Systems and methods for validating hemodynamic data

Abstract

A system for determining catheter placement in a patient includes a hemodynamic sensor and a display. The hemodynamic sensor generates a signal representative of a right ventricular pressure waveform of the patient. The catheter is connected to the hemodynamic sensor. The system further includes one or more processors and a computer-readable memory encoded with instructions that, when executed by the one or more processors, cause the system to receive the signal representative of the right ventricular pressure waveform and extract features from the right ventricular pressure waveform. The instructions further cause the system to determine that the catheter is correctly placed if the value of the feature is within one or more value ranges of the feature, or to compare the feature to one or more value ranges of the feature to determine that the catheter is not correctly placed in the patient or is not properly connected to the system, or that a signal quality problem is occurring.

Description

[Technical Field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of U.S. Provisional Application No. 63 / 492,159, filed March 24, 2023, entitled "SYSTEMS AND METHODS FOR VALIDATING HEMODYNAMIC DATA," the disclosure of which is incorporated herein by reference in its entirety. This application also claims the benefit of U.S. Provisional Application No. 63 / 492,176, filed March 24, 2023, entitled "SYSTEMS AND METHODS FOR DETERMINING FLOW AND CARDIAC OUTPUT," the disclosure of which is incorporated herein by reference in its entirety. This application also claims the benefit of U.S. Provisional Application No. 63 / 492,173, filed March 24, 2023, entitled "SYSTEMS AND METHODS FOR DETERMINING FILTERED CARDIAC OUTPUT," the disclosure of which is incorporated herein by reference in its entirety. [Background technology]

[0002] The present disclosure relates generally to hemodynamic monitoring, and particularly to determining blood flow and cardiac output in a patient using monitored hemodynamic data.

[0003] Monitoring patient hemodynamic variables enables improved patient care. Cardiac output, the volume of blood pumped by the heart per minute, is an important hemodynamic variable. Cardiac output is monitored by clinicians to make diagnoses and provide interventions in critically ill patients. However, the accuracy of cardiac output monitoring often decreases at high cardiac output levels. Furthermore, delays in cardiac output measurement can affect the accuracy of cardiac output information. The equipment required to monitor cardiac output is often expensive, difficult to obtain, or too invasive. Summary of the Invention [Means for solving the problem]

[0004] In one example, a system for determining catheter placement in a patient includes a first hemodynamic sensor and a display. The first hemodynamic sensor continuously generates a first hemodynamic sensor signal representative of the patient's right ventricular pressure waveform. The catheter is connected to the first hemodynamic sensor. The system further includes one or more processors and a computer-readable memory encoded with instructions that, when executed by the one or more processors, cause the system to receive the first hemodynamic sensor signal representative of the patient's right ventricular pressure waveform and extract features from the patient's right ventricular pressure waveform. The instructions further cause the system to compare the features extracted from the patient's right ventricular pressure waveform with one or more specified value ranges of the features to determine that the catheter is correctly placed if the value of the features extracted from the patient's right ventricular pressure waveform is within one or more specified value ranges, or to determine that the catheter is not correctly placed in the patient, is not properly connected to the system, or is experiencing a signal quality problem if the value of the features extracted from the patient's right ventricular pressure waveform is not within the one or more specified value ranges. The instructions further cause the system to output to a display or module an indication of whether the catheter is properly positioned within the patient based on whether the value of the feature extracted from the patient's right ventricular pressure waveform is within one or more specified value ranges.

[0005] In another example, a method for determining catheter placement in a patient includes receiving, by a hemodynamic monitoring system, sensed hemodynamic data representing the patient's right ventricular pressure waveform, and performing, by the hemodynamic monitoring system, waveform analysis of the hemodynamic data to extract features from the patient's right ventricular pressure waveform. The method further includes determining, by the hemodynamic monitoring system, whether values ​​of the features extracted from the patient's right ventricular pressure waveform are within the one or more specified value ranges, or whether values ​​of the features extracted from the patient's right ventricular pressure waveform are not within the one or more specified value ranges, by comparing the features extracted from the patient's right ventricular pressure waveform with one or more specified value ranges for the features.

