Long-term air leakage perception

Abstract

A device may include a processor. The processor may be configured to receive data during a procedure. A processor may receive data representative of an intraoperative air exchange of a patient. The data may represent an air exchange of a breathing cycle of the patient. For example, the data may include any of the following: respirator inlet flow, respirator inlet pressure, respirator output pressure, thoracic duct flow, or thoracic duct pressure. The processor may also receive data representative of surgical parameters other than the surgical parameters associated with the air exchange. For example, the data may include any of patient medical record data, intra-operative report data, surgical data, etc. The processor may be configured to receive an updated machine learning model from a cloud resource. And the processor may output information indicative of a likelihood of long-term air leakage during the procedure based on the data.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims priority to and the benefit of U.S. Provisional Application No. 63 / 363,631, filed on April 26, 2022, under 35 USC § 119(e), which is incorporated herein by reference in its entirety. background

[0002] A prolonged air leak occurs when air escapes from a patient's lungs into the chest cavity for an unacceptable length of time (usually 5 to 7 days). This leak may occur after lung surgery, traumatic injury, lung biopsy, etc. In addition, a significant number of patients may experience a prolonged air leak after lung resection. This complication can be associated with increased hospital stays and major postoperative morbidity (e.g., lobar collapse, nosocomial pneumonia, and pleural effusion). All of these lead to higher healthcare costs and worse patient outcomes.

[0003] Although preventive strategies have been investigated to mitigate the risk of prolonged air leaks (preventive strategies include, for example, surgical techniques, sealants, and buttress materials), no single strategy has been shown to be unequivocally effective. Overview

[0004] The device may include a processor. The processor may be configured to receive data during surgery. The processor may receive data representing intraoperative air exchange of the patient. The data may represent air exchange of the patient's breathing cycle. For example, the data may include any of the following: ventilator inlet flow, ventilator inlet pressure, ventilator output pressure, chest tube flow, or chest tube pressure. The processor may also receive data representing surgical parameters other than surgical parameters related to air exchange. For example, the data may include any of patient medical record data, intraoperative report data, surgical operation data, etc. The processor may be configured to receive an updated machine learning model from a cloud resource. And the processor may output information indicating the possibility of long-term air leak based on the data during surgery. BRIEF DESCRIPTION OF THE DRAWINGS

[0005] Figure 1 is a depiction of a computed tomography (CT) scan of a patient's chest.

[0006] Figure 2 is a depiction of a lung resection with a chest tube.

[0007] Figure 3 An example signal of a lung without prolonged air leak is shown.

[0008] Figure 4 An example signal of a lung that does experience a prolonged air leak is shown.

[0009] Figure 5 is a graph of an example signal.

[0010] Figure 6 is a graph of an example signal including two time series.

[0011] Figure 7 is a block diagram illustrating an example system for determining the likelihood of a prolonged air leak.

[0012] Figure 8 is a block diagram of an example device for collecting signals and for determining the likelihood of a prolonged air leak.

[0013] Fig. 9 is a flow chart of an example process for collecting signals and for determining the likelihood of a prolonged air leak.

[0014] Fig.10 is a diagram of an example device for collecting signals and for determining the likelihood of a prolonged air leak.

[0015] Fig.11 is a block diagram of an example device for collecting signals and for determining the likelihood of a prolonged air leak.

[0016] Fig.12 is a flow chart of an example process for collecting signals and for determining the likelihood of a prolonged air leak.

[0017] Fig.13 is a diagram of an example device for collecting signals and for determining the likelihood of a prolonged air leak.

[0018] Fig.14 is a block diagram of a system including an exemplary plurality of surgical computing devices.

[0019] Fig.15 is a block diagram illustrating an example system for determining the likelihood of a prolonged air leak.

[0020] Fig.16 is a flow chart of an example process for collecting signals and for determining the likelihood of a prolonged air leak. Detailed Description

[0021] Figure 1 is a depiction of a computed tomography (CT) scan of a patient's chest (e.g., thorax) cavity 100. The thoracic cavity 100 contains lungs 102, heart 104, and other organs (such as major blood vessels). The surface 106 of the lungs and the interior of the chest wall 108 are covered by multiple membranes. In a healthy patient, the lungs 102 can fully expand within the cavity 100 because the pressure within the lungs 102 is generally higher than the pressure within the pleural space 110.

[0022] However, here, the CT scan shows a pneumothorax 112 on the left side of the patient (right side of the image). In certain circumstances, such as damage to the chest wall 108 and / or damage to one or both lungs 102, a pneumothorax 112 may occur. Air may enter the pleural cavity 110. Intrapleural pressure may increase. This increase in pressure may cause the pressure difference between the lung pressure and the intrapleural pressure to normalize. This may cause one or both lungs 102 to collapse, which is a life-threatening condition for the patient.

[0023] During thoracic surgery, particularly where portions of the lungs are removed, a typical outcome is that some air leaks from the lungs 102 and into the pleural cavity 110. These air leaks generally decrease as any surgical wounds to the lungs 102 heal. However, when such air leaks do not decrease within an appropriate amount of time, they are considered to be prolonged air leaks (PALs), and the consequences to the patient can be significant. Therefore, a chest tube may be used post-operatively to allow the leaked air to escape the pleural cavity 110.

[0024] Figure 2 is a depiction of a lung resection with a chest tube 200. Lung resection surgery may involve removing an entire lung, such as a pneumonectomy; removing a lobe, such as a lobectomy; or removing a portion of a lobe, such as a segmentectomy or a wedge resection. The remaining portion of the lung 202 may have a wound from the resection surgery that has been surgically closed. Such closed wound 204 may be closed using any suitable surgical technique, such as sutures, surgical glue, staples, etc., alone or in combination.

[0025] Because there is usually an air leak when the lung 202 and the closed wound 204 heal, a chest tube 200 is usually installed when the lung resection surgery is completed. The chest tube 200 allows air to leak from the lung into the pleural cavity to escape the body. The chest tube 200 can also allow fluid drainage. In some cases, more than one chest tube 200 can be used. The chest tube 200 can be connected to a drainage system, such as an underwater drainage system. The drainage system can be connected to a wall suction device. And when the air leak and / or fluid drainage has stopped or has reached an acceptable low amount, the chest tube can be removed. When this is not the case, the patient has a long-term air leak, which threatens the patient's recovery and may require further treatment or even subsequent surgical intervention. Figure 3 and Figure 4 A method is shown for predicting whether an air leak will heal or will persist for multiple days and become a chronic leak. Such a prediction, especially if available during surgery, could enable the surgeon to take additional steps to mitigate or prevent this dangerous complication.

[0026] Figure 3An example signal 300 of a lung that does not have a long-term air leak is shown. Lung 302 is considered as a system and respiration 304 is considered as an input to the system. Granular measurements of air leakage parameters can be used as signals representing the system. For example, the air leakage parameter can include a granular measurement of the mass air flow through the chest tube. For example, the parameter can be quickly measured and / or sampled during a complete breath. The resulting signal 300 may include a time series of the parameter. One part may be related to inspiration (e.g., in breath) and another part may be related to expiration (e.g., out breath). Respiration 304 itself can be provided by the patient, a ventilator, etc. And the measurement can be performed, for example, after the lung is removed but before closing the patient and ending the operation.

[0027] Here, the leak rate, measured as airflow through the chest tube, is sampled at a significantly high frequency to capture changes in the rate of airflow through the chest tube during a breath 304. These changes over a short duration (e.g., approximately 4000 ms) can provide detailed insight into the health of the lungs 302 and can be used to predict the likelihood of a prolonged air leak. In this example signal 300, a breath cycle 304 through the lungs 302, for example, generates a relatively low and flat signal 300. Analysis of such a signal can predict that a patient will not experience a prolonged air leak after a lung resection.

[0028] As a comparison, Figure 4 An example signal 400 is shown for a lung 402 that does experience a prolonged air leak. This signal 400 is significantly different from the signal 300 produced by a lung that does not experience a prolonged air leak. Here, the example signal 400 generally oscillates, showing a correlation with the inspiration and expiration of breathing 404. The example signal 400 contains multiple defined peaks and valleys. Such peaks and valleys can be characterized by their number, frequency, amplitude, etc. The overall flow is generally high. The air leakage characteristics of the lung 402 can be represented by the nature of these intra-breath variations. Such variations can be made detectable by appropriate measurement granularity (e.g., sampling rate). And analysis of such a signal 400 can predict that the patient will not experience a prolonged air leak after lung resection surgery.

[0029] Figure 3 and Figure 4The air leak parameters shown may include one or more measurable parameters associated with air entering, leaving, and / or passing through the lungs and / or thoracic cavity. For example, the air leak parameters may include one or more of ventilator inlet flow, ventilator inlet pressure, ventilator output flow, ventilator output pressure, chest tube flow, chest tube pressure, etc. For example, the signals 300, 400 may include a time series of one or more such parameters.

[0030] For example, Figure 5 and Figure 6 Several example signals that may be used to assess the likelihood that a patient has a prolonged air leak are shown. Figure 5 500 is a graph of example signals. It is a graph of chest tube leak rate over time for three breathing cycles of lungs with different physical characteristics. The first signal 500 is the signal of an example healthy lung with no holes. The second signal 502 is the signal of a lung with a 3 cm hole in the lower lobe. And the third signal 504 is the signal of a more significantly damaged lung.

