Clot aspiration guidance from angiography

Abstract

An image-processing system (FS) and related method for facilitating removal of an obstruction (EM ) in a system of conduits (CN) through which fluid (FLU) is passable. The system may comprise an input interface (IN) capable for receiving input imagery obtainable by an imaging apparatus (IA) of the system of conduits (CN) including the obstruction. The system of conduits (CN) with the obstruction (EM) and the fluid (FU) therein forming together components of a fluid system (FLS). An estimator component (EC) is capable of estimating, based on the input imagery, an estimate (m) of a response of the fluid system (FLS) to a suction action causable by operation of an in-conduit obstruction removal device (ORD).

Description

[0001] CLOT ASPIRATION GUIDANCE FROM ANGIOGRAPHY

[0002] FIELD OF THE INVENTION

[0003] The invention relates to an image-processing system for facilitating obstruction removal in a system of conduits through which fluid is passable, to a related method, to a related arrangement, to a computer program element, and to a computer readable medium.

[0004] BACKGROUND OF THE INVENTION

[0005] Stroke, thrombosis and similar ailments afflict large parts of the population, particularly the elderly. Common to such types of conditions is an obstruction in blood vessels by a mass of material, such as blood clot or other. X-ray based imaging is an imaging technique that allows acquiring imagery of the internal organs or anatomies of a patient. Thus, the obstruction may be located in the afflicted blood vessel by X-ray based imaging, such as in angiography or fluoroscopy, or similar modalities.

[0006] In addition, such obstructions may be removed by using an aspirator device that can be placed inside the afflicted vessel. Suction may be applied to assist clot removal or removal of other material that may cause health problems. A, if not the most common, technique of clot removal to date is the use of a stent retriever device that entangles or entraps the clot in its mesh after its positioning through the clot, to so mechanically dislodge the clot and pull it away from its position for evacuation outside the vessel. To avoid debris of the clot breaking off and flowing off distally, suction through the aspirator device (e.g. via a guide catheter) is applied during the mechanical pulling action of the stent retriever to also evacuate such broken off clot debris, if any.

[0007] Direct contact clot aspiration methods are increasingly used in various clinical domains like neuro / stroke and venous procedures. New devices like the Penumbra™, lightning system are being released to support this trend. Non-contact aspiration through guiding catheter and balloon guide catheters is often used in addition to stent retriever thrombectomy and direct contact clot aspiration in order to avoid debris being washed further downstream as mentioned above in connection with the sent retriever. A typical question is if a balloon guide (which blocks the artery proximally to the aspiration) is needed or if the aspiration through a standard guide catheter is sufficient to avoid distal flow. Another choice that must be taken is the size of the devices, eg, its diameter, or other specifications such as suction power, etc. While large devices have the advantage of allowing stronger aspiration, they cannot be placed in small lumen vessels.

[0008] For all aspiration techniques, the placement and selection of the devices, the aspiration power, and the duration of aspiration are parameters that should be optimized for the specific procedure. In the previously mentioned Penumbra™ device, aspiration is optimized by a sensor that measures the properties of aspirated blood in order to stop the aspiration automatically once fresh blood is being sensed / extracted.

[0009] Selection and placement, as well as application of aspiration devices, is currently done without guidance, based only on operator experience. Incorrect choices may lead to distal embolization, damage to vessels, or other adverse responses, and thus the patient may come to harm which is undesirable.

[0010] SUMMARY OF THE INVENTION

[0011] There may therefore be a need for improved obstruction removal in image-guided procedures.

[0012] An object of the present invention is achieved by the subject matter of the independent claims where further embodiments are incorporated in the dependent claims. It should be noted that the following described aspects of the invention equally apply to the related method, to the imaging arrangement, to the computer program element and to the computer readable medium.

[0013] According to one aspect there is provided an image-processing system for facilitating removal of an obstruction in a system of one or more conduits through which fluid is passable, comprising: an input interface capable for receiving input imagery obtainable by an imaging apparatus of the system of one or more conduits including the obstruction, the system of one or more conduits with the obstruction and the fluid forming together components of a fluid system; and an estimator component capable of estimating, based on the input imagery, an estimate of a response of the fluid system to a suction action causable by operation of an in-conduit obstruction removal device.

[0014] In embodiments the received imagery is X-ray imagery.

[0015] In embodiments the imaging apparatus is configured for any one or more of: spectral imaging, X-ray imaging. Eg, fluoroscopy, angiography may be used. Input imagery may be reconstructed 3D / section imagery or projection imagery. "Imagery" may include more than one image instances (frames). Preferably, projection imagery is used.

[0016] In embodiments the imagery is contrasted based on presence of contrast agent in a field of view, FOV, of the imaging apparatus.

[0017] In embodiments the system includes a segmentor capable of providing the input imagery as segmented imagery including a segmentation (,s ) for at least a part of the system of one or more conduits.

[0018] In embodiments the system of one or more conduits includes one or more (side) branches, such as collaterals, etc, branching off in a region of interest, ROI, including the object.

[0019] In embodiments the estimator component is capable of so computing the estimate based on a computational fluid dynamics model, the model so based on the segmentation and, optionally, on one or more characteristics of the (at least one) in-conduit obstruction removal device. Such parameters may include type, size (diameter), placement, suction power, etc.

[0020] Optionally there may be input in form or hemodynamic properties like internal pressure or flow measurements e.g. from contrast agent dynamics or through the catheter.

[0021] In embodiments the computational fluid dynamics model is based on any one or more of: a finite element model, a lumped resistor and flow model.

[0022] In embodiments the computational fluid dynamics model includes a component representative of the suction exertable by the suction device.

[0023] In embodiments the estimator is based on a trained machined learning model.

[0024] In embodiments the said response is any one or more of: i) disintegration deformation or motion of the object, ii) flow of fluid away from the obstruction removal device, iii) flow of fluid towards the obstruction removal device, iv) a deformation of at least a part of the system of the one or more conduits.

[0025] For example, a response of the fluid may be flow towards the suction device or flow away from the suction device, or a changing flow in the process of the obstruction removal operation. In embodiments the system includes a graphics display generator capable of generating a graphics display for displaying on a display device, the graphics display including an indication of the estimate so computable.

[0026] In embodiments the system includes a recommender capable of outputting, based on the estimate, or on further such estimate, or on the input imagery, a recommendation in relation to the obstruction removal device.

[0027] In embodiments, the system of the one or more conduits includes at least one of: i) a blood vessel, ii) urinary system, iii) digestive system, iv) endocrine system. For, example, instead of clots or such deposits in blood vessels, calculi, in kidney, liver, gall bladder or ureter, etc, may be removed herein.

[0028] In embodiments, the said object includes a thrombus or embolus, clot, or other mass of deposits, etc.

[0029] In embodiments, the obstruction removal device, or a different such obstruction removal device, is capable of exerting, in addition to such suction, a mechanical contact-based operation on the object to so further facilitate its removal.

[0030] Aspirator-based obstruction removal device may act as facilitator for such removal, together with mechanical remover. They may act together in an integrated device or may act together as separately. They may through a common gateway, such as the same guide catheter, or via separate gateways, catheters, etc.

[0031] In another aspect there is provided an arrangement comprising at least a part of the system of any one of the previous claims, and further comprising any one or more of: i) the imaging apparatus, ii) the obstruction removal device, iii) a storage element including the model as per any of the above embodiments.

[0032] In yet another aspect there is provided a computer-implementable method of facilitating image-based obstruction removal in a system of one or more conduits through which fluid is passable, the method comprising: receiving input imagery obtainable by an imaging apparatus of the system of one or more conduits including the obstruction, the system of one or more conduits with the obstruction and the fluid forming together components of a fluid system; and estimating, based on the input imagery, an estimate of a response of the fluid system to a suction action causable by operation of an in-conduit obstruction removal device.

[0033] In yet another aspect there is provided a computer-implementable method of training the machine learning model.

[0034] In yet another aspect there is provided a computer-implementable method of generating training data for use in the ML based method.

[0035] In yet another aspect there is provided a computer program element, which, when being executed by at least one processing unit, is adapted to cause the at least one processing unit to perform any one of the method above.

[0036] In yet another aspect there is provided at least one computer readable medium having stored thereon any one or more of: i) the program element, ii) the machine learning model, and iii) the computational fluid dynamics model.

[0037] The estimator component may operate automatically, or semi-automatically. In the latter case, the facilitator system may support a user interface through which the estimator component is operable together with user to build, in a first (preparatory) phase, a model, such as a fluid dynamics ("FD") model, based on the input imagery, instead, a machine learning based model may be used. In a second, deployment phase, the model is then used ("run") as in a simulation or other to compute fluid flow pattern of the fluid around or , more commonly, in the area of the obstruction,, in order to compute the response of, for example, the obstruction to operation of the obstruction removal device.

[0038] Instead of, or in addition to, estimating / predicting the obstruction's motion response, a response of the fluid flow pattern in general in the conduit in the area of the obstruction and or the aspirator device may be sufficient as in many cases there is no flow through or around the obstruction, such as in an occlusion with complete blockage so there may be no motion response as such by the obstruction. A response of the obstruction in terms of disintegrating into debris rather than the dislodging and motion of the obstruction itself as a whole by the applied suction may also be of interest herein. In addition, or instead of, estimating response of fluid or obstruction, such response of the system of conduits itself may be of interest in some cases as too strong a suction applied may cause a partial vacuum in places of the conduit which may cause the collapse of the conduit system at places, thus causing damage. Such collapse may be adverse to the objective of removal of the obstruction. Thus, what is proposed herein more generally is to estimate the response of the whole or parts of a fluid system, an aggregation of interacting components formed by the conduit, and the fluid and obstruction in it.

[0039] The estimate is preferably to distinguish favorable vs adverse responses of the fluid system as a whole, or of any one or more of its components. The response to aspirator suction is favorable if, for example, such obstruction does not disintegrate into debris / parts, or if such pattern dislodges and causes the object to move towards the aspirator, or conduit does not collapse, or flow patten of fluid is directed towards rather than away from aspirator without necessarily causing the obstruction to dislodge and being carried by such flow towards aspirator, etc. The latter case may still be favorable for mechanically dislodging and removing obstruction as in this case it is at least the smaller debris pieces that will be carried by favorable flow pattern to aspirator to be so suctioned off by aspirator rather than escaping distally or otherwise away from aspirator. Adverse response is the opposite of any of the foregoing favorable responses. Thus, what is proposed herein is a system or method for n estimating whether the application of the suction by aspirator causes a favorable or adverse environment / response of the fluid communication system n as a whole or any one or more of components. Such response is deemed favorable if is conducive to a safe removal of the obstruction. This can inform therapy decision by the clinician at the outset or during the intervention. The opposite of such favorable response is adverse response. A reference herein to the fluid system shall be construed as a reference to any one or more of its components, or to the fluid system as a whole.

