Endoscope system

Abstract

An endoscopic system is presented, comprising: controller circuitry configured to issue a control signal to coordinate timing or other parameters of a fluid aspiration or irrigation device with timing or other parameters of an ablation member, where the control signal is automatically issued using a trained learning model, the trained learning model coordinates timing or other parameters of the fluid aspiration or irrigation device with timing or other parameters of the ablation member based on at least one of real-time data or historical data from one or more sensors communicatively coupled to the controller circuitry.

Description

[0001] This application is a divisional application of patent application No. 202211302274.7, filed on October 24, 2022, entitled "Automatic Suction Switch Based on Laser Switch".

[0002] Cross-references to related applications

[0003] This application claims the benefit of priority to U.S. Provisional Patent Application Serial No. 63 / 262,924, filed on October 22, 2021, the contents of which are incorporated herein by reference in their entirety. Technical Field

[0004] This invention relates to an automatic suction on / off mechanism based on laser switching. Background Technology

[0005] Medical endoscopes, such as endoscopes, were first developed in the early 19th century and have been used to examine the inside of a patient's body. Such endoscopes can include endoscopes, laparoscopes, ureteroscopes, duodenoscopes, arthroscopes, etc., which may be referred to as endoscopes or endoscopes. Endoscopes can be used to move fluids through organs. Some endoscopes provide flushing, aspiration, or both, for example, to move fluids. Laser energy can also be used inside the body. For example, a laser generator can deliver laser energy into the body, for example, through a laser fiber or bundle of fibers. Laser energy can be used for diagnosis or treatment, such as performing laser lithotripsy to ablate one or more stones. Ablation can break the stone into smaller fragments that can pass through naturally or be grasped, aspirated, or otherwise removed from the body using an endoscope or other assistive instruments. Summary of the Invention

[0006] One aspect of this disclosure provides an endoscope system comprising: a controller circuit system configured to issue control signals to coordinate the timing or other parameters of a fluid aspiration or flushing device with the timing or other parameters of an ablation component, wherein the control signals are automatically issued using a trained learning model, the trained learning model being at least partially based on at least one of real-time or historical data from one or more sensors communicatively coupled to the controller circuit system, to coordinate the timing or other parameters of the fluid aspiration or flushing device with the timing or other parameters of the ablation component. Attached Figure Description

[0007] The accompanying drawings are not necessarily drawn to scale. In the drawings, similar numbers may describe similar parts in different views. Similar numbers with different letter suffixes may represent different instances of similar parts. The drawings are illustrated by way of example, not limitation, of the various embodiments discussed in this document.

[0008] Figure 1 An example of the components of an endoscope system with a suction unit for controlling suction and a laser generator is shown.

[0009] Figure 2 An example of a method for controlling suction and / or flushing based on laser energy is shown.

[0010] Figures 3 to 6 An example of a method for controlling suction and / or flushing based on imaging of the working area is shown.

[0011] Figure 7 An example of a method for determining a treatment plan based on stone imaging is shown.

[0012] Figure 8 An exemplary schematic diagram of a computer-based clinical decision support system (CDSS) is shown. Detailed Implementation

[0013] The systems and techniques described herein relate to medical devices such as endoscopes that use suction or “aspiration” during lithotripsy to move fluids, stones, stone powder, and / or stone fragments through organs such as the kidneys. Endoscopic aspiration can be actuated by laser energy delivery, synchronized with, or otherwise coordinated with laser energy delivery, but aspiration can also be coordinated with one or more additional or different functions. This can be automated—requiring no user input to control aspiration.

[0014] Endoscopes can be used to visualize internal targets for diagnosis, treatment, or both. Endoscopes can be configured to facilitate fluid inflow (e.g., “flushing”) and removal (e.g., “aspiration”), such as washing away debris from an organ and improving visualization during endoscopic procedures. For example, a ureteroscope may include a flushing fluid channel supplying saline to the kidney during laser lithotripsy and an aspiration channel for removing stone powder and fragments. While patients can naturally expel stone powder and small fragments, active removal of debris during the procedure can be beneficial.

[0015] The inflow and outflow of fluid through the kidneys may involve increases and decreases in kidney pressure, which may benefit from monitoring, control, or both. Keeping aspiration continuously “on” throughout the procedure may be undesirable. Always keeping aspiration “on” may involve continuous monitoring and control, or a continuous inflow of flushing fluid to replace the fluid removed from the aspiration. Furthermore, when pulverization or other stone breakage is paused, aspiration and / or fluid passage through the kidneys may not be necessary during that time. Therefore, it is advantageous to control aspiration so that it is only activated when needed.