[0006] In another example, a system for determining data quality from a catheter in a patient includes a first hemodynamic sensor and a display. The first hemodynamic sensor continuously generates a first hemodynamic sensor signal representative of the patient's right ventricular pressure waveform. The catheter is connected to the first hemodynamic sensor. The system further includes one or more processors and a computer-readable memory encoded with instructions that, when executed by the one or more processors, cause the system to receive the first hemodynamic sensor signal representative of the patient's right ventricular pressure waveform and extract features from the patient's right ventricular pressure waveform. The instructions further cause the system to compare the features extracted from the patient's right ventricular pressure waveform with one or more specified value ranges of the features to determine that the data from the catheter is of high quality if the values ​​of the features extracted from the patient's right ventricular pressure waveform are within one or more specified value ranges, or to determine that the data from the catheter is not of high quality if the values ​​of the features extracted from the patient's right ventricular pressure waveform are not within the one or more specified value ranges. The instructions further cause the system to output to a display or module an indication of whether the data from the catheter is of high quality based on whether the values ​​of the features extracted from the patient's right ventricular pressure waveform are within one or more specified value ranges. [Brief explanation of the drawings]

[0007] [Figure 1] FIG. 1 is a schematic block diagram illustrating an exemplary hemodynamic monitoring system for determining a patient's cardiac output and blood flow based on hemodynamic data. [Figure 2] FIG. 1 is a perspective view of an exemplary hemodynamic monitor that analyzes right ventricular and pulmonary artery pressure waveforms to provide cardiac output and blood flow of a patient. [Figure 3] 1 is a perspective view of an exemplary catheter that can be inserted into a patient and connected to one or more hemodynamic sensors to provide hemodynamic data to a hemodynamic monitor. [Figure 4] FIG. 1 is a perspective view of an exemplary minimally invasive pressure sensor that may be attached to a patient to sense hemodynamic data representative of the patient's right ventricular pressure or pulmonary artery pressure. [Figure 5] 1 is a perspective view of an exemplary oximetry module for receiving oximetry data from a catheter inserted within a patient; FIG. [Figure 6] FIG. 1 is a schematic block diagram showing inputs and outputs for modules of a hemodynamic monitoring system. [Figure 7] 1 is a graph showing an exemplary right ventricular pressure waveform trace including exemplary indices of blood flow and cardiac output. [Figure 8] 1 is a graph showing an exemplary pulmonary artery pressure waveform trace including exemplary indices of blood flow and cardiac output. [Figure 9] FIG. 2 is a schematic block diagram illustrating a verification module shown in FIG. [Figure 10] 1 is a graph showing an exemplary right ventricular pressure waveform tracing including indicators of catheter placement and data quality. [Figure 11] 1 is a graph showing an exemplary right ventricular pressure waveform trace and an exemplary pulmonary artery pressure waveform trace, including exemplary indicators of catheter placement and data quality. [Figure 12] FIG. 2 is a schematic block diagram of the flow module shown in FIG. [Figure 13]It is a schematic block diagram showing the conversion of a right ventricular pressure waveform into a flow waveform processed through a flow module. [Figure 14] It is a schematic block diagram showing an autoencoder model of a flow module. [Figure 15] It is a schematic diagram showing a filter of an autoencoder model. [Figure 16] It is a flowchart showing an exemplary process for training an autoencoder model for predicting human blood flow based on a right ventricular pressure waveform. [Figure 17] It is a schematic block diagram showing the COAE module shown in FIG. 1. [Figure 18] It is a schematic block diagram showing the COLR module shown in FIG. 1. [Figure 19] It is a schematic block diagram showing reference features and current features used as inputs to a regression model of a COLR module. [Figure 20A] It is a schematic block diagram showing the COfiltered module shown in FIG. 1, including a first example of a filter submodule. [Figure 20B] It is a schematic block diagram showing the COfiltered module shown in FIG. 1, including a second example of a filter submodule. [Figure 21A] It is a graph showing exemplary predicted values of a COfiltered module. [Figure 21B] It is a graph showing exemplary measurements over time of a COfiltered module. [Figure 21C] It is a graph showing exemplary filtered values over time from a COfiltered module. [Figure 22] It is a graph comparing exemplary cardiac output values from a COAE module, a COLR module, and a COfiltered module with exemplary intermittent cardiac output values over time of a patient in a pre-bypass state, a post-bypass state, and an ICU state. [Figure 23]FIG. 2 is a schematic block diagram illustrating the inputs and outputs of each module of an exemplary implementation of the hemodynamic monitoring system of FIG. 1. DETAILED DESCRIPTION OF THE INVENTION