[0040] The estimated response may be made individually for some or each such components, in which the estimate may include multiple estimates, at least one for each component. Alternatively, the response estimate may relate to the fluid system as a whole. The response estimate may be displayed or otherwise provided individually or concurrently as needed. The individual estimate for some or each component may be combined to obtain a global estimate, or such global estimate is provided natively without producing as intermediate data such individual respective estimate for some or each component. The in-conduit aspirator system has at least a part through which the suction is applied situated within the conduit system. The said part may be situated anywhere, so long as there is fluid communication with the obstruction. The FOV (field of view) is preferably large enough to capture imagery or at least one of the obstruction and or the said part of the aspirator, or preferably both. However, not both of the said part and the obstruction may necessarily be within the FOV of a single frame. Multiple frames may be needed to cover such a FOV.

[0041] It has been found that when placing a larger aspiration device in a proximal location, whilst desirable in some cases, there may be remain more side branches between the device and the location of obstruction (eg, clot). This may increase the risk of collateral flow. The proposed system and method allow better gauging whether the intended fluid system presents and adverse or favorable response to a given selected aspirator size. The system may thus allow finding the largest size that still is consistent with a favorable response. Alternatively, and as briefly mentioned, the first phase includes training a machine learning ("ML") model and this is then deployed to process input imagery to estimate the fluid system 's response. Training or model building in the first phase can be automatic, without user involvement. The second phase is in general automatic, but, again, some user involvement in tweaking estimated results or tweaking input parameters such as the aspirator's specification (size, power, etc) to obtain new estimates is indeed envisaged herein.

[0042] Indeed, in some embodiments, in addition to the input imagery, further input may include one or more operational parameters of the obstruction removal device and / or location in the conduit of the obstruction. The operational parameters may include the location of at least the in-conduit part of the obstruction removal device and / or the suction power to be applied, its size (in particular in lumen diameter of the suction tube, etc). The said location of obstruction removal device may be determined from the input imagery together with the location of the obstruction. Alternatively, the location of the obstruction removal device is specified, eg by user or is retrieved from protocol information, if the obstruction removal device is not yet placed.

[0043] Thus, the estimator may determine one or more obstruction location and / or location of the obstruction removal device automatically from the input imagery, or this is provided as separate data items in addition to the input imagery. In general, the input imagery provides spatio topological information of the conduit system's configuration from which the fluid system's response (such obstruction's response etc) is estimated, given locations in the conduit system of obstruction removal device and of the obstruction (eg clot).

[0044] The obstruction removal device is mainly envisaged herein as an aspirator in-situ that can be placed in-side a blood vessel (being one example of such a conduits system with fluid therein).

[0045] In embodiments, a segmented vessel model from angiographic images is used to extract information about vessel size, number of local branches, possible collateral flow, and clot size. A fluid dynamics model may include a dedicated aspiration model component (eg, a dedicated model element for a source of negative pressure) may be used to suggest based on the estimate, improved or optimal placement of aspirator device, and selection of device types, as well as aspiration power, volume and / or duration of application of suction or other such operational parameters.

[0046] The proposed setup allows avoiding expensive in-situ sensor-based approaches and is mainly, if not exclusively, image based. What is proposed herein in embodiments is to use a segmented vessel model from angiographic images to guide aspiration techniques. Information about vessel size, number of local branches, possible collateral flow, and clot size may be used to optimize, or at least improve, placement, selection, and application of aspiration devices. Such improvement allows avoiding, or at least reducing, a risk of distal embolization which is one example of an adverse response of the fluid system to the obstruction removal (eg, clot) at hand. Distal embolization may be the result of adverse flow, that is, a flow component of fluid that carries the obstruction away from the aspirator device during an obstruction removal operation, as may happen under certain conditions. Thus, whilst removal of obstruction is intended, the obstruction may nevertheless evade such removal as the context may cause an adverse flow pattern to emerge when an interaction (by suction or mechanical) with the obstruction is performed. The said context may include manners of operation of the aspirator device (such manners being related to operational parameters) and the topological configuration of the vessel at or around the obstruction site such configuration being one example of a configuration of the fluid system. The aspiration model component used herein in embodiments may represent one or more of the said operational parameters.

[0047] The proposed system and method can be used with benefit to support stent retriever based thrombectomy, a widely used technique as mentioned above. The proposed system and method can be used with aspirator supported mechanical obstruction removal technique. Thus, as used herein, the aspirator may itself act to remove the obstruction, or the aspirator acts in concert with such mechanical removal, such as by retaining debris that may form during such removal operation, or in any other way that facilitate safe removal of the obstructing object / obstruction.

[0048] The proposed system and method may be used for neuro / stroke or venous procedures, or for other clinical applications, such as pulmonary embolism, coronary thrombus, and others. However, non-medical applications such as the examination of inaccessible subterranean river or drainage systems or inaccessible hydraulic or plumbing systems in a building or in vehicle such as a spacecraft, ship, car or other machinery is not excluded herein.

[0049] "favorable " / "adverse" as used herein are qualifiers that relate to a response of the fluid system in general. In general, the response is favorable if it relates to an environment or condition(s) which are conducive to a safe removal of the obstruction, fluid system's response is adverse, if it is not favorable. In particular and as an example, an adverse response of the fluid is a flow pattern directed away from the aspirator's suction tube, whilst a favorable response if the fluid is a flow pattern directed at least in parts towards aspirator's tube. An adverse response of the system of conduits is it at least partial deformation (eg collapse) during obstruction removal. An adverse response of the obstruction is not moving towards the aspirator (such as a failure to dislodge), or its disintegration. Favorable responses may include the opposite of the aforesaid.

[0050] For example, in more detail, flow pattern is favorable, if there is at least one flow component in a direction parallel (not merely anti- parallel) to a longitudinal axis of the suction tube of the aspirator, or to a vector of the suction flow component in the tube of aspirator. Other definitions may be used herein.

[0051] "fluid" is any matter capable of flowing, such as liquids, but the term as envisaged herein may also include gaseous matter, such as air or any other.

[0052] "conduit (vessel) tree" is a projection view or 3D reconstruction of the spatial configuration of the conduit system, such as of all or part of blood vessel system or other system of conduits. "obstruction" may include any impediment to flow of a fluid in a conduit system such as blood in a vessel tree. Obstruction may include an occlusion where there is no downstream / distal flow, or a situation where there may be some reduced flow remaining, as compared to when obstruction is removed. The term obstruction as used herein is a reference to the object / mass or thing (clot, embolus, etc) that causes a state of being obstructed.

[0053] “contrasted / non / un-contrasted” imagery refers to contrast agent assisted imaging (eg, angiography or MRI). Contrast agent is released into the obstructed vessel / conduit system. Contrasted imagery with contrast agent present in the field of view of imager is acquired. In addition, and in some embodiments, non / un-contrasted imagery is acquired without, such contrast agent being present. Suitable contrast agent (materials) may include Iodine, Gadolinium, or others.

[0054] BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Exemplary embodiments of the invention will now be described with reference to the following drawings, which, unless stated otherwise, are not to scale, wherein:

[0056] Fig 1 shows a medical imaging arrangement, as may be used herein in embodiments;

[0057] Figs. 2A-2C show configurations of a fluid system including an obstruction in a system of conduits;

[0058] Fig. 3 shows a block diagram of a computer implemented system for assisting operation of an in-conduit obstruction removal device;

[0059] Figs 4A-4B illustrate various instances of a fluid dynamic model of the lumped- flow type as may be used in embodiments herein; and

[0060] Fig. 5 shows a flow chart of a computer implemented method of facilitating operation of an obstruction removal device, such an aspirator of the direct or indirect contact type.

[0061] DETAILED DESCRIPTION OF EMBODIMENTS

[0062] Turning first to Fig. 1, this shows a block diagram of a, in embodiments, medical imaging arrangement MIA. The medical imaging arrangement MIA is configured to facilitate image-guided removal of an obstruction EM in a region of interest in a system of conduits. In the medical realm, such a system of conduits may be a system of vessels of the human or mammal heart, or of the brain, or of other anatomies or organs. However, other medical applications, and even non-medical applications, such as the removal of obstructions in inaccessible pumping or hydraulic systems are not excluded herein. But having said that, the following will mainly focus on medical applications for the removal of an obstruction EM such as embolus, thrombus, clot, etc, lodged in a blood vessel system, such as in the heart coronaries, or in blood vessels in the brain, or in a limb, neck, etc.

[0063] Broadly, the medical imaging arrangement MIA may include an imaging apparatus IA operable to acquire, for example, still imagery, or a stream of fluoroscopic or angiographic imagery, comprising frames acquired over time of the relevant anatomy of the patient, such as a region of interest (ROI) where the obstruction is broadly located. Thus, the ROI may include the location of the obstruction and its surroundings, including branches of the vessel system CN, etc. Thus, the imaging arrangement facilities obstruction removal, to so restore (or at least increase) at the affected site flow in a blood vessel or other conduit system.

[0064] More specifically, such facilitation may be enabled by a computing-based facilitator system FS of the medical imaging arrangement MIA. The computing-based facilitator system FS is capable of providing guidance for operation of an obstruction removal tool / device ORD, such as a direct contact or indirect contact aspirator, optionally in conjunction with a mechanical removal tool MR.

[0065] Broadly, such aspirator device ORD may be operable to use suction to remove the obstruction or support its removal. A tube portion thereof, may include an action tip, through which the suction is applicable. The tube portion may be introduced into the vessel CN through an access point AX via a catheter system or some such. A guide catheter itself may be used as such a tube potion. The tube portion may then be advanced though the vessels system preferably under a fluoroscopic guidance and / or angiography guidance by imager IA, until the action tip comes to he in a vicinity, sufficiently close to where the obstruction is located. Again, fluoroscopic guidance and / or angiography guidance may assist user or robot in this navigation task. The said vicinity or "action radius" may depend on the functioning, type of obstruction removal device ORD, the kind of removal procedure envisaged by user / protocol, type or thrombus etc. In fact, the proposed facilitator may assist in embodiments finding a correct "parking" distance (position) from the obstruction, the said position being one of the device ORD's operational parameters, as will explored more fully below.