[0016] The systems and techniques described herein can be used to enable endoscopic aspiration at times deemed important or desirable, and to suspend aspiration at other times. While aspiration can be manually controlled by the endoscopy user (e.g., using a foot pedal or other actuator), these systems and techniques can provide the ability to automatically control aspiration, for example, based on one or more criteria, without requiring separate direct user input to enable and / or deactivate aspiration. For example, aspiration can be automatically controlled based on the state of the laser or other energy source (or associated control signals), such as enabling aspiration in response to the laser source being turned on or off during laser lithotripsy, and deactivating aspiration in response to the laser source being turned off or off.

[0017] A time-shift delay can be provided for suction control relative to the laser generator's activation / deactivation. For example, while suction can be automatically deactivated when the laser is turned off, a programmable deactivation delay can be used to continue suction for a period of time after the laser is turned off or before it is turned on. For instance, suction can be kept "on" for a programmable or otherwise specified amount of time (e.g., ½ second, 1 second, X seconds, etc.) before the laser has started delivering laser energy or after the laser has stopped delivering laser energy, for example, to move stones or fragments to a desired location for ablation via energy delivery, or to provide some time after laser activity to remove powder / fragmentation.

[0018] In addition to time, one or more criteria can be used as the basis for automatically delaying when to pause suction, such as when the image sensor has produced one or more images that are considered sufficient to remove powder, fragments, and / or other debris. For example, image sharpness, haze, or one or more other individual or composite features can be compared to a specified acceptable amount or a specified acceptable image. This can be achieved through algorithms or by using trained models, such as artificial intelligence and / or machine learning (AI / ML) to help determine or improve the results.

[0019] For example, the criteria used for an automatic delay when aspiration is to be paused can be based on measurements derived from sensors of intraoperative pressure. For instance, depending on when the intraoperative pressure stabilizes to an acceptable level or range, aspiration can be maintained for a period of time after laser pulverization has stopped.

[0020] In cases where flushing or other fluid delivery is based on an active mechanism such as a pump rather than a gravity-driven IV, fluid delivery can be similarly controlled, for example, based at least in part on one or more laser-activated times. For instance, both fluid delivery and suction can be automatically paused at the end of laser pulverization / crushing (or after a certain delay) (e.g., without any separate direct user input).

[0021] Figure 1This is a schematic diagram showing a portion of an endoscope system 100 having a suction component 104 for controlling suction and a laser generator 106.

[0022] Endoscopic system 100 may include endoscope 102, a suction component 104, a laser generator 106, and a fluid source 108. Endoscopic system 100 may interact with patient 110. For example, endoscopic system 110 may be used to access organs of patient 110, such as kidney 112. Kidney 112 may have stones 114.

[0023] Endoscope 102 may include one or more channels, such as a working channel, to allow a distal portion of the assistive device to be introduced into a target site within patient 110. Separate irrigation and aspiration channels may be included in the body of endoscope 102, or such functions may be performed using the working channel of endoscope 102. Regardless of the channel used, the distal portion of the aspiration component 104, the distal portion of the laser fiber from laser generator 106, or the distal portion of the irrigation channel in fluid communication with fluid source 108 may provide access to a target site within an organ (e.g., kidney 112). Endoscope 102 may include an illumination source (which may differ from laser generator 106) or may be coupled to an illumination source (which may differ from laser generator 106) via illumination fibers or fiber bundles. This can help provide a method of illuminating the kidney 112 distal to the endoscope, for example, to aid visual observation via endoscopic visualization optics. Endoscope 102 may be disposable or reusable. Endoscope 102 may be flexible, for example, to allow endoscope 102 to enter the kidney 112 or other target organs of patient 110 through an incision in patient 110, through the urethra of patient 110, or by other means.

[0024] Fluid source 108 can deliver saline or other fluids to the surgical site within the kidney 112 or other organs via endoscope 102. Fluid source 108 may include an intravenous infusion (IV) bag on a pole that delivers fluid by gravity. Fluid delivery using fluid source 108 can be controlled by a pressure band, such as squeezing the IV bag, by raising or lowering the IV bag, by a thumbwheel actuator, such as adjusting the valve opening, or by other means capable of modifying the rate of fluid delivery. Fluid source 108 can be controlled to supply fluid continuously. Fluid source 108 can be controlled to supply fluid at a constant rate. Fluid source 108 can be controlled to supply fluid at an adjustable or even variable rate. The rate of fluid from fluid source 108 can be manually controlled, for example, by the endoscope user or other caregiver. The rate of fluid from fluid source 108 can be controlled automatically or semi-automatically, for example, based on one or more detected or anticipated events, without requiring user intervention, although such user intervention may be permitted.