[0008] As described herein, a hemodynamic monitoring system implements a series of modules that generate continuous blood flow and cardiac output measurements of a patient. The hemodynamic monitoring system utilizes a catheter and hemodynamic sensors to generate a right ventricular pressure ("RVP") waveform and, optionally, a pulmonary artery pressure ("PAP") waveform for input into the hemodynamic monitoring system's modules to generate the patient's blood flow and cardiac output during a procedure in, for example, an operating room (OR), an intensive care unit (ICU), or other patient care environment. Medical personnel can use the blood flow and cardiac output information to improve patient care.

[0009] Hemodynamic monitoring system (Figures 1 to 8) FIGURE 1 is a schematic block diagram of a hemodynamic monitoring system 10 that determines a patient's cardiac output and blood flow based on hemodynamic data. FIGURE 1 shows hemodynamic monitoring system 10 including hemodynamic monitor 12 and hemodynamic sensors 14 (including hemodynamic sensors 14A, 14B, 14C, and 14D). Hemodynamic monitor 12 includes a system processor 20, system memory 22, display 24, analog-to-digital converter (ADC) 26, and digital-to-analog converter (DAC) 28. System memory 22 includes an RVP feature module 32, a PAP feature module 34, a verification module 36, a flow module 38, and a CO AE Module 40 and CO LR Module 42 and CO filtered The flow and cardiac output software code 30 includes a flow and cardiac output module 44. The display 24 includes a user interface 46 that includes controls 48 and sensory alarms 50. Figure 1 also shows the patient 16 and healthcare worker 18.

[0010] As shown in Figure 1, hemodynamic monitoring system 10 includes hemodynamic monitor 12 and hemodynamic sensors 14 (including hemodynamic sensors 14A, 14B, 14C, and 14D). Hemodynamic monitoring system 10 can be implemented in a patient care environment, such as an ICU, OR, or other patient care environment, to monitor the hemodynamic status of the patient. As shown in Figure 1, the patient care environment can include a patient 16 and a medical professional 18 trained to utilize hemodynamic monitoring system 10.

[0011] As described below with respect to FIG. 2 , hemodynamic monitor 12 can be an integrated hardware unit including system processor 20, system memory 22, display 24, ADC 26, and DAC 28. In other examples, one or more components and / or described functions of hemodynamic monitor 12 can be distributed across multiple hardware units. For example, in some examples, display 24 can be a separate display device located remotely from and operably coupled to hemodynamic monitor 12. Similarly, at least a portion of the data processing within hemodynamic monitoring system 10 can be performed via a smart cable connected between a catheter (e.g., catheter 54 shown in FIG. 5 ) or sensor and hemodynamic monitor 12. While generally illustrated and described as an integrated hardware unit in the example of FIG. 1 , it should be understood that hemodynamic monitor 12 can include any combination of devices and components electrically, communicatively, or otherwise operably connected to perform the functions attributed to hemodynamic monitor 12 herein.