[0031] The first signal 500 shows a generally flat profile with characteristic peaks during the inspiratory portion of the breathing cycle. The second signal 502 exhibits certain peaks and valleys, an overall oscillation with breathing, and a relatively low overall amplitude. The overall oscillation is shown with the overlay line 506. The third signal 504 is similar to the second signal 502, but has a larger overall leak rate and oscillation amplitude (as shown by the overlay line 508).

[0032] Figure 6 is a graph of an example signal including two time series. Here, the first time series 600 is a time series of chest tube airflow rate. And the second time series 602 is a time series of chest tube pressure. This example signal has two dimensions, one for flow and the other for pressure. Here, not only does a single time series contain information to predict long-term air leakage, but the relationship between the two time series also contains information to predict long-term air leakage. The example signal may include one or more time series. For example, the time series may be time-dependent. For example, the time series may represent the duration of a concurrent event. Here, the correlation and / or relationship between the pressure curve and the flow curve may include the number of peaks, the variance between peaks, the total area under the curve, etc.

[0033] In an example, the signal and / or one or more characteristics of the signal can be used to assess the possibility of a prolonged air leak. For example, the signal and / or one or more characteristics of the signal can be used to determine whether the current lung is in a state where a prolonged air leak is likely or unlikely. The signal characteristics can include the number of peaks in the flow and / or pressure curve during inspiration and / or exhalation, the total leak per breathing cycle (area under the flow curve), the difference between two different peaks of the flow and / or pressure curve, etc.

[0034] like Figure 5 and Figure 6 As shown, the measured parameters are sampled at a relatively high rate. A digital chest tube drainage system can provide a static or "snapshot" view of the flow. And a health care professional can determine whether a patient can safely remove the chest tube through a static view over hours and / or days. However, a static or "snapshot" view cannot capture intra-breathing changes indicative of lung leakage, which can be used to train a predictive model and ultimately used to predict the likelihood of a prolonged air leak. The devices, systems, and methods disclosed herein can employ sampling of the measured parameters at a higher rate that is suitable for capturing these indicative intra-breathing changes.

[0035] The selection of the suitability of the sampling rate can be discerned based on the Nyquist rate associated with the measured analog parameter. For example, chest tube rate and pressure and / or parameters associated with air exchange can represent continuous functions. The continuous function can be characterized by a frequency range. The rate of the sampler is selected to be greater than twice the highest relevant frequency of the continuous function, and the resulting discrete time series can be distortion-free and can be used to preserve and / or recreate the information present in the original relevant signal with the desired fidelity. In an example, the frequency range or spectrum of the continuous signal associated with airflow exchange can be determined for a population, and then the appropriate Nyquist rate for the sampler can be selected for use with the parameter and / or population. In an example, the sampling rate can include a sampling period of 1 second, 500 milliseconds, 100 milliseconds, 50 milliseconds, or faster.

[0036] Signals and / or characteristics of signals may be analyzed for a population. One or more signal profiles may be developed for a population. Such profiles may be associated with different degrees of likelihood of a prolonged air leak. For example, signals and / or characteristics of signals may be analyzed for a population of resection surgeries associated with no prolonged air leak. For example, signals and / or characteristics of signals may be analyzed for a population of resection surgeries associated with the occurrence of a prolonged air leak. Subsequent signals and / or signal characteristics may be compared with one or more profiles and / or other signals to assess whether a particular subsequent signal is more likely to be a signal for a population of resection surgeries associated with no prolonged air leak, or more likely to be a signal for a population of resection surgeries associated with the occurrence of a prolonged air leak.

[0037] The computer system can be configured to perform the analysis of the signal and / or signal characteristics of the colony. The computer system can be configured with an analysis program to assess the likelihood. The analysis program can include any process suitable for comparing the signal and / or signal characteristics with one or more signal and / or signal characteristics colonies. Analysis programming can include methods such as regression analysis, linear regression, nonlinear regression, vector autoregression and / or machine learning. For example, the computer system can be configured to develop a prediction model.

[0038] Figure 7 700 is a block diagram illustrating an example system for determining the likelihood of a prolonged air leak. Surgical data 700 may be captured from one or more resection surgeries. For example, surgical data 700 may be captured from a large number of resection surgeries. For example, data may be collected from at least one hundred resection surgeries. Surgical data 700 may include one or more signals indicating air exchange for a corresponding patient. Surgical data 700 may be combined with medical record data to generate one or more data elements 702, each element comprising at least a signal and / or signal characteristic and a corresponding patient outcome. For example, a patient outcome may include whether a patient has a prolonged air leak. For example, a patient outcome may include the number of days a leak exists and on which day the leak reaches an acceptable level.

[0039] The surgical data 700 and / or data elements 702 may be stored and processed by a surgical computer system 704. The surgical computer system 704 may include a processor 706 and / or a data repository 708. The surgical computer system 704 may operate on a computer network 710. The collected data may populate the data repository 708 so that a predictive model 712 may be generated. The processor 706 may be configured to generate a predictive model 712 based on, for example, the surgical data 700 and / or data elements 702 stored in the data repository 708. The predictive model 712 may be generated by conventional data analysis techniques, machine learning, etc.

[0040] The prediction model 712 can be generated by the processor 706 according to any suitable machine learning technology. For example, the processor 706 can use a supervised learning algorithm. The supervised learning algorithm can create a mathematical model through a training data set (e.g., training data). The training data can be composed of a set of training samples. The training sample can include one or more inputs and one or more labeled outputs. The data element 702 can be used as training data. The labeled output can be used as supervised feedback. In the mathematical model, the training sample can be represented by an array or vector, sometimes referred to as a feature vector. The training data can be represented by the rows of the feature vectors that constitute the matrix. Through iterative optimization of the objective function (e.g., cost function), the supervised learning algorithm can learn a function (e.g., prediction function) that can be used to predict the output associated with one or more new inputs. The prediction function properly trained can determine the output of one or more inputs that may not be part of the training data. Example algorithms can include linear regression, logistic regression, and neutral networks. Example problems that can be solved by supervised learning algorithms can include classification, regression problems, etc.

[0041] In an example, the processor 706 may use an unsupervised algorithm to develop a predictive model 712. An unsupervised learning algorithm may be trained on a data set that may contain inputs, and may find structures in the data. The structure in the data may be similar to grouping or clustering of data points. Therefore, the algorithm may learn from training data that may not yet be labeled. Instead of responding to supervised feedback, the unsupervised learning algorithm may identify commonalities in the training data, and may react based on the presence or absence of such commonalities in each training sample. Surgical data 700 may be used as training data. Example algorithms may include Apriori algorithm, K-means, K-nearest neighbors (KNN), K-median, and the like. Example problems that may be solved by an unsupervised learning algorithm may include clustering problems, anomaly / outlier detection problems, and the like.

[0042] In an example, the processor 706 may use a reinforcement learning algorithm to develop the prediction model 712. Reinforcement learning is a field of machine learning that may involve the concept of how a software agent can take actions in an environment to maximize a cumulative reward. Reinforcement learning algorithms may not assume knowledge of an exact mathematical model of the environment (e.g., represented by a Markov decision process (MDP)) and may be used when an exact model may not be feasible. Reinforcement learning algorithms may be used in autonomous vehicles or to learn to play games with human opponents.

[0043] For example, the output of the training process of machine learning can be a model for predicting results on a new data set, such as a prediction model 712 used with subsequent signal information 714. The linear regression learning algorithm can be a cost function that can minimize the prediction error of the linear prediction function during the training process by adjusting the coefficients and constants of the linear prediction function. When the minimum value can be reached, the linear prediction function with the adjusted coefficients can be considered to be trained and constitute the model that the training process has generated. For example, a neural network (NN) algorithm (e.g., a multilayer perceptron (MLP)) for classification can include a hypothesis function represented by a network of node layers, which are assigned biases and interconnected with weight connections. The hypothesis function can be a nonlinear function (e.g., a highly nonlinear function), which can include linear functions and logic functions nested together, wherein the outermost layer consists of one or more logic functions. The NN algorithm can include a cost function to minimize the classification error by adjusting biases and weights through the process of feedforward propagation and backpropagation. When the global minimum value can be reached, the optimized hypothesis function with its adjusted bias and weight layers can be considered to be trained and constitute the model that the training process has generated.

[0044] Processor 706 can be used to perform any element of the machine learning life cycle, including stages such as data collection, data preparation, model training, model deployment, post-deployment, etc.

[0045] As the first stage of the machine learning life cycle, data collection can be performed for machine learning. For example, data collection can include steps such as identifying various data sources, collecting data from the data sources, and integrating the data.

[0046] Data preparation can include data preprocessing steps, such as data formatting, data cleaning, and data sampling. Data preparation can include data transformation processes (e.g., after preprocessing), such as scaling and aggregation. For example, preprocessed data can include data values ​​at various scales. These values ​​can be scaled up or down, such as between 0 and 1, for model training. For example, preprocessed data can include data values ​​that carry more meaning when aggregated.