[0066] It has been found that sometimes the introduction and very presence of the device ORD in the obstruction's vicinity, and / or in particular its operation and / or any additional interaction to remove the obstruction may have adverse unintended effects. For example, the obstruction EM (which may be referred to herein as a clot for exemplary definiteness) may disintegrate on application of suction, or during a mechanical interaction with the clot. For example, such actions may, counterintuitively, cause the clot to move, in a distal direction, away from the action tip through which the suction is applied. So, in the latter case, a dislocation there may have been achieved, alas the clot EM still escapes removal as it may be urged "deeper" into the vessel where it may be harder to reach and extract still. This effect may be viewed as counterintuitive, but can happen due to the fluid dynamical disturbances, or flow pattern, in the surrounding blood pool. Such flow patterns may conspire unfavorably to cause a net distal blood flow component which may cause the clot being dislodged, if at all, but at any case carried further downstream into the vessel system CN making it potentially harder to remove as a result. This is particularly the case if the clot disintegrates into (smaller) debris parts during mechanical interaction. Also, on occasion, even though suction is applied, the debris or the clot itself may be carried away from aspirator or mechanical removal tool due to adverse flow or pressure patterns emerging in the vessel.

[0067] Thus, what is proposed broadly herein is the facilitator system FS reading in X-ray based imagery of the ROI as provided by the imager IA, for example an angiographic still image or a time sequence of angiographic frames. The facilitator FS may process such imagery to compute, based on such processing, a result m. The result m may be an indication for a response of the clot EM, or the response of the fluid system FLS more generally. The fluid system as used herein relates to the system of conduits CN, with the fluid (eg blood) FLU and the obstruction (eg clot) EM in it, thus forming three components of the fluid system FLS (there may be more than tree such components). Thus, the result m may include for example an estimate / prediction of a likely response of the clot EM to operation of the device ORD, with or without mechanical removal assistance MR. For example, the result may include an indication on a probability for such an adverse event to occur. Such adverse event may include, as said, clot disintegration and / or a net distal flow component to emerge that would force the clot or its debris away from the aspirator's tip rather than towards it, or any other such event that goes counter the intended objective of obstruction removal. As another example, the vessel / conduit system may sustain damage such as collapse, due to an unfavorable pressure distribution caused by the suction. In general, any one or more of the above events may be collectively referred to herein as a response of the fluid system, and such response may be adverse. The indication may indicate a probability or binary indication or a scored for anyone, more than one, or all such events, and whether they are likely to be adverse or favorable. Thus, the indication m= (ml,m2m3) may be a triple, or an w-tuplc (n>3), each mj indicative of whether any one of more of conduit, fluid or obstruction may respond unfavorably to suction, and, optionally, in addition, to mechanical removal. The indication may take any form as needed, such as in graphical form, numerical form, or simply a sensory alert or signal (audio, haptic, etc).

[0068] Broadly, the response of the fluid system FLS may be function of the particular spatial configuration of the vessel in the said vicinity (also referrable herein as action radius). The vicinity comprises the location of the tip of the device and the location of the clot. The response may be a further function of one of more the said operational parameters of the obstruction removal device. Thus, said parameters may include any one or more of manner of placement of device in the vicinity of obstruction, its geometry (eg, size, diameter), the suction power applicable, or use of any features thereof. For example, some aspirator device ORD types may have, as such a feature a "neck" portion, an inflatable / deflatable balloon catheter for sealing off parts of the vessel "behind", that is, distal, of the tip TP. Such deployment of balloon catheter may be used to influence the emerging flow or pressure pattern ( adverse vs favorable , to be described more fully later below, at Figs 2A-2C) that is caused when suction is applied by the device ORD. Herein, any parameters of the obstruction removal device ORD that may have an impact on the success of the operation, may be referred to herein collectively as the "operational parameter(s) P" of the device ORD.

[0069] It will be understood that operation of the facilitator FS is preferably predictive and / or preventive in nature. Thus, the facilitator FS may be processing, via its estimator component EC, the input imagery to supply output m before suction is actually applied. Preferably however, the facilitator FS may be operable after placement of the obstruction removal device ORD, or whilst such placement is ingoing. For example, the facilitator FS may be providing such indication results m=mtwhilst the device ORD is being advanced towards the clot EM. The imagery or field of FOV(s) of the imager is such that is captures at least a part of device ORD's tip and the clot location. Such image capture may not necessarily happen in a single frame. Thus, the information on device ORD placement / advancement, and clot location may be spread across more than one frame. Alternatively, the indication(s) are computed repeatedly as a stream m=mtduring an ongoing application of suction, thus the said stream of indications mt are correspond to possibly dynamically changing responses of the fluid system FLS. In the following we will refer herein merely to indication / result m, whether or not, it is stream or a still data, unless such distinction is necessary.

[0070] In addition, if the result m provided by facilitator FS is indicative of an adverse response by the fluid system FLS, the facilitator may provide, on user request or automatically, without such request, additional output m'. Such output m' may include one or more recommendations for new one or more operational parameters. Such may include a suggestion of re-positioning of device ORD, or of different settings thereof (eg, suction power, use of balloon catheter yes / no, etc). Such recommended parameters are such that they would likely attract a favorable fluid system FLS response as opposed to an adverse one associated with the current operational parameters, to so help user avoid an adverse fluid system FLS, such as an adverse clot response, or of one or more of the other components of fluid system FLS. The recommended new operational parameters may be provided by a recommender RC module. The recommender RC module may sample through variations of the current operational parameter sand effect recomputing by an estimator EC of facilitator FS of such new / updated estimate m', based on the input image X as before, but now using new operational parameters, etc. Such new, sampled parameters may be generated by adding or subtracting in increments perturbations A to some or each such parameters, and have for each the estimate recomputed. A feedback loop may be established between recommender RC and estimator EC to vary the operational parameters until a favorable response is elicited. The variations of operational parameters (collectively referred to herein as " ") into new parameter P' = AP+P may include any one or more of changing position of aspirator device, such as moving its tip from the current in-vessel position to a new such in-vessel position. For example, the position may be moved back or forth, for example past a (side)branch SB, for example to so avoid adverse turbulences or other unfavorable flow or pressure distribution conditions. In addition, or instead, new estimates m ' may be computed with balloon catheter deployed or not. In addition, different negative pressure setting at different magnitudes to be applied by the device ORD may be so sampled. Any one or more of the parameters P may be sampled in isolation, with some other parameters P held constant, or two or more such parameters are varied simultaneously, or in any other variation pattern. A randomizer may be used to choose the pattern and / or the increment size in each or some such operational parameter "channel", such as for position, pressure setting, etc.

[0071] Any of the results, the clot response estimate m and / or related recommended new operational parameter m may be displayed on the display device DD under operation of a visualizer or graphics display generator VIZ. Visualizer VIZ may generate a graphics display GD which includes indication of any one of both of the results m, m'. Any one or both of the results m, m' may be displayed graphically, and / or alphanumerically, by text or numbers, as required. For example, the information may be combined with a current live image or stream as provided by the image apparatus IA.

[0072] The computerized facilitator system FS may be arranged in hardware or software, or partly in both, facilitator system FS may be integrated into an operator console OC of the imaging apparatus IA, or may be otherwise integrated in the machinery of the imaging apparatus or associated therewith. In other embodiments, the facilitator system FS may be implemented on one or more servers or other computing devices (eg, desktop computer), located remote from the imaging apparatus, or located in the exam room with the imager IA, but in a separate housing. The facilitator system FS may be partly run on such a server and / or in addition or instead on a mobile computing device such as a laptop, smartphone or tablet, or any other terminal user device. Whichever one or more computing devices may be used, such computing devices may include one or more data processor circuitry (chip), such as CPU, GPU or any other, and storage media for storing code instructions. Such code instructions so stored may be capable of causing executing the functions of facilitator FS described above and herein below. Distributed computing, or cloud-computing may be used, if required. Some or more of the functions may be performed by one single computing device or multiple such computing devices, may be used, as required and apt in the circumstances.

[0073] Before explaining operation of the obstruction removal facilitator system FS (referred to herein in brief as the "facilitator (system) FS") in more detail, reference is now made first to aspects of the imaging equipment as such, in order to provide context which may assist such explanation.

[0074] The imaging equipment may include the imagery apparatus IA, preferably configured for fluoroscopy or angiographic imaging as mentioned above. The imaging apparatus IA ("imager") is preferably X-ray based. The imager IA may be of the C-arm / U-arm type so may allow collection of projection raw data X' along different projection directions P(a) around the region of interest, such as around the mentioned stricture / obstruction in a coronary vessel. However, non-rational radiography equipment is not excluded herein.

[0075] The imaging apparatus IA includes an X-ray source XS and an X-ray sensitive detector XD. In some embodiments, the X-ray source XS, and opposite thereto, the X-ray sensitive detector XD, are arranged preferably on a rotatable gantry G. The gantry G may have a recess in which an examination region is defined where the patient, or in particular the region of interest, resides during imaging. The rotatable gantry may thus have a “C”, “U” or similar shape, such as in C-arm imaging apparatus envisaged herein. However, such or similar rotational or an otherwise dynamic setup, where there is at least some relative motion between source and detector intended, is not necessarily required herein in all embodiments. Instead, a planar radiographic setup is also envisaged in some embodiments. Instead of the rotatable setup as illustrated in Fig. 1, a bi-planar imaging setup with two detectors with intersecting imaging axes (for example at right angles) may be used. In addition, whilst digital-at-source imaging, such as flat panel detectors is preferred, where the imagery is natively acquired as digital data, such as setup is not necessarily required. More traditional, cassette-operable, analog imaging methods (based on X-ray films) are not excluded herein. The analog imagery may be digitized post-acquisition for example, by using a digital camera. Imaging based on X-ray image intensifiers, etc with post-digitization is also envisaged in some embodiments.