[0025] Laser generator 106 can deliver energy to stone 114, for example, via one or more optical fibers or bundles of optical fibers passing through endoscope 102, any of which can be referred to as laser fibers. For example, laser generator 106 can generate laser pulses to ablate stone 114 into smaller pieces (e.g., dust or fragments) that can be removed from or pass through patient 110. The ablation performed by laser generator 106 can produce stone dust or fragments. The stone dust or fragments can be sized to allow natural passage of the stone and can be removed using suction applied from suction component 104, or by means of a retrieval device or other medical equipment. In some embodiments, laser generator 106 can be replaced (or an additional energy source can be added) by another energy source, such as an ultrasound energy source. Laser generator 106 can be manually controlled by the user to trigger the delivery of laser energy to the target. Alternatively or additionally, the laser generator 106 can be automatically controlled, for example, to trigger the delivery of laser energy to the target without user intervention, such as when an AI / ML or other trained model identifies the stone targeted by the laser. Alternatively or additionally, the laser generator 106 can be semi-automatically controlled, for example, when an AI / ML or other trained model identifies the stone targeted by the laser, but requires user confirmation before actually triggering the delivery of laser energy. Alternatively or additionally, the energy level of the laser pulses emitted by the laser generator 106, or the pulse repetition rate or pattern of the laser generator 106, can be determined automatically or semi-automatically. Such automatic energy levels, repetition rates, or patterns can be issued automatically, semi-automatically, or can be issued by the user. Control of any of the functions of aspiration, flushing, ablation, or other functions can be accompanied by control of the positioning or movement of the endoscope 102 itself, for example in a fully robotic or robot-assisted method, or can be accompanied by control of one or more other end actuators associated with the endoscope 102.

[0026] The suction component 104 can provide suction to a target or working area of ​​an organ (e.g., kidney 112) via the endoscope 102, for example, through a suction channel or working channel of the endoscope 102. The suction component 104 can remove fluid from the kidney 112 using suction. The suction component 104 can also remove stone dust or fragments from the kidney 112 using suction. The suction rate can be established or adjusted to be substantially similar to the rate of flushing or other fluid delivery, such that a substantially constant amount of fluid is present in the working area and organ. For example, if the fluid delivery rate is high, the suction rate can be high. Similarly, if the fluid rate is low, the suction rate can be low. The fluid delivery rate can be established or modified at least in part based on the suction rate. Conversely, the suction rate can be established or modified at least in part based on the fluid delivery rate.

[0027] Endoscopic system 100 may include an imaging device, imaging optics, or both. For example, a camera or other imaging device may be incorporated into endoscope 102. The imaging device may be separated from endoscope 102—completely separated, or separated but insertable toward a target via endoscope 102. The imaging device may provide one or more still or video images of a working area of ​​an organ (e.g., kidney 112). Object recognition or other image processing, such as that described herein, may be performed on the images.

[0028] The endoscope system 100 may include or be coupled to a computing device or other controller circuitry, which may be implemented using hardware, software, firmware, or a combination thereof. The computing device may include a processor and a storage device. The storage device may include instructions that, when executed by the processor, cause the computing device to perform one or more actions or operations. The computing device may be capable of image processing, enabling and / or disabling the suction component 104, changing the suction rate of the suction component 104, enabling and / or disabling the laser generator 106, changing the energy level of the laser generator 106, enabling and / or disabling the fluid source 108, changing the fluid delivery rate of the fluid source 108, receiving sensing indications of one or more characteristics of the working environment from one or more sensors (e.g., temperature sensors, pressure sensors, etc.), or combinations thereof.

[0029] Figure 2A method 200 for controlling suction and / or flushing based on laser energy is shown. A computing device or controller circuitry system may implement method 200. At 202, an indication may be received that laser generator 106 is enabled, meaning laser generator 106 is providing laser energy. A user can enable laser generator 106 to deliver laser energy via direct user input. At 204, suction can be modified. Suction can be modified without direct user input. For example, when laser generator 106 is providing laser energy, in response, the computing device may trigger the opening of suction component 104 to provide suction or, for example, increase suction from a lower ambient or standby level. At 206, an indication may be received that laser generator 106 is deactivated, meaning laser generator 106 is not providing laser energy. At 208, suction can be modified. Suction can be modified without additional user input. When the laser generator 106 stops supplying laser energy, in response, the computing device can trigger the shutdown of the suction component 104 to stop supplying suction or reduce suction from a relatively high "active" suction level to a relatively low ambient or standby suction level. The laser generator 106 can be stopped supplying laser energy via direct user input to the laser generator 106. The fluid delivery rate of the fluid source 108 can be similarly and automatically controllably adjusted by the computing device in a manner similar to method 200. In this example, one or more parameters associated with the relative energy or power of the laser generator 106 can be used alternatively or additionally, for example, to trigger the computing device to change the fluid delivery rate, the suction rate, or both. For example, reducing the laser energy can trigger the computing device to reduce the suction rate of the suction component 104, while increasing the laser energy can trigger the computing device to increase the suction rate of the suction component 104. The computing device can predict when to modify pumping a second time based on the current pumping rate, current fluid velocity, current energy level, and current laser pulse mode, rather than waiting for an indication that the laser energy is deactivated. Therefore, pumping can be modified at 204 based on the indication received at 202, and at 208 without requiring an additional indication at 206.