[0047] Model training involves applying an appropriate machine learning algorithm to the prepared data. After the model has been trained, cross-validated, and tested, the model can be considered to have been properly trained. Therefore, the dataset from the data preparation stage (e.g., the input dataset) can be divided into a training dataset (e.g., 60% of the input dataset), a validation dataset (e.g., 20% of the input dataset), and a test dataset (e.g., 20% of the input dataset). After the model has been trained on the training dataset, the model can be run against the validation dataset to reduce overfitting. If the accuracy of the model decreases when run against the validation dataset while the accuracy of the model is constantly improving, this may indicate an overfitting problem. The test dataset can be used to test the accuracy of the final model to determine whether it is ready for deployment or requires more training.

[0048] Model deployment may include how the model is used. For example, the model may be deployed as part of a standalone computer program. The model may be deployed as part of a larger computing system. In an example, the prediction model 712 may be deployed in: a computer system, an embedded system, a surgical computer system (e.g., a surgical hub), a cloud-based system, etc. For example, the prediction model 712 may be deployed in the systems, devices, and methods disclosed herein.

[0049] Models can be deployed with model performance parameters. Such performance parameters can monitor model accuracy as it is used to predict datasets in production. For example, such parameters can track false positives and false negatives for a classification model. Such parameters can further store false positives and false negatives for further processing to improve the accuracy of the model.

[0050] Post-deployment model updates can be another aspect of the machine learning cycle. For example, when false positives and / or false negatives are predicted on production data, the deployed model can be updated. In the example, for the deployed MLP model for classification, when false positives occur, the deployed MLP model can be updated to increase the possible cutoff value for predicting positives to reduce false positives. In the example, for the deployed MLP model for classification, when false negatives occur, the deployed MLP model can be updated to lower the possible cutoff value for predicting positives to reduce false negatives. In the example, for the deployed MLP model for surgical complication classification, when both false positives and false negatives occur, the deployed MLP model can be updated to lower the possible cutoff value for predicting positives to reduce false negatives, because predicting false positives may not be as critical as predicting false negatives.

[0051] For example, a deployed model may be updated when more real-time production data becomes available as training data. In this case, the deployed model may be further trained, validated, and tested using this additional real-time production data. In an example, the updated biases and weights of the further trained MLP model may update the biases and weights of the deployed MLP model.

[0052] In an example, a prediction model 714 may be generated, validated, and ultimately deployed. For example, a subsequent signal 714 may be input to a prediction model 712 to provide an output 716. The output 716 may include a probability that the subsequent signal 714 is associated with the occurrence of a prolonged air leak. The subsequent signal 714 may be a signal collected from surgical data that is not part of the surgical data 700 used to generate the prediction model 712. For example, the subsequent signal 714 may include a test signal for which the patient's outcome is known. The test signal may be used to confirm the accuracy of the prediction model. For example, the subsequent signal 714 may include a new patient signal from a patient whose outcome is not yet known. The new patient signal may be used so that a surgeon or other health care professional may use the output of the prediction model to assess the likelihood of a patient having a prolonged air leak. In this capacity, such a model 712 may enable intraoperative intervention and / or earlier postoperative intervention.

[0053] Such a model 712 provides substantial clinical and economic benefits to patients, clinicians, healthcare facilities, etc. For example, by having such information in the operating room, a surgeon can perform further surgical tasks to address potential leaks that would otherwise be considered unnecessary. For example, a surgeon can provide additional sealants, sutures, staples, etc. to a wound in the lung. For example, a surgeon can perform further diagnostics associated with a lung leak, such as submerging it in water to discover a leak that has not yet been seen. For example, a surgeon can provide additional surgical relief care. Such information about the likelihood of a prolonged air leak may be particularly helpful during resection surgery, as these additional surgical activities can be completed while the patient is still in the operating room prior to closure, without the need for subsequent surgical intervention.

[0054] Figure 8 800 is a block diagram of an example device 800 for collecting signals and / or for determining the likelihood of a prolonged air leak. For example, the device 800 can be adapted to capture signal information of a population for the purpose of training a predictive model. For example, the device 800 can be adapted to capture subsequent signal information, such as test or new patient signals, in order to consider the output of the predictive model. For example, the device 800 can be used to capture subsequent signals to be input into the predictive model and provide a surgeon with a predicted patient outcome, such as the likelihood of the patient experiencing a prolonged air leak.

[0055] The device 800 may include one or more sensors 802, 804, 806, 808, 810, 812. The device 800 may include any sensor suitable for collecting signals indicating patient air exchange. For example, the device 800 may include any sensor suitable for collecting signals indicating air exchange between a surgical ventilator and a patient. For example, the device 800 may include any sensor suitable for collecting signals indicating that air leaves a patient via a chest tube. For example, the device 800 may include one or more sensors, such as a ventilator inlet flow sensor 802, a ventilator inlet pressure sensor 804, a ventilator outlet flow sensor 806, a ventilator outlet pressure sensor 808, a chest tube flow sensor 810, a chest tube pressure sensor 812, etc. Such sensors may be suitable for recording surgical quality data from a patient. For example, air flow data may be measured with a tolerance of + / - 0.4 liters / minute. For example, pressure measurements may be performed with a tolerance of + / - 2.5 cm H20.

[0056] Flow sensors 802, 806, 810 may include any sensor suitable for capturing the flow of air. For example, flow sensors 802, 806, 810 may include sensors or flow meters designed for medical applications. For example, flow sensors 802, 806, 810 may be used to withstand autoclave procedures. For example, flow sensors 802, 806, 810 may be packaged for single use and / or for multiple use. For example, flow sensors 802, 806, 810 may be designed for medical ventilation or breathing applications. In an example, flow sensors 802, 806, 810 may include analog sensors and / or digital sensors. For example, flow sensors 802, 806, 810 may include one or more silicon sensor chips. For example, one or more flow sensors 802, 806, 810 may include related support circuits, such as amplifiers, integrated A / D converters, EEPROM memories, digital signal processing circuits, and interface circuits. In an example, one or more of the flow sensors 802, 806, 810 may include an SFM3400 digital flow meter from Sensirion (TM).

[0057] Pressure sensors 804, 808, 812 may include any sensor suitable for measuring air pressure. For example, pressure sensors 804, 808, 812 may include any sensor suitable for measuring absolute air pressure, gauge pressure and / or differential pressure. Pressure sensors 804, 808, 812 may include sensors or pressure gauges designed for medical applications. For example, pressure sensors 804, 808, 812 may be used to withstand autoclave procedures. For example, pressure sensors 804, 808, 812 may be packaged for single use and / or for multiple use. For example, pressure sensors 804, 808, 812 may include any pressure sensor suitable for medical applications, such as for air monitors, pneumatic controllers, breathing machines, ventilators, spirometers, etc. In an example, pressure sensors 804, 808, 812 may include analog sensors and / or digital sensors. For example, pressure sensors 804, 808, 812 may include one or more silicon sensor chips. For example, one or more pressure sensors 804, 808, 812 may include associated support circuits, such as amplifiers, integrated A / D converters, EEPROM memory, digital signal processing circuits, and interface circuits, etc. In an example, one or more pressure sensors 804, 808, 812 may include an on-board pressure sensor, such as an on-board pressure sensor from Honeywell (TM) part number 785-HSCDRRN100MD4A3.

[0058] The measurements sensed by the sensors 802, 804, 806, 808, 810, 812 may be converted to digital representations via one or more analog-to-digital converters 814. The analog-to-digital converter 814 may, for example, convert an analog representation of the measurement of the sensor (such as voltage, current, etc.) into a digital representation (such as an 8-bit, 16-bit, 24-bit, 32-bit digital value). The analog-to-digital converter 814 may include any architecture and / or form factor suitable for inclusion in a medical device (such as device 800). For example, the analog-to-digital converter 814 may include a converter integrated with one or more sensors themselves. The analog-to-digital converter 814 may include a subcomponent of the processor 816. For example, the analog-to-digital converter 814 may include an independent electrical component. The analog-to-digital converter 814 may include a separate converter for each sensor, a shared converter, or a combination thereof.

[0059] The analog-to-digital converter 814 can convert analog information captured by one or more sensors 802, 804, 806, 808, 810, 812 into a digital format by sampling the signals received from the sensors at a specific sampling rate. The sampling rate can be any rate suitable for capturing the signal and / or signal characteristics of the patient's air exchange. For example, the sampling rate can be selected to be at least twice the highest relevant frequency of the sampled signal type. For example, the sampling rate can be selected as discussed herein. The digital signal from the analog-to-digital converter 814 can represent a time series of data for each of the sensors 802, 804, 806, 808, 810, 812 of the device 800. The captured data, such as one or more data time series, can be stored in the memory 818 and / or processed by the processor 816.

[0060] Processor 816 may include any device suitable for processing such data. For example, processor 816 may include any device suitable for processing such data, performing numerical operations on such data, storing data to memory 818, running a predictive model with the data as input, processing operations of device 800, etc. Processor 816 may include a general purpose processor, a microcontroller, an application specific integrated circuit (ASIC), etc. In an example, the processor may include, for example, an Arduino Uno microcontroller.