[0076] Generally, imaging includes energizing the X-ray source XS, so that an X-ray beam XB issues forth from a focal spot of the source XS, traverses the examination region (with the region of interest in it) and interacts with patient tissue matter. Interaction of X-ray beam XB with tissue matter causes the beam to be modified. The modified beam may then be detected at the detector XD. Conversion circuitry (not shown) converts detected intensities into digital imagery, in particular into projection imagery. The imagery may be used by user in the medical procedure. The patient may reside on a patient support PS such as a table during imaging. Soft tissue such as the one making up the coronary vessels, provide usually poor contrast in native X-ray imagery. In order to boost contrast, a contrast agent (sometimes “CA” or short), such as an iodine solution, gadolinium, or other, is delivered through an access point AX into the blood flow of the patient prior to or during imaging. The contrast agent CA may be delivered via a CA delivery apparatus (not shown), such as a motorized pump, but manual delivery is not excluded herein. The delivery apparatus may be operable to deliver a defined volume of contrast agent to the patient. The volume of contrast agent so delivered travels with the blood flow. After a certain period of time contrast agent accumulates at the region of interest such as in or around the stricture. Image acquisition is preferably synchronized with arrival of contrast agent at the ROI. Preferably “contrasted” imagery is acquired whilst there is contrast agent CA present at the ROI. Optionally, additional “noncontrasted” imagery is acquired whilst no contrast agent is present at the ROI.

[0077] The in particular contrasted imagery may be acquired in a time series with varying concentration of the contrast agent accumulating at the region of interest. Image acquisition may start before, may continue throughout, and may terminate after accumulation of the contrast agent at the ROI. The acquired stream of projection imagery may be visualized by a visualizer VIZ as a live video feed on display device DD. Alternatively or in addition, still imagery is so visualized.

[0078] X-ray imaging is generally based on the attenuation co-efficient of the matter(s) that makes up the region of interest. The attenuation (coefficient) is dependent on the energy of the X-ray beam. In order to harness this energy dependency, a spectral imager IA may be used, configured for spectral imaging. This may include detector-sided or source-sided solutions. Detector-sided solutions include photon-counting detector hardware or a dual -energy setup with dual -layer hardware for example. More than two detector layers may be used although this is less common. Source-side solutions include multi-source setups, kVp switching, filtration and other.

[0079] The imaging set-up in spectral imaging allows acquisition of multi-dimensional image data. Such multi-dimensional image data includes projection imagery at different energy levels, instead of acquiring energy integrated imagery as provided by a conventional (non-spectral) X-ray imaging set-up, the latter not being excluded herein. For example, such a conventional set-up may be used for digital angiography subtraction (“DAS”) where pairs of projection images at higher and lower concentration of the contrast agent is acquired. Such image pairs are then spatially registered and subtracted from each other to boost contrast of the soft tissue including the contrast agent. DSA may also be practiced with spectral imaging setup.

[0080] Preferably however, for more accurate quantification of the contrast agent concentration / volume at the region of interest, a spectral set-up is preferred. Such a spectral imaging setup includes a spectral image processor SP. The spectral image processor SP may implement a spectral image processing algorithm, such as material decomposition or other. A range of such spectral image processing algorithms are envisaged herein. The spectral processor SP processes the multi-energy projection imagery X acquired by the source- or detector-sided imaging set-up just described, and computes therefrom spectral imagery that differs from the original input multi -energy projection imagery. Spectral imagery is thus computationally derived from the input multi -energy projection imagery. Such spectral imagery, which may be in projection domain, may include for example, a monochromatic image at a desired energy level, or a virtual non-contrast image, or a contrast-only image (e.g. iodine image), a Compton scatter image, and others. The monochromatic image may correspond to the attenuation co-efficient of the contrast agent used. For example, the energy value or energy range for the monochromatic image may be chosen to equal or include the K-edge energy of the particular contrast agent to be used for good contrast for example. However, such as K-edge based choice is not mandatory herein and other energy values or ranges for the mono-chromatic imagery is equally envisaged herein. The monochromatic image represents a computational approximation of an image that may have been obtained had a monochromatic X-ray source XS at the respective energy level been used for imaging. In most practical cases, the actual X-ray source XS is polychromatic. The virtual non contrast (“VNC”) image represents an approximation of an image that may have been obtained if there had been no contrast agent present in the field of view of the imager IA, at the exclusion of other materials. The contrast (“VC”) image represents an approximation of an image that may have been obtained if there had been only contrast agent present in the field of view of the imager IA, at the exclusion of other materials. Thus, image contrast in VC is exclusively or pre-dominantly contrast agent focused. It is preferably the spectral (projection imagery) so computed by spectral processor SP that is processed by the facilitator system SYS, instead or, or in addition, to the original multi -energy input projection imagery, as will be described below in more detail.

[0081] For simplicity, the designation “X” may be used herein to indicate input imagery in general to be processed by the facilitator system FS. That is, “X” may relate to original projection imagery as acquired or, , to the spectral projection imagery as computed by the spectral processor SP based on the original projection imagery as acquired. The spectral processor SP , if used, and the facilitator system FS may be implemented on different computing systems, possibly geographically separated such as in a distributed “cloud” architecture. However, implementation on a single computing system is also envisaged herein. Imaging as envisaged herein is preferably atenuation-based, but this is not at the exclusion of other X-ray modalities, such as dark-field imaging and / or phase contrast imaging. Tomosynthetic or tomographic imaging is not excluded herein, due to the multi -directional imaging capability of the preferably rotational imaging apparatus IA. However, processing of projection imagery in projection domain is mainly envisaged herein in embodiments. Bi-plane imaging systems that have two detectors at an angle may also be used herein, instead of C-arm or fixed radiography system. Whilst spectral imaging has many advantages, the use thereof is not a necessity herein and more traditional angiographic approaches such as DSA or others are not excluded herein, indeed are specifically envisaged in some embodiments. Thus, energy integrating projection-based imaging is sufficient for present purposes, and spectral imaging may not be needed.

[0082] Referring now to the obstruction removal device ORD in more detail, such as a direct or indirect aspirator, this may include the mentioned tube portion / member TB having the action tip portion TP, which is placeable after the introduction into the conduit system CN, in the action vicinity within the vessel CN, in sufficient proximity to the location of the obstruction EM.

[0083] The aspirator device, for as such will the obstruction removal device ORD be referred to herein, may include a vacuum component, such as a pump, capable of generating negative pressure or partial vacuum. The vacuum component is in fluid communication with the mentioned tubular portion TM. Upon generating of the requisite negative pressure as per device seting, suction force is generated at tip AP, capable of engaging and drawing the clot out of its current position, or at least eliminating any debris when atempting to dislodge clot by mechanical remover MR. In one example of a favorable response of fluid system FLS, , the clot or debris thereof may be held by vacuum force against the tip and can then be safely removed. In other embodiments, the clot, if of appropriate size and constitution, may be sucked into the tube by operation of the negative pressure, and the material EM is evacuated outside the patient. It may be ejected into a receptacle for discarding or for further analysis, as required. Two types or modes of operations of such aspirator devices ORD are envisaged, which may be referred herein as the direct or indirect contact type. In the former, the action vicinity is zero as the tip is brought into contact with clot first is driven unto the clot, or even therethrough, and suction is applied then or whilst retracting the tip through the clot, etc. In the later, suction is applied at a distance and no prior contact of tip AP with clot EM is required. Additionally, a combination of two or more aspiration devices may be envisioned. In this case the facilitator considers interaction and operational parameters of some or all devices in its estimation of flow patern and obstruction response, or fluid system FLS response more generally.

[0084] A mechanical removal device MR, either different from, or integrated into, the aspirator device ORD may be used in addition to the aspirator. This second removal device MR may act through direct physical / mechanical contact, in particular engagement, with the clot. Such physical engagement device, referred to here as mechanical remover MR, may include a wire mesh or a grabber arm or other end effector, including opposed hinged member, such as hinged jaws or claws, openable and closable, which exert a mechanical grab or pinch action on the clot EM, thereby allowing securing the clot. The physical engagement device MR may then retract through a separate tube portion , or directly into an aspiration device, in order to assist the removal of the clot EM. If both aspirator and mechanical removal device MR may be used, both may be referred herein as an "obstruction removal device ORD" or "obstruction removal system ORD".

[0085] Use of such physical / mechanical contact engagement based removal device MR in combination with aspirator ORD may be of benefit, because the aspirator and its negative pressure is not directly occluded by the clot when the clot is first moved out of its initial position. During the first dislodgement of the clot the risk of (small) debris breaking off and flowing off distally is the highest. Having an uninterrupted suction and a consistent blood flow towards the aspiration device at this stage may be favorable. Having a large portion of the clot securely entangled in the physical engagement device gives user good control on its motion, and allows user to choose when it is pulled fully into or closer towards the aspiration device.

[0086] Reference is now made to Figs 2A-2C which show configurations of the fluid system FL including, as component, a conduit system CN such as a system of vessels. Such system of vessels or similar conduits may include a main branch MB and one or more side branches SB branching off of the said one or more main branch MB. Thus, the conduit system may comprise one or more bi-or multifurcated structures tubular with the fluid FLU circulatable therein, as a further component of the fluid system FLS, such as a liquid (eg blood) flowable through the conduit system CM. As a yet further component of the fluid system FS, this may include within the conduits system CN an obstruction EM, such as an emboli, clot or any other distinctly defined, foreign or innate, mass that may be lodged into a position inside the vessel system CN, thereby obstructing one or more branches, either the main branch or one or more of the side branches SB. Thus, obstruction EM hampers flow of fluid FLU. Obstruction EM is preferably not an innate or native part of the fluid system FLS. For example, in the medical field, such obstruction EM may be a pathology that should not be there.

[0087] Once such obstruction, eg clot or embolus, may be comprised of deposits, fats, etc that form into a, possibly sticky, conglomerated, discrete mass definition that tends to hinder blood flow through the affected vessel site, such as in main branch MB. Clot EM location and shape is purely and whole illustrative and is in particular in no way limiting the instant disclosure.

[0088] It will be understood that the obstruction EM may not necessarily be a complete obstruction such as to occlude essentially the affected branch portion. It may sufficient herein that simply flow is hindered by the obstructing body EM, causing a stricture thereby reducing or hindering the flow of the surrounding fluid. Whether the obstruction amounts to an occlusion or a blood flow impediment with residual flow, such obstruction may have adverse health effects on patient PAT as the obstruction may starve downstream tissue off oxygen.