[0012] 1, the system memory 22 stores flow and cardiac output software code 30. The flow and cardiac output code 30 includes an RVP feature module 32, a PAP feature module 34, a validation module 36, a flow module 38, and a CO AE Module 40 and CO LR Module 42 and CO filteredand a module 44. Display 24 provides a user interface 46 including control elements 48 that enable user interaction with hemodynamic monitor 12 and / or other components of hemodynamic monitoring system 10. As shown in FIG. 1, user interface 46 also provides sensory alarms 50 that provide alerts to medical personnel based on the flow rate and / or cardiac output of patient 16, as described further below.

[0013] Hemodynamic sensor 14 can be attached to patient 16 to sense hemodynamic data representing the patient's 16 RVP waveform, PAP waveform, blood oxygen saturation ("SvO2"), or cardiac output, or any combination of these hemodynamic data. Hemodynamic sensor 14 is operatively connected (e.g., electrically and / or communicatively connected via a wired or wireless connection, or both) to hemodynamic monitor 12 to provide the sensed hemodynamic data to hemodynamic monitor 12. In some examples, hemodynamic sensor 14 provides the patient's 16 hemodynamic data to hemodynamic monitor 12 as an analog signal, which is converted by ADC 26 into digital hemodynamic data representing the RVP and / or PAP waveforms. In other examples, hemodynamic sensor 14 can provide the sensed hemodynamic data to hemodynamic monitor 12 in digital form, in which case hemodynamic monitor 12 may not include or utilize ADC 26. In yet another example, hemodynamic sensor 14 may provide hemodynamic data of patient 16 as an analog signal to hemodynamic monitor 12, which analyzes the hemodynamic data in its analog form.

[0014] The hemodynamic sensor 14 may include one or more non-invasive, minimally invasive, or invasive sensors attached to the patient 16. For example, the hemodynamic sensor 14 may take the form of an invasive hemodynamic sensor 14A, such as a second pressure transducer 14A providing RVP waveform data sensed at a right ventricular port 68C located within the right ventricle of the patient's heart (shown in FIG. 4). The hemodynamic sensor 14 may take the form of an invasive hemodynamic sensor 14B, such as an oximetry module 14B providing blood oxygen saturation data within the pulmonary artery based on optical pulses emitted from module 14B into the pulmonary artery, reflected, returned, and received by module 14B via the optical connector 62 of the catheter 54 (shown in FIG. 5). Additionally, the hemodynamic sensor 14 may take the form of a non-invasive hemodynamic sensor. In some examples, hemodynamic sensor 14 may be non-invasively attached to an extremity of patient 16, such as the forehead, wrist, arm, finger, ankle, toe, or other extremity of patient 16. Hemodynamic sensor 14 may also take the form of other invasive, minimally invasive, or non-invasive sensors.

[0015] In certain examples, hemodynamic sensor 14 can be configured to sense the patient's 16 RVP, PAP, or both right ventricular pressure and pulmonary artery pressure. In some examples, hemodynamic sensor 14 can also be used to sense the patient's cardiac output, blood oxygen saturation in the pulmonary artery, or both cardiac output and blood oxygen saturation, in addition to the right ventricular pressure waveform and the pulmonary artery pressure waveform. For example, one or more hemodynamic sensors 14 can be attached to the patient 16 via a radial artery catheter inserted in the patient's 16 arm. In other examples, hemodynamic sensor 14 can be attached to the patient 16 via a femoral artery catheter inserted in the patient's 16 leg. Such techniques can similarly enable multiple hemodynamic sensors 14 to provide substantially continuous beat-to-beat monitoring of RVP and PAP, as well as monitoring of the patient's 16 cardiac output and blood oxygen saturation, or any combination of these hemodynamic data, over extended periods of time, such as minutes or hours.

[0016] The system processor 20 includes an RVP feature module 32, a PAP feature module 34, a validation module 36, a flow module 38, and a CO2 analyzer module 39 that utilize the RVP waveform and, optionally, the PAP waveform to generate blood flow and / or cardiac output measurements. AE Module 40 and CO LR Module 42 and CO filtered and executes flow and cardiac output software code 30 that implements module 44. Examples of system processor 20 may include any one or more of a microprocessor, a controller, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), or other equivalent discrete or integrated logic circuitry.