[0061] The memory 818 may include any component suitable for storing such digital data. For example, the memory 818 may include random access memory, read-only memory, volatile memory, and / or non-volatile memory. For example, the memory 818 may include solid-state memory, etc. Depending on the sensors 802, 804, 806, 808, 810, 812 and the processor 816, the size and selection of the memory 818 may be suitable for the required volume and storage speed.

[0062] The device 800 may include one or more auxiliary processors 820. For example, the auxiliary processor 820 may include any component, device, system, computation, and / or resource for providing processing in addition to the processing of the processor 816 and / or access to such components, devices, systems, and / or computational resources. The auxiliary processor 820 may include a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an application programming interface (API) to external processing resources (such as cloud and / or edge processing resources), etc. In an example, the auxiliary processor may be used to handle the computational requirements of developing and / or implementing a predictive model (such as the predictive model discussed herein).

[0063] The device 800 may include a user interface 822. For example, the user interface 822 may provide a user input mechanism (such as a button, a touch screen) and / or access to an external user interface device (such as a keyboard, a monitor, and a mouse). The user interface 822 may provide a user input mechanism to establish the start and / or end of a signal recording session. For example, the user interface 822 may provide the ability to input certain patient-related data. The user interface 822 may provide a user output mechanism (such as an indicator light, a display), access to an external user interface device, and the like. The output mechanism may be used to output all or part of the information captured by the sensors 802, 804, 806, 808, 810, 812. The output mechanism may be used to output a summary of the information captured by the sensors 802, 804, 806, 808, 810, 812. The output mechanism may be used to output the results of the prediction model processing the data captured by the sensors 802, 804, 806, 808, 810, 812. For example, the output mechanism may include a probability of a prolonged air leak based on data captured from one or more of the sensors 802, 804, 806, 808, 810, 812. For example, the display may be used to display information indicating the probability of a prolonged air leak. In an example, the displayed output information indicating the probability of a prolonged air leak may include a numerical value. In an example, the displayed output information indicating the probability of a prolonged air leak may include a qualitative risk level.

[0064] The device 800 may include a communication interface 824 to provide data exchange between the device 800 and one or more other components and / or networks. For example, the communication interface 824 may include a serial interface, a parallel interface, a universal serial bus (USB) interface, etc. For example, the communication interface 824 may include a network communication interface, such as an Ethernet interface, a WiFi interface, a cellular interface, a 5G interface, etc. For example, the communication interface 824 may provide access to a host computer for data recording capabilities. For example, the communication interface 824 may provide data exchange between the device 800 and a surgical computer (such as a surgical hub). For example, the communication interface 800 may provide data exchange with one or more edge and / or cloud computing resources. The communication interface 824 may receive information such as procedural information, intraoperative report information, etc.

[0065] In an example, the communication interface 824 can be used to update the programming of the device 800, including, for example, a predictive model used by the device 800. For example, the predictive model can be stored in the memory 818. The predictive model can be an updateable predictive model. For example, the communication interface 824 and the processor 816 can receive downloads of software, firmware, etc. For a predictive model that includes a neural network, for example, the communication interface 824 and the processor 816 can receive updated coefficients and / or an updated neural network architecture.

[0066] The device 800 can have a housing and connectors suitable for use in an operating room. For example, the device 800 can include a tube assembly and connectors suitable for surgery. The device 800 and / or the connector can be suitable for autoclave sterilization cycles. For example, the device 800 can be manufactured according to a procedure for providing durable medical equipment. The device 800 can be integrated into a piece of medical equipment commonly used in surgery (such as lung resection surgery). For example, the device 800 can be integrated into a ventilator. In an example, the device 800 can be manufactured so that it can be used in surgery to provide an intraoperative indication to the surgeon about the possibility of a prolonged air leak.

[0067] In an example, sensors 802, 804, 806, 808, 810, 812 can record various air exchange data. This information can be digitized at a suitable rate by an analog-to-digital converter 814 to capture changes within the breathing cycle, such as peak values, changes between peak values, amplitudes, etc. The signal can be stored in a memory 818 and processed by a processor 816 and / or an auxiliary processor 820 according to, for example, a predictive model. And a user interface 822 can display the output of the predictive model, which can include a probability associated with the likelihood of a long-term air leak.

[0068] In another example, sensors 802, 804, 806, 808, 810, 812 can record various air exchange data. This information can be digitized at a suitable rate by an analog-to-digital converter 814 to capture changes within the respiratory cycle, such as peak values, changes between peak values, amplitudes, etc. The signal can be stored in a memory 818, processed by a processor 816, and / or transmitted via a communication interface 824. The signal can be paired with one or more other data elements, including patient outcomes, such as patient outcomes regarding the presence of prolonged air leaks. This pairing of signals and outcomes can be used to develop a predictive model for a population.

[0069] Fig. 9 is a flow chart of an example process for collecting signals and / or for determining the likelihood of a prolonged air leak. In an example, the process may be performed by Figure 8 , the device 800 shown in is performed. At 900, first data is received. This data can indicate the air exchange between the surgical ventilator and the patient. For example, the first data can be received via, for example, a ventilator inlet flow sensor, a ventilator inlet pressure sensor, a ventilator outlet flow sensor and / or a ventilator outlet pressure sensor. The first data can include one or more time series. Each time series can be associated with a specific sensor and / or parameter. For example, such data can be sampled at a rate suitable for capturing the change of the air exchange signal between the surgical ventilator and the patient within a breathing cycle.

[0070] At 902, second data may be received. The second data may indicate air leaving the patient via a chest tube. For example, the second data may include data captured from a chest tube flow sensor, a chest tube pressure sensor, etc. The second data may include one or more time series of data. For example, the second data may include a time series associated with each chest tube sensor. For example, such data may be sampled at a rate suitable for capturing changes in air leaving the patient via a chest tube.

[0071] At 904, the first data and the second data may be time correlated. For example, the first data and the second data may be captured in a synchronized manner. For example, the time correlation of the first data and the second data may be performed by simultaneous and / or parallel sampling. For example, the first data and the second data may include a reference to a common clock for time correlation purposes. For example, the first data and the second data may be subject to software-based synchronization. The time correlation of the first data and the second data may be performed by any mechanism or algorithm suitable for capturing time series data so that various time series may be superimposed on a common timeline.

[0072] At 906, the first data and the second data may be processed via a prediction algorithm. The prediction algorithm may include a machine learning algorithm. In an example, the prediction algorithm may be provided by a computer program such as, for example, Figure 8 The prediction algorithm may be implemented and / or operated on a computing device other than the device that collects the first data and the second data. For example, the first data and the second data may be transmitted to another device via a communication interface for off-board processing purposes.

[0073] At 908, the output of the prediction algorithm can be output. The output information can indicate the likelihood of a prolonged air leak. For example, the output information can include probabilities associated with how similar the first data and the second data are to other first data and second data from other populations and their respective patient outcomes.

[0074] Fig.10 1 is a diagram of an exemplary device 1000 for collecting signals and for determining the likelihood of a prolonged air leak. The device 1000 may include a console unit 1002. The console unit 1002 may be configured to connect to a ventilator 1004 and / or a chest tube assembly 1006. The console unit 1002 may include one or more pressure sensor ports 1008. The console unit 1002 may include one or more air flow sensor ports 1010. For example, the console unit 1002 may include data logging hardware and / or software to capture mass air flow and pressure.

[0075] One or more pressure sensor ports 1008 may be connected to a portion of the chest tube assembly 1006 via surgical tube 1012 to measure chest tube pressure. One or more pressure sensor ports 1008 may be connected to an outflow port of the ventilator 1004 via surgical tube 1014 via a "Y" connection 1016 to measure outflow ventilator pressure. One or more pressure sensor ports 1008 may be connected to an inflow port of the ventilator 1004 via surgical tube 1018 via a "Y" connection 1020 to measure inflow ventilator pressure. Located between the "Y" connection 1020 and the ventilator inflow port may be an air flow sensor 1022. The air flow sensor 1022 may be connected to the air flow sensor port 1010 of the console unit 1002. The air flow sensor 1022 may be a digital air flow sensor.

[0076] The chest tube assembly 1006 may include a chest tube 1024, a chest tube flow sensor 1026, and a water seal drainage unit 1028. The chest tube flow sensor 1026 may be connected between the chest tube 1024 and the water seal drainage unit 1028. For example, the water treatment drainage unit may be connected to the vacuum port. The chest tube flow sensor 1026 may be connected to the air flow sensor port 1010 of the console unit 1002 for measuring the chest tube air flow. The air flow sensor 1024 may be a digital air flow sensor.

[0077] The device 1000 can be adapted to collect signal information from various sensors before the end of the surgical procedure. For example, the console unit 1002 can obtain flow information from the chest tube flow sensor 1026 and the ventilator air flow sensor 1022. The console unit 1002 can obtain pressure information from the chest tube assembly 1006, the ventilator inlet, and the ventilator outlet via one or more pressure sensor ports 108.

[0078] The console unit 1002 may record the data in an internal memory. The console unit 1002 may include a communication interface connection 1030, such as a USB interface for connecting the console unit 1002 to a computer. The console unit 1002 may transmit the data to another computing device via the communication interface connection 1030.