[0030] The timing of the suction component 104 can also be established or manipulated, for example, in response to changes in or activation / deactivation of the laser generator 106 to modify the suction rate and / or to coordinate the activation or quantity of suction with changes in or activation / deactivation of the laser generator 106, without requiring precise simultaneous coordination. For example, the computing device can shut down (or de-energize) the suction component 104 at 204 at a certain time difference after (or even before) the laser generator 106 is shut down (or de-energized) at 206. For example, the laser generator 106 may have a predetermined laser energy delivery mode when the laser generator 106 is activated, such that the computing device can shut down (or de-energize) the suction before the laser generator 106 is deactivated (or the laser energy is reduced). The time difference can additionally or alternatively be based, in whole or in part, on time delay, temperature at the working area, pressure at the working area, and / or other measurement characteristics. Therefore, the computing device can enable (at 204) or disable (at 208) the suction component 104 and / or fluid source 108 based on a combination of settings or measurements (e.g., temperature, pressure, laser energy, or other measurement characteristics). For example, at 202, the laser generator 106 can begin delivering laser energy at time zero, and two seconds later, at 204, the suction component 104 and / or fluid source 108 can be enabled or increased. Similarly, at 206, the laser generator 106 can stop delivering laser energy at time x, and two seconds later, at 208, the suction component 104 and / or fluid source 108 can be disabled or reduced. As another example, the laser generator 106 can be enabled, but the laser energy can be set to be delayed, allowing suction to be enabled or modified before the delivery of laser energy. For example, suction can move the stone 114 to a location for the delivery of laser energy. Method 200 can be repeated when the laser generator 106 is enabled and disabled to ablate the stone 114.

[0031] The imaging device can capture one or more images of the working area for analysis by a computing device, for example, to determine how to automatically or semi-automatically control suction, flushing, or other operating parameters. Figure 3An example method 300 for controlling suction and / or flushing based on imaging of a working area is illustrated. At 302, the imaging device may capture one or more images of the working area. Stone dust may cause image blurring. At 304, image processing performed by a computing device may be used to determine the degree of image blurring. If no certain degree of blurring is found, the imaging device may continue to capture images at 302, and the image processing device may continue to determine the degree of image blurring at 304. At 306, when a certain degree of blurring is determined in the image, the computing device may enable the suction component 104 or modify the suction rate. Similarly, when a low degree of blurring is determined in the image, the computing device may deactivate the suction component 104. For example, the computing device may compare two or more images of the working area based on the degree of blurring, for example, to determine a blurring trend. The computing device may predict when to enable or deactivate the suction component 104 based on the blurring trend. Alternatively or additionally, method 300 may be used to control flushing or both flushing and suction of fluid source 108. Method 300 or a portion thereof can be repeated throughout the duration of the procedure. The computing device can calculate comprehensive indicators, such as the degree of ambiguity, aspiration rate, fluid rate, laser energy or state, or one or more other conditions of the working area, to determine or predict when to activate or deactivate the aspiration component 104.

[0032] Figure 4 An example of a method 400 for controlling aspiration and / or flushing based on imaging of a working area is shown. At 402, the imaging device can provide one or more images of the working area. At 404, the computing device can determine or track the proximity of the stone 114 or powder or fragments caused by the stone 114 to the distal end of the endoscope 102. At 406, based on such distance or proximity, when the distal end of the endoscope 102 is within a desired distance from the powder or fragments caused by the stone 114, the computing device can activate the aspiration component 104 or modify the aspiration rate. When the distal end of the endoscope 102 is beyond a specified or desired distance from the powder or fragments caused by the stone 114, the computing device can deactivate the aspiration component 104. Alternatively or additionally, method 400 can be used to control flushing of fluid source 108 or both flushing and aspiration. Method 400 or portions thereof can be repeated throughout the duration of the process.