[0017] The system memory 22 can be configured to store information within the hemodynamic monitor 12 during operation. The system memory 22, in some examples, is described as a computer-readable storage medium. In some examples, the computer-readable medium can include non-transitory media. The term "non-transitory" can indicate that the storage medium is not embodied in a carrier wave or propagated signal. In certain examples, the non-transitory storage medium can store data that may change over time (e.g., in RAM or cache). The system memory 22 can include volatile and non-volatile computer-readable memory. Examples of volatile memory can include random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), and other forms of volatile memory. Examples of non-volatile memory can include, for example, a magnetic hard disk, an optical disk, flash memory, or forms of electrically programmable memory (EPROM) or electrically erasable programmable memory (EEPROM).

[0018] Display 24 may be a liquid crystal display (LCD), light-emitting diode (LED) display, organic light-emitting diode (OLED) display, or other display device suitable for providing information in graphical form to a user. User interface 46 may include graphical and / or physical control elements that enable user input to interact with hemodynamic monitor 12 and / or other components of hemodynamic monitoring system 10. In some examples, user interface 46 may take the form of a graphical user interface (GUI) that presents graphical control elements presented on a touch-sensitive and / or presence-sensitive display screen of display 24, for example. In such examples, user input may be received in the form of gesture input, such as touch gestures, scrolling gestures, zoom gestures, or other gesture input. In certain examples, user interface 46 may take the form of and / or include physical control elements, such as physical buttons, keys, knobs, or other physical control elements configured to receive user input to interact with components of hemodynamic monitoring system 10.

[0019] An example of a hemodynamic monitor 12 is shown in Figure 2. An example of a catheter is shown in Figure 3. An example of a minimally invasive pressure sensor is shown in Figure 4. An example of an oxygen saturation measurement module is shown in Figure 5.

[0020] FIG. 2 is a perspective view of a hemodynamic monitor 12 that analyzes RVP and PAP waveforms to provide cardiac output and blood flow of a patient 16. As shown in FIG. 2, the hemodynamic monitor 12, in the example of FIG. 1, includes a display 24 that presents a graphical user interface including control elements 48 (e.g., graphical control elements) that enable user interaction with the hemodynamic monitor 12. The hemodynamic monitor 12 may also include multiple input and / or output (I / O) connectors 52 configured for wired connection (e.g., electrical and / or communicative connection) with one or more peripheral components, such as one or more hemodynamic sensors 14. While the example of FIG. 2 shows five separate I / O connectors 52, it should be understood that in other examples, the hemodynamic monitor 12 may include fewer than five I / O connectors 52 or more than five I / O connectors 52. In still other examples, the hemodynamic monitor 12 may not include I / O connectors 52 and may instead communicate wirelessly with various peripheral devices.

[0021] 1 , hemodynamic monitor 12 includes one or more system processors 20 and a computer-readable system memory 22 that stores flow and cardiac output software code 30 executable to generate continuous blood flow and cardiac output measurements. Hemodynamic monitor 12 can receive sensed hemodynamic data representing RVP and PAP waveforms, such as via one or more hemodynamic sensors 14 connected to hemodynamic monitor 12 via I / O connector 52. Hemodynamic monitor 12 executes flow and cardiac output software code 30 to obtain blood flow and cardiac output measurements using the received hemodynamic data and a plurality of profiling parameters (e.g., input features), as described further below.

[0022] 1, hemodynamic monitor 12 can present a graphical user interface on display 24. Display 24 can be an LCD, LED display, OLED display, or other display device suitable for providing information in a graphical format to a user. In some examples, such as the example of FIG. 2, display 24 can be a touch-sensitive and / or presence-sensitive display device configured to receive user input in the form of gesture input, such as a touch gesture, a scroll gesture, a zoom gesture, a swipe gesture, or other gesture input.