[0079] The console unit 1002 may include a user interface 1032. The user interface 1032 may include one or more buttons and / or a display on the console unit 1002. The console may provide information indicating output information from the predictive model via the user interface 1032. For example, the user interface 1032 may indicate to the surgeon the likelihood that the leak will become a long-term air leak based on, for example, the predictive model.

[0080] Fig.111 is a block diagram of an example device 1100 for collecting signals and for determining the likelihood of a prolonged air leak. For example, the device 1100 may be adapted to capture signal information of a population for the purpose of training a predictive model. For example, the device 1100 may be adapted to capture subsequent signal information, such as test or new patient signals, in order to consider the output of the predictive model. For example, the device 1100 may be used to capture subsequent signals to be input into the predictive model and provide a surgeon with a predicted patient outcome, such as the likelihood of the patient experiencing a prolonged air leak.

[0081] The device 1100 may include one or more sensors 1102, 1104. The device 1100 may include any sensor suitable for collecting signals indicating patient air exchange. For example, the device 800 may include any sensor suitable for collecting signals indicating air leaving the patient via a chest tube. For example, the device 1100 may include one or more sensors, such as a chest tube flow sensor 1102, a chest tube pressure sensor 1104, etc. Such sensors may be suitable for recording surgical quality data from the patient. For example, air flow data may be measured with a tolerance of + / - 0.4 liters / minute. For example, pressure measurements may be made with a tolerance of + / - 2.5 cm H20.

[0082] Flow sensor 1102 may include any sensor suitable for capturing the flow of air. For example, flow sensor 1102 may include a sensor or flow meter designed for medical applications. For example, flow sensor 1102 may be used to withstand autoclave procedures. For example, flow sensor 1102 may be packaged for single use and / or for multiple use. For example, flow sensor 1102 may be designed for medical ventilation or breathing applications. In an example, flow sensor 1102 may include an analog sensor and / or a digital sensor. For example, flow sensor 1102 may include one or more silicon sensor chips. For example, flow sensor 1102 may include associated support circuits, such as amplifiers, integrated A / D converters, EEPROM memories, digital signal processing circuits, and interface circuits. In an example, flow sensor 1102 may include an SFM3400 digital flow meter from Sensirion (TM).

[0083] Pressure sensor 1104 may include any sensor suitable for measuring air pressure. For example, pressure sensor 1104 may include any sensor suitable for measuring absolute air pressure, gauge pressure and / or differential pressure. Pressure sensor 1104 may include a sensor or pressure gauge designed for medical applications. For example, pressure sensor 1104 may be used to withstand high pressure sterilization procedures. For example, pressure sensor 1104 may be packaged for single use and / or for multiple use. For example, pressure sensor 1104 may include any pressure sensor suitable for medical applications, such as for air monitors, pneumatic controllers, breathing machines, ventilators, spirometers, etc. In an example, pressure sensor 1104 may include analog sensors and / or digital sensors. For example, pressure sensor 1104 may include one or more silicon sensor chips. For example, pressure sensor 1104 may include related support circuits, such as amplifiers, integrated A / D converters, EEPROM memories, digital signal processing circuits, and interface circuits, etc. In an example, the pressure sensor 1104 may include an on-board pressure sensor, such as an on-board pressure sensor from Honeywell (TM) part number 785-HSCDRRN100MD4A3.

[0084] The measurements sensed by the sensors 1102, 1104 can be converted into digital representations via one or more analog-to-digital converters 1106. The analog-to-digital converter 1106 can, for example, convert an analog representation of the measurement of the sensor (such as voltage, current, etc.) into a digital representation (such as an 8-bit, 16-bit, 24-bit, 32-bit digital value). The analog-to-digital converter 1106 may include any architecture and / or form factor suitable for inclusion in a medical device (such as device 1100). For example, the analog-to-digital converter 1106 may include a converter integrated with one or more sensors themselves. The analog-to-digital converter 1106 may include a subcomponent of the processor 1108. For example, the analog-to-digital converter 1106 may include an independent electrical component. The analog-to-digital converter 1106 may include a separate converter for each sensor, a shared converter, or a combination thereof.

[0085] The analog-to-digital converter 1106 can convert analog information captured by one or more sensors 1102, 1104 into a digital format by sampling the signals received from the sensors at a specific sampling rate. The sampling rate can be any rate suitable for capturing the signal and / or signal characteristics of the patient's air exchange. For example, the sampling rate can be selected to be at least twice the highest relevant frequency of the signal type being sampled. For example, the sampling rate can be selected as discussed herein. The digital signal from the analog-to-digital converter 1106 can represent a time series of data for each sensor 1102, 1104 of the device 1100. The captured data, such as one or more data time series, can be stored in the memory 1110 and / or processed by the processor 1108.

[0086] The processor 1108 may include any device suitable for processing such data. For example, the processor 1108 may include any device suitable for processing such data, performing numerical operations on such data, storing the data to the memory 1110, running a predictive model with the data as input, processing the operation of the device 1100, etc. The processor 1108 may include a general purpose processor, a microcontroller, an application specific integrated circuit (ASIC), etc. In an example, the processor may include, for example, an Arduino Uno microcontroller.

[0087] The memory 1110 may include any component suitable for storing such digital data. For example, the memory 1110 may include random access memory, read-only memory, volatile memory, and / or non-volatile memory. For example, the memory 1110 may include solid-state memory, etc. Depending on the sensors 1102, 1104 and the processor 1108, the size and selection of the memory 1110 may be suitable for the required volume and storage speed.

[0088] The device 1100 may include one or more auxiliary processors 1112. For example, the auxiliary processor 1112 may include any component, device, system, computation, and / or resource for providing processing in addition to the processing of the processor 1108 and / or access to such components, devices, systems, and / or computational resources. The auxiliary processor 1112 may include a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an application programming interface (API) to external processing resources (such as cloud and / or edge processing resources), etc. In an example, the auxiliary processor may be used to handle the computational requirements of developing and / or implementing a predictive model (such as the predictive model discussed herein).

[0089] The device 1100 may include a user interface 1114. For example, the user interface 1114 may provide a user input mechanism (such as a button, a touch screen) and / or access to an external user interface device (such as a keyboard, a monitor, and a mouse). The user interface 822 may provide a user input mechanism to establish the start and / or end of a signal recording session. For example, the user interface 1114 may provide the ability to input certain patient-related data. The user interface 1114 may provide a user output mechanism (such as an indicator light, a display), access to an external user interface device, and the like. The output mechanism may be used to output all or part of the information captured by the sensors 1102, 1104. The output mechanism may be used to output a summary of the information captured by the sensors 1102, 1104. The output mechanism may be used to output the results of the prediction model processing the data captured by the sensors 1102, 1104. For example, the output mechanism may include the possibility of a long-term air leak based on the data captured from one or more of the sensors 1102, 1104.

[0090] The device 1100 may include a communication interface 1116 to provide data exchange between the device 1100 and one or more other components and / or networks. For example, the communication interface 1116 may include a serial interface, a parallel interface, a universal serial bus (USB) interface, etc. For example, the communication interface 824 may include a network communication interface, such as an Ethernet interface, a WiFi interface, a cellular interface, a 5G interface, etc. For example, the communication interface 1116 may provide access to a host computer for data recording capabilities. For example, the communication interface 1116 may provide data exchange between the device 1100 and a surgical computer (such as a surgical hub). For example, the communication interface 1116 may provide data exchange with one or more edge and / or cloud computing resources. The communication interface 1116 may receive information such as procedural information, intraoperative report information, etc.

[0091] In an example, the communication interface 1116 may be used to update the programming of the device 1100, including, for example, a prediction model used by the device 1100. For example, the prediction model may be stored in the memory 1110. The prediction model may be an updateable prediction model. For example, the communication interface 1116 and the processor 1108 may receive a download of software, firmware, etc. For a prediction model that includes a neural network, for example, the communication interface 1116 and the processor 1108 may receive updated coefficients and / or an updated neural network architecture.

[0092] The device 1100 can have a housing and connectors suitable for use in an operating room. For example, the device 1100 can include a tube assembly and connectors suitable for surgery. The device 1100 and / or the connector can be suitable for autoclave sterilization cycles. For example, the device 1100 can be manufactured according to a procedure for providing durable medical equipment. The device 1100 can be integrated into a piece of medical equipment commonly used in surgery (such as lung resection surgery). For example, the device 1100 can be integrated into a ventilator. In an example, the device 1100 can be manufactured so that it can be used in surgery to provide an intraoperative indication to the surgeon about the possibility of a prolonged air leak.

[0093] The device 1100 may include a power management subsystem 1118. The power management subsystem 1118 may include an onboard power source, such as a battery. The power source and power management subsystem 1118 may enable the device 1100 to be a portable device. Such a portable device may be sent home with the patient to continuously and / or intermittently collect information via the sensors 1102, 1104, and transmit the information via a communication interface 1116 (such as, for example, a wireless cellular interface). The power management subsystem 1118 may provide power management for the device 1100, including scheduling measurements between periods of low-power "sleep" mode. The power management subsystem 1118 may enable the device 1100 to provide several days of accurate signal recordings in a portable setting, and provide these recordings to health care professionals.