[0089] Upon introduction and / or operation of the aspirator ORD, the suction power generated at its action tip TP causes a negative pressure any may in some cases drive flow or debris in the tube member TM. As shown in the configuration in Fig. 2 A, operation of such aspirator device ORD may unfortunately cause under certain condition a net adverse flow Vacomponent in the surrounding blood (or more generally, fluid) pool in which clot is situated that is directed away from the action point TP of the aspirator ORD. Thus, the net adverse flow vsscan potentially drive the embolus EM or its debris deeper into vessel, away from aspirator device ORD. Whilst some favorable flow components Vf towards aspirator may exist, there may however be a net adverse flow vaaway from the device, thus causing an unfavorable / adverse clot response, which is to be avoided. It is suspected that such adverse response with a net adverse flow component vamay be caused by fluid dynamic effects and turbulences that may form at the ostia of the one or more side branches SB, and fluid, such as blood, entering thereto into the branch in which the pressure is applied.

[0090] Fig. 2B shows a favorable situation where the placement and operation of the aspirator ORD is such that there is at least a net favorable flow components Vf, which may cause a favorable response in respect to clot EM or fluid's flow patten , in which upon dislodgement, the clot or any debris thereof is carried , by the surrounding favorable fluid flow patten Vf, towards the aspirator's tip TP, at which it can be retained and safely evacuated from the affected vessel system CN.

[0091] In an example of a favorable response situation as illustrated in Fig. 2B, this may provide a safe situation / environment in the fluid system FLS to use a stent retriever or other mechanical device MR for (further) dislodging, entrapping, and removal of the obstruction outside the conduit CN, as any debris that may be caused by such mechanical removal operation may be evacuated by the suction of the aspirator ORD. Thus, the aspirator ORFD may either act on its own to remove the obstruction itself in some cases where size and consistency of clot so allows, or the aspirator ORD acts to facilitate such removal by the mechanical removal device MR. For example, debris that may be so formed during mechanical interaction (grab, entrap) by mechanical device MR is absorbed by the aspirator ORD. Such a stent retriever type device MR is schematically shown in inset Fig. 2C, and the fluid system FLS's response may be conducive for such use of mechanical retriever MR, as per the flow pattern in Fig. 2B, or in other similar situations / responses. As mentioned earlier, grabber function is provided by the mesh MSH, deployable out of a micro-catheter MC, that may have a smaller lumen than the tube portion TP of the aspirator ORD. Microcatheter MC may run longitudinally and essentially coaxially inside the larger tube portion TP of the aspirator ORD. Thus, in some embodiments the mechanical remover MR and the aspirator ORD may form an integrated device, although this may not necessarily be so in all embodiments contemplated herein. A further or alternative suction may be applied through the microcatheter MC, so that suction is applied after deployment of mesh MSH, and whilst mesh engages with clot and is retracted, thus dragging clot EM with it, after microcatheter MC has been positioned through the clot EM initially. Thus, any debris that mas so form during grab-and-drag may be absorbed safely by aspirator action through micro catheter and / or aspirator proper tube TP, should such debris form, thanks to the favorable flow pattern Vf in Fig. 2B. Instead of using dedicated aspirator only device ORD with its own tube TP, this may not be needed, and suction may be solely applied through microcatheter of stent retriever MR, or other such mechanical device with own aspirator functionality. Thus, any reference herein to aspirator ORD or parts thereof includes is a reference to either dedicated aspirator / tube TB, or microcatheter MC, or any other part through which suction is applied in the fluid system FLS, in which ever form, with or without mechanical remover MR. The stent retriever arrangement in Fig. 2C or similar is merely one example embodiment, and may not necessarily be used in other embodiments. Any obstruction removal facilitator / device ORD, whether or not including mechanical removal assistance MR, is envisaged herein, so long as there is an aspirator component TP, MC present. Indeed, instead of, or in addition to, mechanical remover MR, removal may be based on other principles such as electrical, adhesives, etc, also envisaged herein. Removal of obstruction EM solely by aspirator suction TP, MC, without any mechanical MR or other assistance, is also contemplated, as mentioned above.

[0092] Exactly which one of the adverse (Fig. 2A) or favorable responses (Fig. 2B) respectively is causable in fluid system FLS by operation of aspirator ORD, may not be precisely understood. It is suspected thar the nature of the response (adverse or favorable) may be function of the spatial configuration of the vessel system around the obstruction EM, and, the manner in which the suction is applied or is to be applied by aspirator device ORD, in particular the pressure gradient caused, and possibly other operational parameters of device ORD.

[0093] The facilitator system FS is configured herein to take some, or all, of these factors into account and estimate / predict the said indication m on whether a favorable or adverse response of the fluid system FLS (such of the obstructing EM) can be expected.

[0094] It will be understood that the proposed system may also be used for removal of obstructions from other conduit or vascular structures in the human body, such as in the urinary system for example for removal of calculi from urinary tract, from kidneys, gallbladder, etc.

[0095] Reference is now made to Fig. 3 which shows a block diagram of the facilitator system FS, illustrating its operation in more detail. Projection raw data X' is receivable at input interface IN'. This may then be pre-processed as described in more detail below to obtain input image data X which may then be received at downstream input interface IN and may then be processed herein by facilitator FS into an indication m on an estimated response of the obstruction EM (eg, a clot etc), such as an adverse response. Thus, such indication may relate to an adverse or favorable response of fluid system FLS, such as distal flow outcome, or disintegration risk of the obstruction EM, etc. Distal flow outcome may lead to distal embolization. Operation in raw projection data raw data X' is also envisaged herein, with some or all of the pre-processing being optional.

[0096] Preferably, the computation of the indication m may be computed by facilitator FS prior to operating the aspirator device ORD. In addition, or alternatively, the system FS may be used a realtime guidance system during application of negative pressure by the aspiration device ORD to so provide a stream of output data. As said before, the output data may include a time series mt of such risk indications, which are updated over time at suitable sampling rate (eg, at frame rate (fps) of imager IA) during the imaging / intervention. This may be useful if the removal device ORD is being moved itself during it operation, or indeed when the clots sets into motion, as either event may change the fluid dynamics boundary condition, which in turn may influence, potentially unfavorably, the fluid system FLS's response. For example, even though the fluid system FLS response may have started out favorable, with favorable flow, such flow may then change into an adverse flow pattern, later in the aspiration procedure.

[0097] The pertinent input X-ray imagery such as the 2D or 3D angiogram, is recorded by the imaging system IA. The facilitator system FS may include the mentioned obstruction response estimator component EC (referred to herein for short as the "estimator (component)". The input imagery is received at input interface IN of facilitator FS, to be processed by estimator EC into the response estimate m, which may be output as an indication thereof at output interface OUT. The imaging system IA providing the imagery may be a C-arm CT scanner or any other, such as C-arm, Bi-planar system, etc. Imagery may be acquired long different projection direction to better estimate the geometry of obstruction location and the surrounding vessel configuration. A C-arm system may be used or the said bi-plane system including two sources and detectors at an angle, usually at 90°. Projection imagery may be sufficient herein in most cases, but input imagery topographically reconstructed from such projections are not excluded herein. Thus, in some embodiment, the input image is in image domain and represents sections through or volumes of the ROI, rather than projection imagery.

[0098] Optional spectral imaging may be used for better visualization of contrast agent, by computing a contrast only image (virtual contrast image, "VC") based on a material decomposition algorithm, such as described by Alvarez et al in "Energy dependent reconstruction in X-ray computerized tomography", published in "Computers in Biology and Medicine". vol 6, Issue 4, October 1976, pages 325-334. A dual energy setup, dual source, kVp switching, etc may be used as imager IA hardware for spectral imaging.

[0099] Contrast agent may be administered by an administration device (not shown) into the vessel structure CN. Thus, the fluid comprises a mix of blood and contrast agent, and it is such a mixed fluid that is flowing through and around the region of interest whilst the imager IA records contrasted input imagery at suitable times. Use of such contrast agent (such as Iodine) allows conferring image contrast to the vessel system CN, which, it being soft tissue, is natively radio-transparent and may thus be difficult, if not impossible, to ascertain in the imagery.

[0100] In some contrast agents based protocols, at least one image may be recorded prior to administration of the contrast agent, or at least prior to such contrast agent reaching the region of interest so as to obtain at least two exposures, contrasted and contrasted. These may be subtracted from each other so as to reveal a digital subtraction angiogram (DSA), which may be processed herein by facilitator FS as input imagery for better contrast in respect of the vessel tree. An optional segmentor SEG may be used to segment the input imagery X for the vessel tree in the ROI including the clot location. Preferably, the segmented ROI includes some or all surrounding side branches or main branch around the location of the clot EM. The segmented ROI preferably also includes the branch of the vessel structure where the tip TP of the aspirator ORD is located or will be located. Choosing the ROI in this manner allows for a better estimating and distinguishing favorable or adverse responses of the fluid system FLS upon application of the suction.

[0101] An in-image obstruction locator OL may be used locate the obstruction in the image. The obstruction locator OL may be able to operate automatically as an in-image recognition / detection component, possibly ML (machine learning) based, or based on analytical image processing techniques such as edge detection, segmentation, etc. The in-image obstruction locator OL may operate on the segmented imagery for better performance. The obstruction locator OL may provide an image-coordinate descriptor, a bounding box. or other image neighborhood to indicate the obstruction location. However, any one or both, the obstruction location and the aspirator tip location may be provided, such as specified by user using a user interface UI, as needed and required, specifically, depending on the level or automation needed.

[0102] In embodiments, the mentioned DSA input imagery may facilitate identifying the clot location. This can be done manually or automatically by image processing such as by obstruction locator OL. The image clot location may betray itself by producing branches of the vessel tree that are only partly contrasted due to the obstruction EM hindering blood flow. In case of occlusion, a sharp contrast gradient may emerge along the direction of a branch in the vessel tree, thus marking the location of the clot. In case the obstruction does not occlude, but does allow for some non-zero flow therethrough, the contrast drop across the obstruction location will be less pronounced, but may still remain clearly detectable for user or automatically by obstruction locator OL. If detection of clot location is by user, the obstruction locator OL may be of the user-interactive type rather than automatic. In this embodiment, the obstruction locator OL may support an interactive user interface UI. The UI may allow user to pin-point the location by touchscreen action, mouse or stylus click definition, keyboard stokes, or any other, as required.