[0023] Hemodynamic monitor 12 receives hemodynamic data from patient 16 via one or more hemodynamic sensors 14A, 14B, 14C, and 14D (collectively hemodynamic sensors 14). In response to receiving the hemodynamic data of patient 16, hemodynamic monitor 12 executes flow and cardiac output software code 30 to determine blood flow and / or cardiac output and display the blood flow and / or cardiac output on display 24. Hemodynamic monitor 12 can also activate a sensory alarm, such as an audible, tactile, or other sensory alarm (e.g., sensory alarm 50 shown in FIG. 1 ), in response to the blood flow and / or cardiac output measurements. Thus, hemodynamic monitor 12 can provide an alarm to alert medical personnel regarding the blood flow or cardiac output measurements.

[0024] FIG. 3 is a perspective view of a catheter 54 that can be inserted into a patient 16 and connected to one or more hemodynamic sensors 14 to provide hemodynamic data to the hemodynamic monitor 12. For example, the catheter 54 can be connected to one or more pressure-sensing hemodynamic sensors 14A to detect the patient's 16 RVP, PAP, or both right ventricular pressure and pulmonary artery pressure. The catheter 54 can also interface with an oxygen saturation measurement module 14B to sense the patient's mixed venous oxygen saturation. Protected by a sheath 56, the catheter 54 includes multiple lumens 58 that allow a fluid connector 60, an optical connector 62, a thermistor connector 64, and a thermal filament connector 66 to be connected to one of ports 68, an implantable hemodynamic sensor 14C (e.g., a thermistor), or an implantable hemodynamic sensor 14D (e.g., a thermal filament). To facilitate insertion of the catheter 54 into the patient 16 or for certain hemodynamic measurements, the catheter 54 includes a balloon 70 located at a tip 72 of the catheter 54.

[0025] As shown in FIG. 3, the catheter 54 includes a distal port connector 60A that communicates with a port 68A at the tip 72. A proximal infusion connector 60B communicates with a proximal port 68B located approximately 30 cm from the tip 72 and can be used to administer fluids and medications into the patient's heart. A right ventricular pacing connector 60C communicates with a right ventricular port 68C, which can be located approximately 19 cm from the tip 72 or approximately 12–13 cm from the tip 72. The connector 60C can be used to sense RVP. A thermistor connector 64 electrically connects to a hemodynamic sensor 14C (e.g., a thermistor) located near the tip 72 of the catheter 54 to measure core blood temperature in the pulmonary artery. In some examples of the catheter 54, a thermal filament connector 66 electrically connects to a hemodynamic sensor 14D (e.g., a thermal filament) embedded within the catheter 54 located in the patient's right ventricle. In some examples, catheter 54 does not include a thermal filament or a corresponding thermal filament connector 66. Balloon connector 60D communicates with balloon 70 and can be used to inflate and deflate balloon 70 using syringe 74.

[0026] For example, after insertion into the patient 16 via the introducer, the distal port connector 60A and the right ventricular paging connector 60C can be connected to separate pressure transducer sensors 14A. The first pressure transducer 14A provides the hemodynamic monitor 12 with PAP waveform data sensed at a distal port 68A located in the pulmonary artery of the patient's heart, and the second pressure transducer 14A provides the hemodynamic monitor 12 with RVP waveform data sensed at a right ventricular port 68C located in the right ventricle of the patient's heart. Blood oxygen saturation data in the pulmonary artery can be provided by the oximetry module 14B based on light pulses emitted from the oximetry module 14B into the pulmonary artery and reflected light received by the oximetry module 14B via the optical connector 62 of the catheter 54. Additionally, utilizing the hot filament connector 66, the thermistor connector 64, and associated wiring, the hemodynamic monitor 12 can receive cardiac output data from the patient 16, for example, using thermodilution techniques. Cardiac output measured via the thermal filament and corresponding thermal filament connector 66 can be considered continuous cardiac output ("CCO"). If the catheter 54 does not include a thermal filament, cardiac output can be determined using a thermistor after injecting a liquid bolus (or set of boluses) of known volume and temperature through the proximal injection port 68B using thermodilution techniques. Cardiac output measured via the thermistor and corresponding thermistor connector 64 after fluid bolus injection can be considered intermittent cardiac output ("iCO"). Intermittent cardiac output iCO measurements can be obtained as frequently as every few minutes, hours, or even longer, depending on the level of monitoring required by the patient. For example, a clinician may administer a bolus set of three to four fluid boluses, with one fluid bolus of the set administered approximately every minute, so that the complete bolus set lasts approximately three minutes. In one example, a clinician may administer a fluid bolus very frequently, such as every minute or every few minutes, as they assess a patient's response to a medication or another medical intervention.In another example, if a patient is relatively stable in the ICU, fluid boluses may be administered less frequently, such as every hour, every six hours, etc. Thus, catheter 54 may be used to provide CCO and / or iCO, as described below with reference to Figures 6, 16, and 18-20A.