[0033] Figure 5An example of a method 500 for controlling aspiration and / or flushing based on an imaging area is shown. At 502, the imaging device can provide one or more images of the working area. At 504, image processing performed by a computing device can be used to count the number of fragments in the images of the working area. Other image processing can be performed by the computing device, for example, to determine one or more fragments or the size of each fragment. At 506, using this information or similar information, the computing device can activate the aspiration component 104, for example, when the number of fragments is greater than a certain number. The computing device can activate the suction component 104 when the size of at least one fragment is less than a certain size. For example, the computing device can activate the suction component 104 when the size of at least one fragment is small enough to pass through the endoscope 102, for example, through its working channel or a separate suction channel. Additionally or alternatively, using this information or similar information, the computing device can issue a control signal to the suction component 104 to deactivate or reduce the suction applied by the suction component 104, for example, when the number of fragments is less than a certain number. For example, when the number of fragments reaches zero, the computing device can send a control signal to the suction component 104 to deactivate or deactivate the suction component 104. In this example, when no fragments below a specified threshold size are detected, the computing device can send a control signal to the suction component 104 to deactivate or deactivate the suction component 104. Alternatively or additionally, method 500 can be used to control flushing of fluid source 108, or both flushing and suction. Method 500 or a portion thereof can be repeated throughout the duration of the process.

[0034] Figure 6 An example method 600 for imaging-controlled aspiration and / or flushing based on a working area is illustrated. At 602, the imaging device can provide one or more images of the working area. At 604, image processing performed by a computing device can be used to distinguish between the stone 114 or powder or fragments of the stone 114 and the patient's organ. For example, the computing device can determine that powder or fragments of the stone 114 have settled on the tissue of the organ. Thus, at 606, when the distal endoscope 102 is within a specific distance from the powder or fragments of the stone 114, the computing device can issue a control signal to activate the aspiration component 104, and when the distal endoscope 102 is beyond a specific distance from the tissue, the computing device can issue a control signal to deactivate the aspiration component 104. For example, when image processing indicates that the tissue has been cleared or substantially cleared of the powder or fragments of the stone 114, the computing device can issue a control signal to deactivate the aspiration component 104. Alternatively or additionally, method 600 can be used to control the flushing of fluid source 108, or both flushing and suction. Method 600 or parts thereof can be repeated throughout the duration of the process.

[0035] Figure 7An example method 700 for determining a treatment plan based on imaging of stone 114 is illustrated. At 702, the imaging device can provide one or more images of the working area, including stone 114. At 704, a computing device can be programmed to determine one or more characteristics of stone 114 and / or powder or fragments of stone 114, for example, using information from one or more images from the imaging device, or by using spectral information from the response of the target stone 114 to illumination. The computing device can determine the size of stone 114 and / or powder or fragments of stone 114. The computing device can determine the material composition or properties of stone 114 and / or powder or fragments of stone 114. At 706, based on one or more characteristics of stone 114 and / or powder or fragments of stone 114, the computing device can determine a treatment plan. For example, the computing device can determine the energy level of a laser pulse emitted by laser generator 106. As another example, the computing device can determine the pulse repetition rate or pattern of laser generator 106. As another example, the computing device can determine the aspiration rate and / or aspiration pattern of the aspiration component 104. The computing device can determine the fluid delivery rate and / or fluid delivery pattern of the fluid source 108. Method 700 or a portion thereof can be repeated throughout the duration of the procedure. For example, the treatment plan can be modified throughout the procedure based on the effectiveness of the current treatment plan. Modification of the treatment plan may include pausing the delivery of ablation energy. For example, treatment may be paused until the image becomes substantially clear and the stone 114 (or the location of the stone 114) can be identified again. Figures 2 to 7 The methods described herein can be used independently or in combination.

[0036] Figure 8 A schematic diagram of an exemplary computer-based clinical decision support system (CDSS) 800 is shown, which is configured to enable fluid or aspiration, disable fluid or aspiration, modify the rate of fluid flow or aspiration, or predict when to provide such functionality based on the laser energy state of the work area and / or characteristic information, such as information about the kidney 112 or stone 114. Characteristic information may include images, intraoperative temperature, or intraoperative pressure. In various embodiments, CDSS 800 includes: an input interface 802 through which patient-specific characteristic information (e.g., images, temperature, pressure) is provided as input features to an artificial intelligence (AI) model 804; a processor that performs inference operations, wherein the characteristic information (e.g., images, temperature, pressure) is applied to the AI ​​model to generate control signals to enable, disable, or adjust the rate of aspiration or fluid flow; and a user interface (UI) through which fluid and aspiration rates and / or statuses are communicated to a user, such as a clinician.