[0103] The estimator component EC (referred to herein for short as the "estimator") may operate on the, optionally segmented s. input imagery X, to estimate response of obstruction EM. The estimator EC may further use the detected location of the obstruction as per the input image and as provided by the obstruction locator OL to provide the estimate. In addition, the location of the device ORD may be used by the estimator EC. Thus, at least 3 input data components may be used in embodiments as per the input imagery: i) the vessel segmentation s, ii) the location of the obstruction in the vessel, and iii) the location of the removal device ORD, in particular of its tip portion. In some embodiments, size (weight) or other properties of the obstruction, its composition etc, may be used as additional input to inform the estimate m. As said, as an option, spectral imaging may be used to obtain such information concerning the obstruction EM. In addition to processing by the estimator EC of input imagery, contextual data may be optionally so coprocessed to compute the clot response. Such optional contextual data may be include hemodynamic properties like internal pressure or flow measurements e.g. obtainable by analyzing contrast agent dynamics in the image data, or obtainable by local measurements in a catheter for example.

[0104] Operation of the estimator component EC may be based on a model M to estimate response of fluid system FLS. Such a model M may include a fluid dynamics model, such as a lumped resistor-and-flow model, as illustrated in Figs 4A-4B. Such model M may use vessel sizes, cross-sections, its geometry, etc, as ascertainable from the segmented imagery, fluid dynamic laws as per differential equations, applicable boundary conditions, etc to estimate expected blood flow pattern and hence the response of the obstruction.

[0105] Operation of the estimator component EC may include solving such differential equations that govern the lumped model. Any numerical methods may be used to solve such differential equations. Difference equations approximating such differential equations may be used. The lumped resistor-and- flow model M may allow estimating the fluid dynamics vector fields around the obstruction EM. In particular, it is the flow vector field in the vicinity including clot location and the device tip that may of interest.

[0106] In Figs 4A-4B such a fluid dynamics model M is illustrated for the two configurations corresponding to Figs 2A-2B, respectively. Such a model M may include various lumped flow related components CM that represent features or aspects of the fluid dynamics. The features or aspects are localized at discrete locations (hence "lumped") within the affected vessel system CN, in particular within the ROI, yet more particularly in said action vicinity. Such localization / lumping is an idealized model assumption, but have been found to yield good results, in sufficient approximation, with relatively low computational and memory overhead.

[0107] Some such components CM may include positive pressure components VP, resister elements RE, and respective connections CN therein between. A zero-level pressure modelling component P0 may be provided that represents an average ambient pressure level. The positive pressure component models effect of the heart, and the connections CN the geometry of the vessel tree, its branches and their connective patten. The resistor elements RE model expected flow resistance as may be estimated from the size, width, geometry of the vessel branches, and optionally from patient biocharacteristics, such as age. Effects of deposits in the vessel walls may be estimated given patient's biocharacteristics such as age, BMI, sex etc. Blood samples may be used to establish blood viscosity to refine the fluid dynamics model. One such resistor element RE' (shown at bottom of Figs 4A-4B) represent the resistance effect of the clot EM over and above the native resistance in the vessel branch. The resistance strength of clot resistor element RE' may be estimated from the DSA for example, with contrast downstream clot location being an indication for such resistance strength. Preferably, the model may include a dedicated specific pressure component DE(ORD) that represents the negative pressure exerted by the aspirator device ORD during operation. This dedicated component DE(ORD) of the lumped model M representing negative pressure (suction, partial vacuum caused by the aspirator device ORD) may be connected / coupled with the remaining components through suitable flow connectors CN representing the vessel configuration at hand, in particular the branching pattern in the ROI of main and side branches, if any. Thus, the model may include at least one dedicated pressure source of the aspirator and at least one dedicated resistance element for obstruction (clot, etc) EM.

[0108] The model M may be dynamically rewired and recomputed depending on the position of the aspirator device ORDM, should this change, as per Figs 4A-4B where the aspirator ORD is moved past a side branch ostium. Also, the clot EM removal may be modeled as a dynamic change by changing its corresponding resistance element. It can be expected that the flow patterns in the conduits change as the resistance element of the clot is changed. Thus, there may be the case that before the clot removal a favorable flow patter is estimated, but as soon as the obstruction is moved the risk of an adverse pattern is increased.

[0109] In addition to the number of side branches SB, a specific type and label may be identified in order to refine the fluid dynamic model M, and more accurately estimate the risk of collateral / side- branch flow occurrence. For example, depending on the larger anatomical context it is more likely to get collateral inflow from a larger branch connected closely to the circle of Willis or the anterior communicating artery.

[0110] Once the model is set up (on which more below), it can be run in simulation mode to compute the flow components f, vaalong the respective connections, in particular at the location of obstruction EM, but not only there. The flow components Vf vacan be matched, for example they may be labelled with the respective branches in the vessel structure to which they pertain.

[0111] Some or all parts of the fluid system model M=M(LFS), in particular the routing of its connections CN and the placement of the pressure PP, PO, DE(ORD) and resistor components RE, may be overlaid as a graphics component onto the input imagery (eg, on a frame thereof) that represents the view of the vessel structure CN. The simulation of the model yields flow vectors vp, Vd as solution to the underlying fluid dynamics differential equation(s). This graphics display including the model circuitry and the input imagery overlaid may be visually displayed for a quality check by the human user.

[0112] Alternatively, once the flow vectors Vf and / or vaare established, the system FS can check whether there are only distal flow components Vaor whether there is net distal flow component Vd that effects fluid system FLS such as the clot EM. Thus, it is established whether there is at least a net distal flow Vd component in connector CN that represents the (one or more) branch MB in which the clot is lodged. If there is, an alert signal m or message may be issued through system FS's output which may be used to drive a transducer, a display device, etc, as needed. The alert signal may be qualified, in that this is only issued if the vais computed to exceed a pre-set tolerable threshold. If, on the other hand, there are only proximal flow(s) Vf , or proximal net flow Vpin the affected branch MB / connection CN, then an all clear signal may be issued, or the system remains dormant in this case, that is, no notification is issued. A similar alert may be issued when there is risk of clot forming with adverse flow va, vessel damage, etc.

[0113] In addition, or instead to producing output in terms of the above per branch / connector flow components , an overall probability map, probability distribution, or simple single probability value 6 [0,1] may be computed indicating that an adverse event may occur. In some embodiments, a probability of adverse flow pattern va(distal embolization) may be determined by response estimator EC as follows using the fluid dynamics (FD) model, of the resistor flow type, or other. For example, multiple FD model simulations are done, including at least one with the clot present in the vessel CN, and at least one with the clot virtually remove from the model. For example, the dedicated clot resistance element RE' may partially removed (eg by reducing the associated resistance value), and may be completely removed. Thus, at least two, better three or more flow simulations are run, resulting in two or three or more respective flow vector fields: one with clot element RE' present, one with clot element RE's removed (value = native vessel resistance), and one or more simulations with clot element RE' having respective reduced resistance values (=pressure / flow, with units chosen accordingly). An observed increase of distal flow over such at least two simulations with varying values of RE's may be used to define a probability for distal flow / distal embolization. The above mentioned probability values may arise as uncertainty out of the model M assumptions, or may be based on a training data set comparing estimations to actual outcome such as in machine learning based embodiments, also envisaged herein a described in more detail below.

[0114] Setting up of the model may be done automatically using the branch pattern as per the segmented vessel tree. Optionally, physiological parameters, such as ambient blood pressure (estimated or measured), blood viscosity (measure or estimated) can be used to define the model and its components CM. The physiological parameters can be measured (vitals readings), or looked up in physiological reference works, or may be indirectly estimated from patent bio-characteristics such as age, sex, BMI, patient health record, etc. Thus, the model M may be translated into from segmented vessel tree and such optional parameters by estimator component EC. Thus, the EC may be operable in two phases: in model translation phase and in a subsequent model deployment phase where the model is run to simulate prevailing flow pattern in form of flow vector field. The flow vector field may be mapped onto the input imagery for a suitable (eg, the latest) frame, and can be displayed concurrently with the imagery on display device DD.

[0115] The deployment phase may include solving underlying systems of FD differential equations such as Euler's equation, Darcy-Weisbach equation, Navier-Stokes equation, or others, to obtain the spatially resolved flow vector fields Vj(x ), va(x ), with x an in-image position vector.

[0116] Alternatively, the system FS may support an interactive user interface UI, configured to allow the translation phase to be at least partly user driven. The UI may allow user to set elements (widgets, etc) for the model components CM (such as in Figs 4A-4B above), with flow determinative properties which the user can manipulate. Thus, the UI may allow user to use touch screen or pointer tool interactions to set in a display field, such widgets w(CM) for the elements CM. In this way, fluid dynamic boundary conditions may be set, with connectors w(CN) therein between. The user may thus replicate graphically the topological pattern of the vessel tree. The field may be formed by the segmented image, with the user selecting elements CN as overlays thereover. The user can move and combine such widgets w(CM) representative of such CMs component. Once so set, the governing diff equations can be set up and solved, using any suitable numerical method to run the fluid dynamics simulation and obtain the flow vector field including any one or more of Vp(x ),Vd((x ) relevant for the location of the obstruction EM.

[0117] In embodiments, the lumped resistor and flow model or other FD model may use vessel sizes and a scaling law as boundary conditions to estimate expected blood flow vector fields Vf,va.. There are some general anatomical laws that relate size of a vessel to the blood flow for which it is intended. See for example G S Kassab " Scaling laws of vascular trees: of form and function" in Am J Physiol Heart Circ Physiol, vol 290, pp H894 -H903, ((2006). Such law may be apply or the non-pathological case, but may nevertheless serve to set boundary conditions for better estimation of flow conditions in the case of pathologies, external perturbation, and blockages, etc.

[0118] Instead of flow-resistor lumped models, a spatially more detailed modelling may be done, for example based on finite element methods ("FEM") computing, although computational and memory demands may likely be higher than for the lumped approach. Also, whilst in the above the output is described with main reference to flow fields Vf,a, instead a corresponding field of (net) pressure gradients Pf,a„ acting distal or proximal at the obstruction EM may be computed instead by estimator component EC and used instead in place of flow vp,d. Such pressure distribution may be useful to establish fluid system FLS response ins term of vessel damage.