[0094] In an example, sensors 1102, 1104 can record various air exchange data. This information can be digitized at a suitable rate by an analog-to-digital converter 1106 to capture changes within the breathing cycle, such as peak values, changes between peak values, amplitudes, etc. The signal can be stored in a memory 1110 and processed by a processor 1108 and / or an auxiliary processor 1112 according to, for example, a predictive model. And a user interface 1114 can display the output of the predictive model, which can include a probability associated with the likelihood of the formation of a long-term air leak.

[0095] In another example, sensors 1102, 1104 can record various air exchange data. This information can be digitized at a suitable rate by analog-to-digital converter 1106 to capture changes within the respiratory cycle, such as peak values, changes between peak values, amplitudes, etc. The signal can be stored in memory 1110, processed by processor 1108, and / or transmitted via communication interface 1116. The signal can be paired with one or more other data elements, including patient outcomes, such as patient outcomes regarding the presence of prolonged air leaks. This pairing of signals and outcomes can be used to develop predictive models for populations.

[0096] Fig.12is a flow chart of an example process for collecting signals and for determining the likelihood of a long-term air leak. For example, the process can be employed by device 1100.

[0097] At 1200, chest tube pressure can be measured. For example, chest tube pressure can be measured by a pressure sensor. At 1202, chest tube flow can be measured. For example, chest tube flow can be measured by a mass air flow sensor. Chest tube pressure and chest tube flow can be measured by corresponding analog sensors with analog-to-digital converters. They can be measured by digital sensors with integrated analog-to-digital converters.

[0098] At 1204, data can be sampled from the corresponding sensors at an appropriate rate for capturing signals associated with the occurrence of prolonged air leaks for the population. For example, the signals can be sampled at a rate at least twice the highest relevant frequency of the sensor and / or the signal being captured. For example, the rate can be suitable for capturing pressure and flow changes within a particular or individual breathing cycle.

[0099] At 1206, the signal can be processed according to a prediction algorithm. The prediction algorithm can include a machine learning algorithm trained based on chest tube pressure and chest tube flow and patient outcomes. And at 1208, information indicating the possibility of a prolonged air leak can be output to the user. For example, information indicating the possibility of a prolonged air leak can be displayed to the user. In an example, adjustments to postoperative care can be determined based on the information indicating the possibility of a prolonged air leak. In an example, adjustments to intraoperative care can be determined based on the information indicating the possibility of a prolonged air leak. In an example, a risk level for at least one surgical outcome can be determined based at least in part on the information indicating the possibility of a prolonged air leak. For example, the risk level can be compared to a predetermined threshold. The predetermined threshold can be predetermined based on historical data related to a particular surgery, clinical trials, expert opinions, etc.

[0100] Fig.13 is an illustration of an exemplary device 1300 for collecting signals and for determining the likelihood of a prolonged air leak. For example, the device 1300 can be used in conjunction with a chest tube and drainage system. For example, the device 1300 can be located, for example, between a chest tube and a chest tube drainage tube. The device can include a first connection 1302. The first connection 1302 can be connected to a chest tube. For example, the first connection 1302 can be connected to the chest tube via a quick connect fitting. The device can include a second connection 1304. The second connection 1304 can be connected to a chest tube drainage unit. The second connection 1304 can be connected to a chest tube drainage unit, for example, via a quick connect fitting.

[0101] The device 1300 may include a tube 1306 between the first connection 1302 and the first side of the filter 1308. The filter 1308 may include a second side opposite the first side. The device 1300 may include a main module 1310. And the tube 1312 may connect the input end of the main module 1310 to the second side of the filter 1308. The main module 1310 may have an output end. The output end may be connected to the second connection 1304.

[0102] The main module 1310 may include an integrated system for collecting signals and for determining the likelihood of a prolonged air leak. The main module 1310 may include a housing 1314. Inside the housing 1314 may be components such as: a wireless communication module, a pressure transducer, a flow meter, a memory, an EEPROM, a processor, and / or an integrated battery. The device 1300 may include, for example Fig.11 The main module 1310 may have one or more user interface elements 1316. For example, the user interface element 1316 may be an element such as a power switch, an indicator, a display, and / or an audible alarm.

[0103] In use, the device 1300 can be connected between a patient's chest tube and a corresponding drainage tube. And the device 1300 can be used, for example Fig.12 The process shown in . In an example, the device 1300 can be connected to the patient's chest tube and corresponding drainage tube before the patient's surgery is completed. The device 1300 can be operated by turning on the power switch. In an example, configuration information can be provided to the device 1300 through the wireless communication module. The device 1300 can collect pressure and flow signal information. In an example, the device 1300 may include an onboard prediction model for processing. In an example, the device 1300 can capture and / or receive information and send the information to an external computing device such as a surgical hub, an edge server, a cloud computing system, etc. via the wireless communication module. And the device 1300 can display the predicted probability of a long-term air leak via a user interface element based on the collected information.

[0104] Fig.14 1 is a block diagram of a system including an exemplary plurality of surgical computing devices 1400, 1402. The devices 1400, 1402 may be located in separate locations 1404, 1406. The locations 1404, 1406 may be physically and / or logically separated, such as in different operating rooms in the same hospital, in different hospitals in the same healthcare system, in different healthcare systems having a common data network, and the like.

[0105] The surgical computing devices 1400, 1402 may be used in an operating room. In an operating room, the surgical computing devices 1400, 1402 may coordinate and / or aggregate data and / or control of surgical equipment. For example, the surgical computing devices 1400, 1402 may coordinate and / or aggregate data associated with surgical equipment such as surgical visualization systems 1408, 1410, surgical robotic systems 1412, 1414, one or more smart instruments 1416, 1418, patient record systems 1420, 1422, air exchange monitoring devices 1424, 1426 (e.g., as disclosed herein), and the like. In an example, the surgical computing devices 1400, 1402 may coordinate and / or aggregate data associated with surgical equipment such as imaging devices, lighting sources, displays, intelligent surgical instruments, patient monitoring equipment, anesthesia machines operating tables, surgical microscopes, vital sign monitors, EKG machines, ultrasound machines, endoscopic equipment, electrosurgical equipment, surgical staplers, smoke evacuators, neurostimulators, central gas and suction controllers, etc. Such devices may generate and / or consume data feeds associated with a surgical procedure.

[0106] In an example, the surgical computing device 1400, 1402 may include one or more modules for use in a surgical procedure. For example, the surgical computing device 1400, 1402 may include a monitor, an imaging module, a generator module, a smoke evacuation module, a suction / irrigation module, a communication module, a processor module, a storage array, an operating room mapping module, etc. The combined generator module may include two or more of an ultrasonic energy generator component, a bipolar RF energy generator component, and a monopolar RF energy generator component housed in a single unit. Such a module may be included in a modular housing.

[0107] The modules and surgical equipment may generate and / or consume data feeds related to the surgical procedure. Such data feeds may represent information about the surgical procedure, such as patient medical record information, preoperative planning, intraoperative activities including equipment usage, surgical tasks, patient reactions, patient outcomes, etc. Also, incorporating surgical computing devices 1400, 1402 in a surgical setting may enable the collection of such relevant surgical data 1428. For example, surgical data 1428 may include data such as patient medical record data, non-surgical report data, surgical procedure data, etc.

[0108] Such devices may interact with surgical computing devices 1400, 1402. Surgical computing devices 1400, 1402 may include a surgical hub. For example, surgical computing devices 1400, 1402 and corresponding surgical data 1428 may include the following disclosures, which are incorporated herein by reference:

[0109] U.S. Patent Application Serial No. 15 / 640,656, filed on March 29, 2018, entitled “SURGICAL HUB COORDINATION OF CONTROL AND COMMUNICATION OF OPERATING ROOM DEVICES” (Attorney Docket No. END8499USNP2) (U.S. Patent Application Publication No. 2019-0201141 A1) U.S. Patent Application Serial No. 15 / 940,668, filed on March 29, 2018, entitled “AGGREGATION AND REPORTING OF SURGICAL HUBDATA” (Attorney Docket No. END9502USNP2) (U.S. Patent Application Publication No. 2019-0201115 A1) U.S. Patent Application Serial No. 16 / 209,385, filed on December 4, 2018, entitled “METHOD OF HUB COMMUNICATION, PROCESSING, STORAGE AND DISPLAY” (Attorney Docket No. END9495USNP) (U.S. Patent Application Publication No. US 2019-0200844 A1) U.S. Patent Application Serial No. 16 / 209,407, filed on December 4, 2018, entitled “METHOD OF ROBOTIC HUB COMMUNICATION, DETECTION, AND CONTROL” (Attorney Docket No. END8497USNP) (U.S. Patent Application Publication No. US 2019-0201137 A1) U.S. Patent Application Serial No. 16 / 209,403, filed December 4, 2018, entitled “METHOD OF CLOUD BASED DATA ANALYTICS FOR USEWITH THE HUB” (Attorney Docket No. END8496USNP) (U.S. Patent Application Publication No. US 2019-0206569 A1) U.S. Patent Application Serial No. 16 / 209,416, filed on December 4, 2018, entitled “METHOD OF HUB COMMUNICATION, PROCESSING, DISPLAY, AND CLOUD ANALYTICS” (Attorney Docket No. END8505USNP) (U.S. Patent Application Publication No. US 2019-0206562 A1) U.S. Patent Application Serial No. 15 / 940,671, filed on March 29, 2018, entitled “SURGICAL HUB SPATIAL AWARENESS TO DETERMINED EVICES IN OPERATING THEATER” (Attorney Docket No. END8502USNP) (U.S. Patent Application Publication No. US 2019-0201104 A1) U.S. Patent Application Serial No. 15 / 940,654, filed on March 29, 2018, entitled “SURGICAL HUB SPATIAL AWARENESS TO DETERMINED EVICES IN OPERATING THEATER” (Attorney Docket No. END8501USNP) (U.S. Patent Application Publication No. US 2019-0201140 A1) U.S. Patent Application Serial No. 16 / 209,478, filed on December 4, 2018, entitled “METHOD FOR SITUATIONAL AWARENESS FOR SURGICAL NETWORK OR SURGICAL NETWORK CONNECTED DEVICE CAPABLE OF ADJUSTING FUNCTIONBASED ON A SENSED SITUATION OR USAGE” (Attorney Docket No. END9015USNP1) (U.S. Patent Application Publication No. US 2019-0104919 A1) U.S. Patent Application Serial No. 16 / 209,490, filed December 4, 2018, entitled “METHOD FOR FACILITY DATA COLLECTION AND INTERPRETATION” (Attorney Docket No. END9017USNP1) (U.S. Patent Application Publication No. 2019-0206564 A1) U.S. Patent Application Serial No. 17 / 358,270, filed on June 25, 2021, entitled “METHOD OF HUB COMMUNICATION, PROCESSING, DISPLAY, AND CLOUD ANALYTICS” (Attorney Docket No. END8505USCNT2) (U.S. Patent Application Publication No. 2021-0315580 A1) U.S. Patent Application Serial No. 17 / 210,839, filed on March 24, 2021, entitled “SURGICAL NETWORK RECOMMENDATIONS FROM REALTIME ANALYSIS OF PROCEDURE VARIABLES AGAINST A BASELINE HIGHLIGHTING DIFFERENCES FROM THE OPTIMAL SOLUTION” (Attorney Docket No. END9018USCNT1) (U.S. Patent Application Publication No. 2021-0205020 A1) U.S. Patent Application Serial No. 17 / 217,163, filed on March 30, 2021, entitled “DATA STRIPPING METHOD TO INTERROGATE PATIENT RECORDS AND CREATE ANONYMIZED RECORD” (Attorney Docket No. END8500USCNT1) (U.S. Patent Application Publication No. 2021-0240852 A1) U.S. Patent Application Serial No. 17 / 363,944, filed on June 30, 2021, entitled “METHOD OF HUB COMMUNICATION” (Attorney Docket No. END8495USCNT2) (U.S. Patent Application Publication No. 2021-0322018A1) U.S. Patent Application Serial No. 16 / 182,278, filed on November 6, 2018, entitled “COMMUNICATION OF DATA WHERE A SURGICAL NETWORK IS USING CONTEXT OF THE DATA AND REQUIREMENTS OF A RECEIVING SYSTEM / USER TO INFLUENCE INCLUSION OR LINKAGE OF DATA AND METADATA TO ESTABLISHCONTINUITY” (Attorney Docket No. END9018USNP4) (U.S. Patent Application Publication No. 2019-0201130 A1) U.S. Patent Application Serial No. 16 / 182,260, filed on November 6, 2018, entitled “AUTOMATED DATA SCALING, ALIGNMENT, ANDORGANIZING BASED ON PREDEFINED PARAMETERS WITHIN SURGICAL NETWORKS” (Attorney Docket No. END9012USNP8) (U.S. Patent No. 11,056,244) U.S. Patent Application Serial No. 16 / 172,198, filed on October 26, 2018, entitled “METHOD OF HUB COMMUNICATION WITH SURGICALINSTRUMENT SYSTEMS” (Attorney Docket No. END8571USNP) (U.S. Patent Application Publication No. 2019-0125455 A1) U.S. Patent Application Serial No. 15 / 940,649, filed on March 29, 2018, entitled “DATA PAIRING TO INTERCONNECT A DEVICEMEASURED PARAMETER WITH AN OUTCOME” (Attorney Docket No. END8500USNP3) (U.S. Patent Application Publication No. 2019-0205567 A1) U.S. Patent Application Serial No. 16 / 931,712, filed on July 17, 2020, entitled “WIRELESS PAIRING OF A SURGICAL DEVICE WITH ANOTHER DEVICE WITHIN A STERILE SURGICAL FIELD BASED ON THE USAGE AND SITUATIONAL AWARENESS OF DEVICES” (Attorney Docket No. END9032USCNT1) (U.S. Patent No. 11,179,204) U.S. Patent Application Serial No. 16 / 182,256 (Attorney Docket No. END9016USNP2), entitled “ADJUSTMENT OF A SURGICAL DEVICE FUNCTION BASED ON SITUATIONAL AWARENESS,” filed on November 6, 2018 (U.S. Patent Application Publication No. 2019-0201127 A1).

[0110] In an example, such surgical data 1428 may be stored at a storage array at the surgical computing devices 1400, 1402. In an example, such surgical data 1428 may be aggregated from one or more surgical computing devices 1400, 1402 at a data repository 1430. Such surgical data 1428 may be used to develop a predictive model.

[0111] Fig.15 is a block diagram illustrating an example system for determining the likelihood of a prolonged air leak. Surgical data 1500 may be captured from one or more resection surgeries. For example, surgical data 1500 may be captured from a large number of resection surgeries. For example, data may be collected from at least one hundred resection surgeries. Surgical data 1500 may include one or more signals indicative of air exchange for a respective patient. For example, surgical data 1500 may include non-signal surgical data, such as data collected from one or more surgical computing devices (such as Fig.14 The surgical computing devices 1400 , 1402 ) shown collect and / or aggregate non-signal surgical data.

[0112] The surgical data 1500 may be prepared as data elements 1502. The data elements 1502 may include at least one or more parameters of a signal and / or signal characteristics, a corresponding patient outcome, and corresponding non-signal surgical data.

[0113] In an example, non-signal surgical data may include data such as patient medical record data, non-surgical report data, surgical data, and the like. For example, patient medical record data may include any data associated with a patient and / or his or her medical treatment. For example, patient record data may include identifying information such as date of birth, name, marital status, social security number, etc., patient number, patient identifier. Patient medical record data may include medical history, such as allergies, previous treatments, previous medical care, current and past diagnoses. Patient medical record data may include medication information, which is a record of medications currently being taken by the patient. Medication information may include prescription medications as well as over-the-counter medications such as herbal remedies, illegal substances, over the counter medication, and the like.

[0114] Patient medical record data may include information about demographic information, race, ethnicity, family history, etc. For example, family history may include family history related to the lungs. Patient medical record data may include information about treatment history. Treatment history may reflect information about treatments that a patient has undergone and / or their results. For example, treatment history information may include chief complaint, medical history, vital signs, physical examination notes, surgical history, obstetric history, allergies, family history, immunization history, habits, including diet, alcohol intake, exercise, drug abuse, smoking, developmental history, age, weight, gender, etc.

[0115] The patient's medical record data may include laboratory results. The laboratory results may include, for example, laboratory results related to cells, tissues, and / or body fluids. Imaging records (such as x-ray scans, CT scans, ultrasound) may be included in such laboratory results. The patient's medical record data may include a medical record. For example, a medical record may indicate specific information during a treatment process, during surgery, and / or during recovery from the surgery. For example, a medical record may include information related to bowel and bladder function, observations of the patient's mental and physical condition, sudden changes in physiology or behavior, food intake, and / or vital signs. For example, a medical record may include information related to air leak metrics, such as chest tube air quality metrics, etc. In an example, the patient's medical record data may include preoperative data and / or the results of one or more tests. For example, patient spirometry, diffusion capacity, myocardial stress test, echocardiogram, quantitative perfusion scan, imaging, specific signs of abnormal anatomical structures and lesion locations, such as the presence of chemotherapy, radiotherapy, smoking, COPD, etc.

[0116] Patient record data may include information related to postoperative information such as patient activity, breathing exercises, use of an incentive spirometer, chest tube drainage rate, total chest tube fluid drainage rate, whether the chest tube was operated under a vacuum or water seal, and use of a portable pleural vacuum device.

[0117] Intraoperative report data may include any information related to the performance of the surgical procedure. For example, such intraoperative report data may be automatically collected, for example, by a surgical hub. Intraoperative report data may be captured as notes from a surgeon and / or surgical nurse. Intraoperative report data may include information such as tumor location and adhesions. For example, intraoperative report data may include changes in patient positioning and / or patient position during surgery. The intraoperative report may include information about the general progress of the surgery. For example, the intraoperative report may include whether a lung resection surgery was performed as an open procedure or VATS. The intraoperative report may include whether a lung resection surgery was started as VATS but converted to open during surgery. In an example, the intraoperative report may include VATS port site placement.

[0118] The report may include information regarding wound closure, such as how the wound in the lung was closed, the size of the wound, the location, etc. Intraoperative report data may indicate the surgical stapler used, the type of staples, the number of staple firings, the pressure level, the firing timing, etc. To the extent that sutures were used, non-surgical report data may include the number of sutures, the size of the incisions that were sutured, the type of sutures, etc. Regarding procedural intraoperative report data regarding the presence of a bronchial stump, non-surgical report data may include whether the bronchial stump is covered by a vascularized tissue flap (such as a vascularized tissue flap from pericardial fat tissue), a pleural flap, an intercostal muscle flap, or a pedicled diaphragmatic flap, etc.