[0037] In some implementations, the input interface 802 may be a direct data link between the CDSS 800 and one or more medical devices that generate at least some of the input features, such as imaging devices, spectroscopic devices, thermometers, pressure sensors, or other sensors that provide imaging, spectral, temperature, pressure, or other feature information. For example, during treatment and / or diagnostic medical procedures, the input interface 802 may directly transmit laser energy status and / or feature information (e.g., images, temperature, pressure) to the CDSS. Alternatively or additionally, the input interface 802 may be a conventional user interface that facilitates interaction between a user and the CDSS 800. For example, the input interface 802 may facilitate a user interface through which a user can manually input the laser energy treatment plan for the laser generator 106. Alternatively or additionally, the input interface 802 may provide the CDSS 800 with access to an electronic patient record from which one or more input features, such as previous imaging of the work area, can be extracted. In any of these cases, the input interface 802 is configured to collect one or more of the following input features associated with a specific patient at or before the time when the laser energy status and / or characteristic information (e.g., information about the kidney 112 or stone 114 as described above) is being assessed using the CDSS 800 to evaluate the work area.

[0038] Based on one or more of the above input features, the processor uses the AI ​​model to perform inference operations to generate control signals to enable or disable fluid or suction, modify the rate of fluid flow or suction, or predict when to provide such functionality. For example, input interface 802 can feed an image of the working area, including stone 114, to the input layer of the AI ​​model, which propagates these input features to the output layer through the AI ​​model. The AI ​​model can infer based on patterns discovered in data analysis, thereby providing the computer system with the ability to perform tasks without explicit programming. The AI ​​model explores the research and construction of algorithms (e.g., machine learning algorithms) that can learn from existing data and make predictions about new data. Such algorithms operate by building AI models from example training data to make data-driven predictions or decisions expressed as outputs or evaluations.

[0039] There are two common paradigms for machine learning (ML): supervised ML and unsupervised ML. Supervised ML uses prior knowledge (e.g., examples that correlate inputs with outputs or results) to learn the relationship between inputs and outputs. The goal of supervised ML is to learn a function that, given some training data, best approximates the relationship between training inputs and outputs, so that the ML model can achieve the same relationship given an input to generate the corresponding output. Unsupervised ML trains the ML algorithm using information that is neither classified nor labeled, allowing the algorithm to operate on that information without guidance. Unsupervised ML is useful in exploratory analytics because it can automatically identify structures in the data.

[0040] Common tasks used in supervised ML are classification and regression problems. Classification problems—also known as categorization problems—aim aim to classify an item into one of several category values ​​(e.g., is the object an apple or an orange?). Regression algorithms aim to quantify some items (e.g., by providing scores for some input values). Some examples of commonly used supervised ML algorithms are logistic regression (LR), Naive Bayes, random forest (RF), neural networks (NN), deep neural networks (DNN), matrix factorization, and support vector machines (SVM).

[0041] Some common tasks used in unsupervised ML include clustering, representation learning, and density estimation. Examples of commonly used unsupervised ML algorithms are K-means clustering, principal component analysis, and autoencoders.

[0042] Another type of machine learning is federated learning (also known as collaborative learning), which trains algorithms on multiple distributed devices that store local data without exchanging data. This approach contrasts with traditional centralized machine learning techniques, where all local datasets are uploaded to a single server, and with more classic distributed methods that typically assume local data samples are uniformly distributed. Federated learning enables multiple participants to build general, robust machine learning models without sharing data, thus allowing for the resolution of critical issues such as data privacy, data security, data access permissions, and access to heterogeneous data.

[0043] In some examples, the AI ​​model can be trained continuously or periodically before the processor performs inference operations. Then, during the inference operation, patient-specific input features provided to the AI ​​model can propagate from the input layer through one or more hidden layers and ultimately to the output layer, which corresponds to control signals that enable or disable fluid or aspiration, modify the rate of fluid flow or aspiration, or predict when to provide such functionality. For example, AI model 804 can be trained to achieve this individually or in combination. Figures 2 to 7The method described in any of them.

[0044] The methods described herein can be implemented, at least in part, by a machine or computer. Some examples may include computer-readable or machine-readable media encoded with instructions operable to configure electronic devices to perform the methods described in the examples above. Implementations of such methods may include code, such as microcode, assembly language code, high-level language code, etc. Such code may include computer-readable instructions for performing various methods. The code may form part of a computer program product. Furthermore, in the examples, the code may be tangibly stored, for example during execution or at other times, on one or more volatile, non-transitory, or non-volatile tangible computer-readable media. Examples of such tangible computer-readable media may include, but are not limited to, hard disks, removable disks, removable optical discs (e.g., compact discs and digital video discs), magnetic tape cartridges, memory cards or memory sticks, random access memory (RAM), read-only memory (ROM), etc.

