality of identified tumor positions can effectively target a tumor trajectory up to 13 cm long.

The method of any one of modes 70 to 76, wherein the plurality of identified tumor positions comprise positions that together permit a tumor to be targeted during at least 80 % of the respiratory period.

The method of any of modes 71 to 72, wherein electrode locations are identified using the 3D model.

The method of any of modes 70 to 81, wherein two sets of electrode elements are configured to effectively target a tumor when the tumor is in a position between the two sets, wherein the two sets are separated by at least 10 mm, and wherein the distance between the two sets of electrode elements is minimized.

The method of any of modes 70 to 81, wherein two sets of electrode elements are configured to effectively target a tumor when the expected field strength in a tumor is at least 1 V/cm from an electric field applied between the two electrodes.

The method of any of modes 70 to 81, wherein two sets of electrode elements are configured to effectively target a tumor when the expected power density in a tumor is at least 1 mW/cm3 from an electric field applied between the two electrodes.

The method of any one of modes 70 to 77, wherein the trajectories (187) of two or more tumors are considered when identifying tumor position.

The method of any one of modes 70 to 85, wherein a location map is determined by using several electrodes (130) of the same dimensions and configuration.

The method of mode 86, wherein the several electrodes (130) each contains 3-15 electrode elements.

The method of any one of modes 70 to 85, wherein a location map is determined by adapting the dimensions and configuration of one or more electrodes (130).

The method of any one of modes 70 to 85, wherein a location map is determined by adapting the number of rows of electrode elements (184) of an electrode (130), wherein each row has the same width. 36/ 39 90. The method of mode 89, wherein each row is equally spaced. 91. The method of mode 89 or 90, wherein each electrode element has the same size. 92. The method of any one of modes 70 to 89, wherein the step of determining a location map comprises the step of identifying one or more sets of electrode elements (184) in one or more electrodes (130) that together form a virtual electrode. 93. The method of any of modes 70 to 92, wherein the location map is configured to be printed. 94. The method of any of modes 70 to 92, wherein the location map is configured to be shown on one or more electronic displays. 95. The method of any of modes 70 to 92, wherein the location map is configured to be generated as part of a piece of clothing comprising one or more electrodes. 96. A personal computer or server configured to execute a program that implements a method according to any one of modes 70 to 95.

Closing comments

[0159] While this specification contains many implementation details, these should not be construed as limitations on the scope of the invention or of what may be claimed, but as descriptions of features specific to implementations of the invention. Certain features that are described in this specification in the context of separate embodiments can also be implemented in combination in a single embodiment. Conversely, various features that are described in the context of a single embodiment can also be implemented in multiple embodiments separately or in any suitable sub combination. Moreover, although features may be described above as acting in certain combinations and even initially claimed as such, one or more features from a claimed combination can in some cases be excised from the combination, and the claimed combination may be directed to a sub combination or variation of a sub combination. Thus, unless explicitly stated otherwise, or unless the knowledge of one of ordinary skill in the art clearly indicates otherwise, any of the features of the embodiment described above can be combined with any of the other features of the embodiment described above. Thus, many variations to the above examples lie well within the scope of the attached claims and within the capabilities of a person having ordinary skill in the art.

Claims

Claims A method for applying tumor treating fields to a tumor of a subject's body, the method comprising: applying an electric field between a first pair of electrodes targeting a first location of the su bject's body, sampling measurements from one or more sensors (195), estimating one or more metrics including a metric related to the subject's respiratory phase based on the sampled measurements, estimating a second location of the tumor based on the metrics, applying an electric field between a second pair of electrodes targeting the second location of the subject's body. The method of claim 1, wherein the measurements are sampled at a rate higher than one hertz. The method of claim 1, wherein the measurements are sampled at a rate higher than ten hertz. The method of any one of the preceding claims, wherein the metrics comprise a measure of respiratory volume. The method of claim 1, wherein during application of an electric field targeting a location, the field is applied in a sequence between two pairs of electrodes targeting the location. The method of any one of the preceding claims, wherein one or more of the measurements or metrics include a measure of measurement uncertainty. The method of claim 1, wherein the estimation of a second location of a tumor is configured to consider measurement uncertainty or missing data during the estimation. The method of claim 1, wherein the metrics comprise respiratory period, and the electric fields are delivered in bursts of a certain lengths separated by separation of certain lengths, and the 38/ 39 lengths of bursts and lengths of separations are configured to fit an integer number of bursts and separations into a respiratory period. The method of claim 8, wherein the burst lengths are of a pre-specified length, and the separation lengths are configured to fit the bursts and separations into the respiratory period. The method of claim 8, wherein the metrics comprise an inha|ation period and an exhalation period, and the bursts are configured to be synchronized with the inha|ation and exhalation periods by varying one or more of the burst lengths and the burst separations. The method of any one of the preceding claims, wherein the estimated second location of a tumor includes a surrounding clinical tumor volume. The method of claim 1, wherein the estimation of a second tumor location considers initial measurements that are made to determine the location of a tumor in the subject's body. The method of claim 12, wherein the initial measurements are made using at least one of magnetic resonance imaging, computer tomography or ultrasound. The method of claim 12, wherein the initial measurements include sampling measurements that comprise a signal for respiratory phase. The method of claim 14, wherein some or all of the initial measurements are made while applying electric fields.