[0119] Intraoperative report data may include information related to tissue separation, such as electrocautery, bipolar current, ultra-precision harmonic scalpel use, etc. Report data may include information about such tools, including, for example, instrument type, instrument settings, and number and / or duration of use, etc. Reports may include information about bleeding, such as amount of bleeding, nature of bleeding, method of coagulation, etc. Intraoperative report data may include visual imaging to collect situational awareness around the positioning, location, and / or use of incision tools and sealants. Intraoperative report data may include lung pressure at a specific sealing time.

[0120] Other non-surgical report data may include whether air leaks were detected by inflation of the lungs underwater, the number and placement of chest tubes, for example, the presence and / or treatment of isolated leaks, the use of lasers (such as CO2 lasers) to cap air leaks.

[0121] The surgical data may include any information related to the planned surgery. This may include the nature of the surgery, whether it is trauma related, cancer related, etc. To the extent that the surgical procedure is a lung resection, the surgical data may include the scale of the surgery, such as whether the surgery includes a lung resection, a lobectomy, a sublobectomy, etc. For example, in a lobectomy, the surgical data may include information such as right upper lobe surgery performed open, right upper lobe surgery performed via VATS, open middle lobectomy, right middle lobe surgery performed via VATS, right middle lobe surgery performed open, right lower lobe surgery performed via VATS, right lower lobe surgery performed open, left upper lobe surgery performed via VATS, left upper lobe surgery performed open, etc.

[0122] Intraoperative data may include other information such as which instruments are being used, how long the surgery is ongoing, recorded data from any visual records, supplemental intraoperative imaging records, and / or data records from one or more wearable devices (such as those worn by the healthcare professional during surgery), etc.

[0123] The surgical data 1500 and / or data elements 1502 may be stored and processed by a surgical computer system 1504. The surgical computer system 1504 may include a processor 1506 and / or a data repository 1508. The surgical computer system 1504 may operate on a computer network 1510. The surgical computer system 1504 may be located within the Health Insurance Portability and Accountability Act (HIPAA) boundary or outside the HIPPA boundary. The collected data may populate the data repository 1508 so that a prediction model 1512 may be generated. The processor 1506 may be configured to generate a prediction model 1512 based on the surgical data 1500 and / or data elements 1502 stored in the data repository 1508, for example. The prediction model 1512 may be generated by traditional data analysis techniques, machine learning, etc. The prediction model 1512 may be generated by the processor 1506 according to any appropriate machine learning techniques, such as those discussed above.

[0124] In an example, a predictive model 1514 may be generated, validated, and ultimately deployed. For example, input 1514 may include a subsequent signal and corresponding subsequent non-signal surgical data. The input 1514 may be input to a predictive model 1512 to provide an output 1516. The output 1516 may include a probability of whether the subsequent signal and the subsequent non-signal surgical data are associated with the presence of a prolonged air leak. For example, including non-signal surgical data as an input to the model may improve the predictive ability of the model for a population.

[0125] Fig.16 is a flow chart of an example process for collecting signals and for determining the likelihood of a prolonged air leak.

[0126] At 1600, during surgery, first data may be received. The first data may represent intraoperative air exchange of a patient in a given breathing cycle. For example, the first data may be received by, for example Figure 8 and / or Fig.11 The first data may include a signal.

[0127] At 1602, also during the procedure, second data can be received. The second data can represent surgical parameters other than surgical parameters related to air exchange. For example, the second data can include non-signal surgical data. For example, the second data can include data such as patient medical record data, non-surgical report data, surgical operation data, etc. Here, the second data related to the procedure can be used to further enhance the effectiveness of the prediction algorithm.

[0128] At 1604, the first data and the second data may be processed. For example, the first data and the second data may be processed via a predictive algorithm. And, at 1606, during surgery, information indicating the possibility of a prolonged air leak may be output to a surgeon. For example, information indicating the possibility of a prolonged air leak may be displayed to a user. In an example, adjustments to postoperative care may be determined based on the information indicating the possibility of a prolonged air leak. In an example, adjustments to intraoperative care may be determined based on the information indicating the possibility of a prolonged air leak. In an example, a risk level for at least one surgical outcome may be determined based at least in part on the information indicating the possibility of a prolonged air leak. For example, the risk level may be compared to a predetermined threshold. The predetermined threshold may be predetermined based on historical data, clinical trials, expert opinions, and the like related to a particular surgery.

[0129] In an example, a treatment recommendation may be generated and / or output. For example, the treatment recommendation may be based at least in part on information indicating the likelihood of a prolonged air leak. The treatment recommendation may be intended to improve surgical outcomes. For example, the treatment recommendation may be based on historical data related to a particular surgery, clinical trials, expert opinions, etc.

Claims

1. A device comprising: A processor, the processor being configured to: receiving first data during surgery, the first data representing intraoperative air exchange in a breathing cycle of a patient, receiving second data during the procedure, the second data representing a parameter of the procedure other than a parameter of the procedure related to air exchange, and During the surgery, information indicating a possibility of a long-term air leak is output based on the first data and the second data.

2. The device according to claim 1, wherein: The first data includes any one of the following: ventilator inlet flow, ventilator inlet pressure, ventilator output pressure, chest tube flow or chest tube pressure.

3. The device according to claim 1, wherein: The second data includes patient medical record data.

4. The device according to claim 1, wherein: The second data includes intraoperative report data.

5. The device according to claim 1, wherein: The second data includes surgical data associated with the surgery.

6. The device according to claim 5, wherein: The surgical data includes information characterizing a type of lung resection performed during the surgery.

7. The device according to claim 1, wherein: The information indicating the possibility of a long-term air leak is based on a machine learning model, and the first data and the second data are input to the machine learning model.

8. The device according to claim 1, wherein: The processor is configured to receive an updated machine learning model from a cloud resource, and wherein the information indicating the possibility of a long-term air leak is based on the updated machine learning model, and the first data and the second data are input to the updated machine learning model.

9. A method comprising: receiving first data during surgery, the first data representing intraoperative air exchange in a breathing cycle of a patient, receiving second data during the procedure, the second data representing a parameter of the procedure other than a parameter of the procedure related to air exchange, and During the surgery, information indicating a possibility of a long-term air leak is output based on the first data and the second data.

10. The method according to claim 9, wherein: The first data includes any one of the following: ventilator inlet flow, ventilator inlet pressure, ventilator output pressure, chest tube flow or chest tube pressure.

11. The method according to claim 9, wherein: The second data includes patient medical record data.

12. The method according to claim 9, wherein: The second data includes intraoperative report data.

13. The method according to claim 9, wherein: The second data includes surgical data associated with the surgery.

14. The method according to claim 13, wherein: The surgical data includes information characterizing a type of lung resection performed during the surgery.

15. The method according to claim 9, wherein: The information indicating the possibility of a long-term air leak is based on a machine learning model, and the first data and the second data are input to the machine learning model.

16. The method of claim 9, further comprising receiving an updated machine learning model from a cloud resource, wherein: The information indicating the possibility of a long-term air leak is based on the updated machine learning model, and the first data and the second data are input to the updated machine learning model.

17. The method of claim 9, further comprising displaying the information indicating the likelihood of a prolonged air leak.

18. The method of claim 9, further comprising determining adjustments to post-operative care based on the information indicating the likelihood of a prolonged air leak.

19. The method of claim 9, further comprising determining adjustments to intraoperative care based on the information indicating the likelihood of a prolonged air leak.

20. The method of claim 9, further comprising determining a risk level for at least one surgical outcome based at least in part on the information indicative of a likelihood of a prolonged air leak.

21. The method of claim 20, further comprising comparing the risk level to a predetermined threshold.

22. A device comprising: A computer, the computer being configured to: receiving first data representing a plurality of intraoperative air exchanges corresponding to individual patients in the population, receiving second data representing respective patient outcomes corresponding to individual patients in the population, receiving third data representing respective surgical parameters, other air exchanges, and outcomes corresponding to individual patients in the population, updating a model of long-term air leakage probability based on the first data, the second data, and the third data, and The model is sent to a surgery hub, wherein the surgery hub is configured to: receiving fourth data during surgery, the fourth data representing intraoperative air exchange in a breathing cycle of the patient, receiving fifth data during the procedure, the second data representing a surgical parameter other than a surgical parameter related to air exchange, During the surgery, information indicating the possibility of a long-term air leak is output based on the fourth data, the fifth data, and the model.

23. The apparatus of claim 22, wherein: The third data includes patient medical record data.

24. The apparatus of claim 22, wherein: The third data includes intraoperative report data.

25. The apparatus of claim 22, wherein: The third data includes operation data corresponding to a corresponding surgical operation of a corresponding intraoperative air exchange.

26. The apparatus of claim 22, wherein: The surgical hub is further configured to generate a treatment recommendation based at least in part on the information indicating a likelihood of a prolonged air leak, wherein the treatment recommendation is intended to improve a surgical outcome.

Images