[0119] It should be understood herein, that the modelling of flow dynamic components explicitly, classically and analytically, as described above is not necessarily required herein in all embodiments. Instead, a machine learning (ML) model M, such as of the artificial neural network type, may be used. This may be trained on suitable training image data previously annotated by a human medical expert for example. The annotation may indicate whether or not adverse fluid system FLS effects / responses occurred at that time. Such training image data may be drawn from prior aspirator interventions, and may be sourced from various medical databases such as PACS and others. The training image data may include historical fluoroscopy / angiography video footage as were recorded at the time. For the preparation of the training image data, such footage can be replayed and studied by the human expert, The expert, such as a clinician, may then observe the response of the obstruction to the aspirator action, debris formed, the nature of flow / pressure pattern, any vessel damage occurring, etc., in order to identify / annotate whether or not adverse reactions have occurred as each of the specific historical interventions. Parameters of the machine learning model such as weights in a neural network model are adjusted in a training procedure. For example, based on cost function or other objective function, parameters are adjusted so as to improve output of the objective function. This allows adjusting the model parameters so that the training imagery applied to the model matches the targets, that is, the annotations provided by the annotating physician. Preferably, a large enough pool of such training imagery ought be used in the training, representing in good balance and variations adverse and favorable responses of the obstruction, so that the model is trained to generalize well in later deployment.

[0120] After training the model may be used for deployment, where different image data as encountered in new clinical practice is used, different from the image data used in training. Optionally, there is a testing phase where the performance of the machine learning model is tested on data also different from the training data set. The machine learning model, such as of the neural network type, may include various layers including input layer, output layer and in between one or more plural hidden layers, such as may be used in deep learning architectures. The size of the input layer should correspond to the size of image so is essentially 2 or 3D depending on the image dimension. A time series, that is 4D or 3D imagery, with one dimension being the time component respectively, may be fed at once as input into the machine learning model for better robustness although frame wise processing is not excluded herein.

[0121] Preferably, a convolutional neural network is used as such models have been observed to perform well with spatial data such as is the image data provided as input. One or more convolutional operators apply convolutional filters in the various layers to obtain intermediate data in between layers. The intermediate data is propagated through network from the input layer across the one or more hidden layers. The output layer may combine the intermediate output of the various convolutional operators to produce the desired output. The output may be a regression or a classification result, as required. For example, the model may be configured to merely classify, based on the respective received image, whether adverse or favorable response of the obstruction EM is to be expected. The respective probabilities may be output as a result instead, or in addition. Instead of neural networks, any other model may be used such as support vector machines, other classifiers, decision trees / forests, etc.

[0122] A similar such set up may be used to assess the risk for clot EM disintegration which is again an event that can be ascertained from historical image data as may be found in medical image records, and can be annotated by expert accordingly. Generative type ML models may be used, that may generate estimated, synthesized fluoroscopic / angiography footage that may serve to illustrate by a short movie as it were, the likely response of the fluid system FLS. Such a generative type ML models may include GANs or GAN-type networks / models, variational autoencoders (VAE), transformer type networks, models with or without attention mechanism, etc. GANs have been described by Ian Goodfellow et al in "Generative Adversarial Networks", published on arXiv: 1406.2661, 2014, available only at https: / / arxiv.org / abs / 1406.2661. In such and similar approach, two models, a generator and a discriminator network as set to compete against each other to so generate data. A corresponding cost function is used.

[0123] Reference is now made to Fig. 5 which shows a flow chart of a computer implemented method which may be used to implement, at least parts of, the above-mentioned facilitator system FS. It will also be understood herein that the proposed method may be also understood as a teaching in its own right and it is not necessarily tied to the facilitator system architecture described above.

[0124] At step S505, raw data X', such as projection data is acquired by imaging apparatus of a ROI. The imaging apparatus with which such data is acquired can be configured for angiographic or fluoroscopic imaging. Contrast agent may be used and released into the vessel / conduit system. Thus, imagery of contrasted and, optionally, of non-contrasted type of the region of interest may be acquired. X- ray based imaging is preferred but other modalities such as US, MRI, phase contrast, dark-field imaging etc, is not excluded herein. Imagery so acquired my record the ROI in suitable area size, in particular may include the clot location, and some surrounding vessel structure wherein the clot or other obstruction is lodged. Preferably, such imagery may also record the location where the obstruction removal device, as aspirator device, in particular its action tip, is placed or will be placed. The action tip is the part of the aspirator device through which suction is applied. In some designs, such action tip or another implement is used to secure the obstruction once dislodged, so it can be evacuated outside the vessel or other conduit system, as the application case may be.

[0125] Such raw data of step S505 is received for optional pre-processing phase.

[0126] For example, at step S510 an optional spectral processing may be applied to the raw image data of step S505 in order to enhance image contrast of the vessel tree. In addition or instead, DSA may be used to yet further increase contrast. A contrast only image / virtual contrast ("VC") may be so computed using any suitable material decomposition algorithm such as Alvarez et al (supra) or other.

[0127] At step S520, pre-processing may be applied to the imagery to identify the obstruction location. However, this may be done instead by user input. In addition, the location of the aspirator device (in particular its tip or component through which suction is to be applied (or is applied) is identified, either automatically of by user input. DSA of contrasted / versus non-contrasted imagery may be used to identify obstruction location.

[0128] At step S530 imagery is optionally segmented for the conduit such as for the vessel structure with the clot location and the aspirator tip placed. Any type of segmentation algorithm may be used, ML based or classical such as edge detector based or other.

[0129] The pre-processing for the obstruction and / or device location may be done after segmentation in the segmented imagery.

[0130] The steps S510-S530 may be considered an optional pre-processing phase to produce input imagery X to be processed below.

[0131] Thus, at step S540 the imagery, optionally pre-processed as above in one or more steps, is received as input imagery X for processing to estimate fluid system FLS response, such as obstruction response, for example in order to distinguish adverse from favorable response of the fluid system FLS, such as favorable / adverse movement of obstruction or debris on mechanical or aspirator only action, and / or clot's disintegration, vessel damage, etc. The imagery may be a sole, still image, at a given projection direction, or may be acquired overtime as video footage from one or multiple projection directions a, as needed.

[0132] The input imagery X considered herein is not necessarily in the projection domain as envisaged herein in most embodiments. An imaging apparatus for tomographic imaging such as a CT apparatus or an imager of the C-arm may be used, in combination with tomographic reconstruction algorithm to compute 3D sectional / volumetric images as input. A single such volume may be used, or a time series (4D) of such voluminal may be used and provided as input image data X. Preferably however, because of the lower computational overhead, and hence better responsiveness, projection domain processing is preferred and sufficient herein.

[0133] Additional input data, in addition to the input imagery, may include one or more operational parameters of aspirator device ORD, such as placement of its tip relative to clot EM location and amount of negative pressure (suction) to be applied or a currently applied. The placement may be as per the imagery (see step S520). The pressure setting may be indicated by user or may be acquired by suitable machine interfacing with a controller interface of device ORO.

[0134] At step S550 an estimate m is computed based on input data including input imagery X and the operational parameters P of aspirator device ORD. Prior segmentation is preferred herein, as per step S530. Thus, at step S550 an estimate m is computed for the likely response of the fluid system FLS (eg response of clot, vessel or flow / pressure), when the aspiration device is operated. Thus, the estimate m is preferably supplied before operation of the aspirator device ORD, but after spatial placement of the aspirational device ready for operation. Alternatively, if the vessel location of where the aspirator's tip will be placed is known beforehand, this location may be provided as per step S520, and the estimate is obtained based on the prospective location. Alternatively, in a dynamic in-use application rather than in a pre-use application, a series of such estimates may be computed during operation of aspirator device ORD.

[0135] The estimations step S550 may be done with a machine learning model or based on other modelling such as FEM, or lumped resistor-flow models, or any other fluid dynamics models, as described above. The model use may include in particular an explicit, dedicated, representation of, / model component for, the aspirator device ORD parameters, in particular of the negative pressure to be applied by the device ORD during its intended or ongoing operation. Thus, in a fluid dynamics model, such as in the lumped model (Figs 4A-4B ) or similar, there is dedicated model component indicative of the said negative pressure and, optionally, of its spatially placement relative to the obstruction location. Preferably there is also such explicit model component of the obstruction (eg, no flow (occlusion), or some through- flow, etc) is terms of a resistor element placed at a topology consistent with the vessel tree, and this, consistent with the manner of fluid communication with the aspirator model component, as per the input image. If ML model is used, the negative pressure data may be supplied as contextual data alongside the input imagery, such as in a separate channel as an additional dimensional component. For example, if the input imagery is 2D, another 2D layer may be included that represents the negative pressure value. The additional 2D layer may be populated with the intended pressure value. Thus, this may result in a 3D data dimension of the input data.

[0136] Thus, it will be understood herein that the input data processed at step S550 into the estimate m is based on two data components: i) the input imagery that represents that spatial topological layout of the conduit and vessel system with location of obstruction and location of device (eg, operational / action tip thereof) ii) operational parameter(s) so the device including the negative pressure specification (a vector or number such as mbar, etc) to be used. A representation of the topological layout in projection view is sufficient herein in most cases.

[0137] At step S560, if the estimate of fluid system FLS response is unfavorable, various new or alternative aspirator device ORD parameters may be recommended. This can be done by sampling various positions in the vessel tree of the aspirator device ORD placement and / or sampling various settings of the aspirator device ORD, such as the to be applied negative pressure in order to obtain different such estimates m'.

[0138] If a favorable fluid system FLS estimate is obtained, the associated aspirator device ORD parameters are then captured and brought to user's attention such as by displaying same for example. The user can then apply the new aspirator device ORD settings to the aspirator device ORD to perform a safe removal of the clot. In addition or instead, the recommendation at step S560 is applied through a control interface to the aspirator device, to re-set the settings automatically and apply the same.

[0139] Thus, the proposed system can be used in a fully automatic fashion, such as in a robotic surgery system, without or with little user interaction. However, keeping the human user in a loop is preferred herein so all the data mentioned above is displayed and it is the user who then initiates appropriate action based on the display information the respective actions.

[0140] The method may further include providing an indication of the estimate and / or of the recommendation (if any) as output at step S570. Such providing may include displaying as information on a display device, or by using any other transducer to produce a sensory (audio, visual, haptic, etc) notification signal to the user, such as an alert if response is likely to be adverse. The output of estimate indication may be binary (adverse / favorable) or as a probability value, a probability distribution, or by providing an explicit vector field representation of flow vaor pressure paat location of obstruction, as needed.