[0027] Catheter 54 is one example of a catheter that can be used to measure RVP, PAP, CCO, and / or iCO. In other examples, any catheter configured to measure RVP, PAP, CCO, and / or iCO can be used.

[0028] FIG. 4 is a perspective view of a hemodynamic sensor 14A that can be attached to a patient 16 to sense hemodynamic data representative of the patient's RVP or PAP. As shown in FIG. 4, the hemodynamic sensor 14A includes a housing 76, a fluid input port 78, a catheter-side fluid port 80, and / or an I / O cable 82. The fluid input port 78 is configured to be connected to a fluid source, such as a saline bag or other fluid input source, via tubing or other hydraulic connection. The catheter-side fluid port 80 is configured to be connected to a catheter inserted in the patient's arm (i.e., a radial artery catheter) or a catheter inserted in the patient's leg (i.e., a femoral artery catheter) via tubing or other hydraulic connection. The I / O cable 82 is configured to connect to the hemodynamic monitor 12, for example, via one or more of the I / O connectors 52 (shown in FIG. 2). The housing 76 of the hemodynamic sensor 14A encloses one or more pressure transducers, communication circuitry, processing circuitry, and corresponding electronic components for sensing fluid pressure corresponding to the RVP or PAP of the patient 16, which is transmitted via an I / O cable 82 to the hemodynamic monitor 12 (shown in FIG. 2).

[0029] During operation, a column of fluid (e.g., saline) is introduced from a fluid source (e.g., a saline bag) through fluid input port 78, through hemodynamic sensor 14A, and into catheter fluid port 80 toward a catheter inserted into patient 16. RVP or PAP is transmitted through the fluid column to a pressure sensor disposed within housing 76, which senses the pressure of the fluid column. Hemodynamic sensor 14A converts the sensed fluid column pressure into an electrical signal via a pressure transducer and outputs the corresponding electrical signal to hemodynamic monitor 12 (shown in FIG. 1) via I / O cable 82. Hemodynamic sensor 14A thus transmits analog sensor data (or a digital representation of the analog sensor data) representing substantially continuous beat-to-beat monitoring of patient 16's RVP or PAP to hemodynamic monitor 12 (shown in FIG. 1).

[0030] FIG. 5 is a perspective view of an oximetry module 14B for receiving oximetry data from a catheter inserted within a patient 16. As shown in FIG. 5, the hemodynamic sensor 14B includes an optical transmitter and an optical receiver disposed within a housing 86 and in communication with the catheter via an I / O connector 84 accessible through a protective door 88. Within the housing 86, as shown in FIG. 5, the hemodynamic sensor 14B includes communication circuitry, processing circuitry, and corresponding electronic components for sensing blood oxygen saturation data derived from optical emissions transmitted to the patient via the catheter and corresponding return light received from the patient 16 via the catheter. An electrical signal indicative of the patient's blood oxygen saturation level is transmitted to the hemodynamic monitor 12 via a cable 90 and a connector 92 that interfaces with one of the I / O connectors 52 (shown in FIG. 2).