[0045] Furthermore, the present invention can also be configured as follows.

[0046] (1). An endoscopic system for coordinating the operation of at least one of an ablation mode and an assisted aspiration or flushing mode, the endoscopic system comprising:

[0047] A controller circuit system configured to issue control signals to coordinate the timing or other parameters of at least one of the auxiliary aspiration or flushing modes with the ablation timing or other ablation parameters of the ablation energy emitted by the ablation mode toward the ablation target within the patient, wherein the control signals are issued automatically without requiring direct user input to control at least one of the auxiliary aspiration or flushing modes, to provide a relative increase in at least one of the aspiration or flushing in time in coordination with the ablation energy emitted by the ablation mode.

[0048] (2). The endoscope system according to (1) further includes:

[0049] At least one of an imaging device, a spectroscopic device, or other sensor is configured to provide imaging, spectral, or other characteristic information regarding at least one of the ablation target within the patient or the ablation target environment within the patient; and

[0050] The controller circuitry generates the control signal based at least in part on information from the sensor regarding at least one of the ablation target within the patient or the ablation target environment within the patient.

[0051] (3). The endoscope system according to (2), wherein the controller circuit system performs the following operations:

[0052] Receive a first image of the ablation target environment from the imaging device;

[0053] Identify the powder or fragments of the ablation target from the first image; and

[0054] The control signal is generated based on feature information about the powder or fragments identified in the first image.

[0055] (4). The endoscope system according to (3), wherein the feature information is the degree of blurring in the first image caused by the powder of the ablation target.

[0056] (5). The endoscope system according to (3), wherein the feature information is the number of fragments of the ablation target in the first image.

[0057] (6). The endoscope system according to (3), wherein the feature information is the amount of powder or fragments of the ablation target remaining in the ablation target environment in the first image.

[0058] (7). The endoscope system according to (3), wherein the controller circuit system performs the following operations:

[0059] Receive a second image of the ablation target environment from the imaging device;

[0060] Identify the powder or fragments of the ablation target from the second image; and

[0061] The control signal is generated based on feature information about the powder or fragments identified in the second image.

[0062] (8). The endoscope system according to (7), wherein the feature information is the change in the degree of blurring caused by the powder of the ablation target between the first image and the second image.

[0063] (9). The endoscope system according to (7), wherein the feature information is based on the trend of the degree of blurring caused by the powder of the ablation target in the first image and the second image.

[0064] (10). The endoscope system according to (9), wherein the controller circuitry predicts when to generate the control signal based on the trend of the degree of blurring caused by the powder of the ablation target based on the first image and the second image.

[0065] (11). The endoscope system according to (7), wherein the feature information is the change in the number of fragments of the ablation target between the first image and the second image.

[0066] (12). The endoscope system according to (7), wherein the feature information is the change in the amount of powder or fragments of the ablation target remaining in the ablation target environment between the first image and the second image.

[0067] (13). The endoscope system according to (2), wherein the feature information is the intraoperative pressure or intraoperative temperature of the ablation target environment.

[0068] (14). The endoscope system according to (2), wherein the controller circuit system predicts the duration between the first control signal and the second control signal based on the feature information.

[0069] (15). The endoscope system according to (1), wherein the control signal for coordinating the timing or other parameters of at least one of the auxiliary aspiration or flushing modes with the ablation timing or other ablation parameters of the ablation energy emitted by the ablation mode is emitted with a time offset from the ablation timing of the ablation energy.

[0070] (16). A method for coordinating the operation of at least one of an ablation mode and an auxiliary aspiration or flushing mode, the method comprising:

[0071] A control signal is issued to coordinate the timing or other parameters of at least one of the auxiliary aspiration or flushing modes with the ablation timing or other ablation parameters of the ablation energy emitted by the ablation mode toward the ablation target within the patient, wherein the control signal is issued automatically without requiring direct user input to control at least one of the auxiliary aspiration or flushing modes, to provide a relative increase in at least one of the aspiration or flushing in time in coordination with the ablation energy emitted by the ablation mode.

[0072] (17). The method according to (16) further includes:

[0073] Receive imaging, spectral, or other characteristic information regarding at least one of the ablation target within the patient or the ablation target environment within the patient; and

[0074] The control signal is generated based at least in part on information from the sensor regarding at least one of the ablation target within the patient or the ablation target environment within the patient.

[0075] (18). The method according to (17) further includes:

[0076] Receive a first image of the ablation target environment;

[0077] Identify the powder or fragments of the ablation target from the first image; and

[0078] The control signal is generated based on feature information about the powder or fragments identified in the first image.