[0141] In addition or instead to the above clot response, a risk for vessel damage from too much aspiration could be estimated as an optional additional output computable at step S550. If the model reveals that pressure inside the vessel CN drops below a predefine threshold the risk of vessel collapse increases. Also, parts of the vessel may become wedged into the aspiration device. E.g. at a bifurcation, or similar. Again, instead of using an FD model, ML may be used instead, based on training data that include cases that are known to have led to such damage. The later may be researched by database queries in medical image or medical record databases or repositories, etc. In embodiments, clot properties like size and composition can be used to estimated / predicted, as additional output at step S550 the likelihood for clot disintegration, such as the occurrence of debris and the speed of clot intake. This embodiment may include adding additional model parameters (eg, on composition of clot) in the probability prediction of distal embolization. Alternatively, as mentioned above, machine learning may be used to estimate response, such as disintegration.

[0142] In the above embodiments, not all operational parameter need to be optimized / considered and controlled / adjusted. This method could be limited to only device location and / or aspiration strength alone.

[0143] In some cases, aspiration is performed manually with a syringe, thus there may be an additional aspiration guidance on the suitable volume of the syringe, and the duration over which the partial vacuum is applied .

[0144] In case that the input image X is a diagnostic image was used to set up the fluid dynamics model in order to recommend the new operational parameters, registration may be used to transfer this information into an interventional imaging procedure and to guide the interventional procedure, e.g. by overlaying the output information directly on a live or diagnostic image.

[0145] Depending on which one or more aspect of fluid system FLS response is of interest, the indication output may be a scalar, doublet or triple, such as m = (ml, m2, m3), such as for flow / pressure distribution indicated as favorable or adverse, clot response, or vessel response, for example, the estimated flow may cause movement of clot or may be enough to carry of at least debris that may be released due to mechanical extraction of clot EM. As to vessel damage, the pressure distribution may be compared against clinically known threshold values that are known to cause collapse. Pertinent material property data may be retrieved from known reference works via interfacing with a medical knowledge database, optionally, using patient's health records and / or their bio-characteristics as such material property vessel values may change with age, or other factors. Instead, anyone, more than one, or all such indications may be estimated using machine learning, based on patterns learned from training data. Such ML estimation may be implicit such as by regression or classification, or by ostension (examples), such as by generating X-ray footage representative of how fluid system FLS response events may unfold. Such generative models can be trained from historical X-ray footage from prior exams from a range of patients. Such historical X-ray footage may be retrievable from medical databases.

[0146] If machine learning is used, such a machine learning model, a neural network model, such as of the convolutional type may be used. Training data may be obtained at step 610, either annotated historical data retrieved from medical database, as mentioned above. Alternatively, the step of obtaining such training data may include synthesizing such as using generative networks (GAN, etc), and / or augmenting to so obtain data at larger variation. At step S620 a model may be trained, such as by adapting parameters of the model to improve an objective function. Gradient based techniques, such as gradient descent, or stochastic gradients may be used. Backpropagation may be used or similar to spread any training errors among current parameters and adjust same accordingly. The training may be iterative. Instead of using a supervised setup, generative type models, such as GANS, variational autoencoders (VAE) may be used, which may be trained on existing interventional X-ray footage as mentioned above.

[0147] The components of system FS may be implemented as one or more software modules, run on one or more general-purpose processing units PU such as a workstation associated with the imager IA, or on a server computer associated with a group of imagers. The workstation or other data processing unit(s) PU may be part of the imaging apparatus IA.

[0148] Alternatively, some or all components of system FS may be arranged in hardware such as a suitably programmed microcontroller or microprocessor, such an FPGA (field-programmable-gate- array) or as a hardwired IC chip, an application specific integrated circuitry (ASIC), integrated into the imaging system FS. In a further embodiment still, system FS may be implemented in both, partly in software and partly in hardware.

[0149] The different components of system FS may be implemented on a single data processing unit PU. Alternatively, some or more components are implemented on different processing units PU, possibly remotely arranged in a distributed architecture and connectable in a suitable communication network such as in a cloud setting or client-server setup, etc.

[0150] One or more features described herein can be configured or implemented as or with circuitry encoded within a computer-readable medium, and / or combinations thereof. Circuitry may include discrete and / or integrated circuitry, a system-on-a-chip (SOC), and combinations thereof, a machine, a computer system, a processor and memory, a computer program.

[0151] In another exemplary embodiment of the present invention, a computer program or a computer program element is provided that is characterized by being adapted to execute the method steps of the method according to one of the preceding embodiments, on an appropriate system.

[0152] The computer program element might therefore be stored on a computer unit, which might also be part of an embodiment of the present invention. This computing unit may be adapted to perform or induce a performing of the steps of the method described above. Moreover, it may be adapted to operate the components of the above-described apparatus. The computing unit can be adapted to operate automatically and / or to execute the orders of a user. A computer program may be loaded into a working memory of a data processor. The data processor may thus be equipped to carry out the method of the invention.

[0153] This exemplary embodiment of the invention covers both, a computer program that right from the beginning uses the invention and a computer program that by means of an up-date turns an existing program into a program that uses the invention.

[0154] Further on, the computer program element might be able to provide all necessary steps to fulfill the procedure of an exemplary embodiment of the method as described above.

[0155] According to a further exemplary embodiment of the present invention, a computer readable medium, such as a CD-ROM, is presented wherein the computer readable medium has a computer program element stored on it which computer program element is described by the preceding section.

[0156] A computer program may be stored and / or distributed on a suitable medium (in particular, but not necessarily, a non-transitory medium), such as an optical storage medium or a solid- state medium supplied together with or as part of other hardware, but may also be distributed in other forms, such as via the internet or other wired or wireless telecommunication systems.

[0157] However, the computer program may also be presented over a network like the World Wide Web and can be downloaded into the working memory of a data processor from such a network. According to a further exemplary embodiment of the present invention, a medium for making a computer program element available for downloading is provided, which computer program element is arranged to perform a method according to one of the previously described embodiments of the invention.

[0158] It has to be noted that embodiments of the invention are described with reference to different subject matters. In particular, some embodiments are described with reference to method type claims whereas other embodiments are described with reference to the device type claims. However, a person skilled in the art will gather from the above and the following description that, unless otherwise notified, in addition to any combination of features belonging to one type of subject matter also any combination between features relating to different subject matters is considered to be disclosed with this application. However, all features can be combined providing synergetic effects that are more than the simple summation of the features.

[0159] While the invention has been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. The invention is not limited to the disclosed embodiments. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing a claimed invention, from a study of the drawings, the disclosure, and the dependent claims.

[0160] In the claims, the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. A single processor or other unit may fulfill the functions of several items re-cited in the claims. The mere fact that certain measures are re-cited in mutually different dependent claims does not indicate that a combination of these measures cannot be used to advantage. Any reference signs in the claims should not be construed as limiting the scope. Such reference signs may be comprised of numbers, of letters or any of alphanumeric combination.

Claims

CLAIMS:

1. An image-processing system (FS) for facilitating removal of an obstruction (EM) in a system of one or more conduits (CN) through which fluid (FLU) is passable, comprising: an input interface (IN) capable for receiving input imagery obtainable by an imaging apparatus (IA) of the system of one or more conduits (CN) including the obstruction, the system of one or more conduits (CN) with the obstruction (EM) and the fluid (FU) forming together components of a fluid system (FLS); and an estimator component (EC) capable of estimating, based on the input imagery, an estimate (m) of a response of the fluid system (FLS) to a suction action causable by operation of an inconduit obstruction removal device (ORD).

2. The system of claim 1, wherein the system of one or more conduits includes one or more branches branching off in a region of interest, ROI, including the obstruction (EM).

3. The system of any one of the previous claims, wherein the estimator component (EC) is capable of so computing the estimate based on a computational fluid dynamics model, the model so based on the segmentation and, optionally, on one or more characteristics of the in-conduit obstruction removal device (ORD).

4. The system of claim 3, wherein the computational fluid dynamics model is based on any one or more of: a finite element model, a lumped resistor and flow model.

5. The system of any one of the previous claims 3-4, wherein the computational fluid dynamics model includes a model component representative of the suction exertable by the suction device (ORD).

6. The system of any one of the previous claims, wherein the estimator (EC) is based on a trained machined learning model (M).

7. The system of any one of the previous claims, wherein the said response is any one or more of: i) disintegration, deformation or motion of an object forming the obstruction, ii) flow of fluid away from the obstruction removal device (ORD), iii) flow of fluid towards the obstruction removal device (ORD), iv) a deformation of at least a part of the system of the one or more conduits.

8. The system of any one of the previous claims, including a graphics display generator(VIZ) capable of generating a graphics display (GD) for displaying on a display device (DD), the graphics display including an indication of the estimate so computable.

9. The system of any one of the previous claims, including a recommender (RC) capable of outputting, based on the estimate, or on further such estimate, or on the input imagery, a recommendation in relation to the obstruction removal device (ORD).

10. The system of any one of the previous claims, wherein the system of the one or more conduits includes at least one of: i) a blood vessel, ii) urinary system, iii) digestive system, iv) endocrine system, and wherein, in case of i), the said obstruction (EM) includes a thrombus or embolus.

11. An arrangement comprising at least a part of the system of any one of the previous claims, and further comprising any one or more of: i) the imaging apparatus, ii) the obstruction removal device (ORD), iii) a storage element including the model of any one of claims 3-6.

12. A computer-implementable method of facilitating image-based obstruction removal in a system of one or more conduits (CN) through which fluid is passable, the method comprising: receiving (S540) input imagery obtainable by an imaging apparatus (IA) of the system of one or more conduits (CN) including the obstruction, the system of one or more conduits (CN) with the obstruction (EM) and the fluid (FU) forming together components of a fluid system (FLS); and estimating (S550), based on the input imagery, an estimate (m) of a response of the fluid system (FLS) to a suction action causable by operation of an in-conduit obstruction removal device (ORD).

13. A computer-implementable method of training or setting up the machine model of any one of claims 3-6.

14. A computer program element, which, when being executed by at least one processing unit, is adapted to cause the at least one processing unit to perform the method as per any one of claims 12 and 13.

15. At least one computer readable medium having stored thereon any one or more of: i) the program element of claim 14, ii) the model as per any one of claims 3-6.

Images