[0079] (19). According to the method of (18), wherein the feature information is the degree of blurring in the first image caused by the powder of the ablation target.

[0080] (20). According to the method of (18), wherein the feature information is the number of fragments of the ablation target in the first image.

[0081] (21). According to the method of (18), wherein the feature information is the amount of powder or fragments of the ablation target remaining in the ablation target environment in the first image.

[0082] (22). The method according to (18) further includes:

[0083] Receive a second image of the ablation target environment;

[0084] Identify the powder or fragments of the ablation target from the second image; and

[0085] The control signal is generated based on feature information about the powder or fragments identified in the second image.

[0086] (23). According to the method of (22), wherein the feature information is the change in the degree of blurring caused by the powder of the ablation target between the first image and the second image.

[0087] (24). According to the method of (22), wherein the feature information is based on the trend of the degree of blurring caused by the powder of the ablation target in the first image and the second image.

[0088] (25). The method according to (24) further includes: predicting when to generate the control signal based on the trend of the degree of blurring caused by the powder of the ablation target based on the first image and the second image.

[0089] (26). According to the method of (22), wherein the feature information is the change in the number of fragments of the ablation target between the first image and the second image.

[0090] (27). According to the method of (22), wherein the feature information is the change in the amount of powder or fragments of the ablation target remaining in the ablation target environment between the first image and the second image.

[0091] (28). According to the method of (17), wherein the feature information is the intraoperative pressure or intraoperative temperature of the ablation target environment.

[0092] (29). An endoscopic system for coordinating the operation of at least one of an ablation modality and an assisted aspiration or flushing modality, the endoscopic system comprising:

[0093] A controller circuit system configured to issue a treatment plan to coordinate the timing or other parameters with the ablation timing or other ablation parameters of the ablation energy emitted by the ablation modality toward the ablation target within the patient;

[0094] At least one of an imaging device, a spectroscopic device, or other sensor is configured to provide imaging, spectral, or other characteristic information regarding at least one of the ablation target within the patient or the ablation target environment within the patient; and

[0095] The controller circuitry generates the treatment based at least in part on information from the sensors regarding at least one of the ablation target within the patient or the ablation target environment within the patient.

[0096] (30). The endoscope system according to (29), wherein the treatment plan further coordinates the timing or other parameters of at least one of the auxiliary aspiration or flushing modes.

Claims

1. An endoscope system comprising: controller circuitry configured to issue control signals to coordinate timing or other parameters of a fluid aspiration or irrigation device with timing or other parameters of an ablation member, wherein the control signals are issued automatically using a trained learning model that coordinates the timing or other parameters of the fluid aspiration or irrigation device with the timing or other parameters of the ablation member based at least in part on at least one of real-time data or historical data from one or more sensors communicatively coupled to the controller circuitry.

2. The endoscope system of claim 1, wherein, the controller circuitry is configured to control aspiration based on an intraoperative pressure of an ablation target environment.

3. The endoscope system of claim 2, wherein, the controller circuitry is configured to maintain aspiration for a period of time after laser ablation has stopped in dependence on the intraoperative pressure being in an acceptable level or range.

4. The endoscope system of claim 1, wherein, the controller circuitry is configured to make one or more predictive adjustments to at least one of aspiration, irrigation, or ablation based on trends derived from the historical data.

5. The endoscope system of claim 4, wherein, the controller circuitry is configured to predict when to generate the control signals based on trends in a degree of obscuring caused by ablation of an ablation target.

6. The endoscope system of claim 1, further comprising: an illumination light source coupled to an endoscope of the endoscope system via an illumination optical fiber or fiber bundle to enhance imaging of a target site.

7. The endoscope system of claim 1, wherein, the controller circuitry is configured to automatically control at least one of i) a position of an endoscope coupled to the endoscope system or ii) an end effector connected to the endoscope.

8. The endoscope system of claim 1, wherein, the controller circuitry is configured to automatically pause both fluid delivery and aspiration at the end of laser ablation.

9. The endoscope system of claim 1, wherein, the controller circuitry is configured to determine a treatment plan based on one or more characteristics of an ablation target, wherein the one or more characteristics of the ablation target include at least one of a size, a material composition, or other properties of the ablation target.

10. The endoscope system of claim 9, wherein, the treatment plan includes at least one of an energy level of laser pulses from a laser generator coupled to the endoscope system; at least one of a pulse repetition rate or pattern of the laser generator; at least one of an aspiration rate or aspiration pattern; or at least one of a fluid delivery rate and / or a fluid delivery pattern.

Images