[0031] 6 is a schematic block diagram showing the inputs and outputs of each module of hemodynamic monitoring system 10. Although hemodynamic monitoring system 10 is described with reference to catheter 54, any suitable catheter may be used with hemodynamic monitoring system 10. FIG. 6 illustrates RVP feature module 32, PAP feature module 34, verification module 36, flow module 38, and CO AE Module 40 (a module that estimates cardiac output based on an autoencoder model) and CO LR Module 42 (a module that uses a linear regression model to estimate cardiac output) and CO filtered Shown are a module 44 (a module for filtering one or more estimates of cardiac output) and an output device 102. The RVP feature module 32, the PAP feature module 34, the validation module 36, the flow module 38, the CO AE Module 40, CO LR Module 42, and CO filtered Each of the modules 44 is a functional module of the flow and cardiac output software code 30, as shown in FIG. 1. While the flow and cardiac output software code 30 is described herein as being divided into seven modules, in other examples, the functionality of the flow and cardiac output software code 30 may be described as more or fewer modules, which may depend on how the code is written or organized in some examples. Also, any module and / or sub-module may be an entirely separate collection of code. While the modules of the flow and cardiac output software code 30 are described in order, the modules may include overlapping or intermixed functionality.

[0032] The RVP feature module 32 is the first module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. The RVP feature module 32 generates the RVP waveform ("RVP"). waveform " ) to RVP features (" RVP featuresThe RVP feature module 32 includes in its code a method for extracting the RVP waveform RVP receives as input. waveform corresponds to the hemodynamic data sensed by one of the hemodynamic sensors 14A and received by the hemodynamic monitor 12. waveform The sensed hemodynamic data corresponding to the RVP is passed to the RVP feature module 32. waveform From RVP features The RVP feature module 32 extracts the RVP features to the verification module 36. The RVP feature module 32 features CO LR It also outputs to module 42.

[0033] The PAP feature module 34 is the second module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. The PAP feature module 34 is responsible for generating the PAP waveform ("PAP"). waveform " ) to PAP features (" PAP features The PAP feature module 34 includes a method for extracting the PAP waveform It receives as input. waveform corresponds to the hemodynamic data sensed by one of the hemodynamic sensors 14A and received by the hemodynamic monitor 12. waveform The corresponding hemodynamic data is passed to the PAP feature module 34. The PAP feature module 34 waveform From PAP features The PAP feature module 34 extracts the PAP features to the verification module 36. The PAP feature module 34 features CO LR It also outputs to module 42.

[0034] The validation module 36 is the third module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. The validation module 36 cleans the data and validates the RVP.waveform and PAP waveform The verification module 36 includes a method in the code to ensure that the RVP is valid and trustworthy. waveform and PAP waveform and RVP features and PAP features and RVP as input. waveform and PAP waveform The sensed hemodynamic data corresponding to RVP is passed to the validation module 36. features The features are provided from the RVP feature module 32 to the verification module 36. Similarly, the PAP features are provided from the PAP feature module 34 to the validation module 36. The validation module 36 determines whether a valid RVP ("RVP_valid") and a valid PAP ("PAP_valid") are true or false (i.e., RVP waveform and PAP waveform 6 (i.e., flow module 38, CO AE Module 40, CO LR Module 42, and CO filtered module 44) and RVP waveform and PAP waveform A determination indicating whether RVP_valid and PAP_valid are valid may be output to an output device 102, such as display 24 (shown in FIG. 1). An indication that RVP_valid and PAP_valid are true (e.g., "RVP_valid=true" and "PAP_valid=true") indicates that the RVP waveform and PAP waveform means that the hemodynamic data corresponding to RVP_valid and PAP_valid are valid for further processing by the modules contained within the dashed box. An indication that RVP_valid and PAP_valid are false (e.g., "RVP_valid=false" and "PAP_valid=false") indicates that RVP waveform and PAP waveform This means that the hemodynamic data corresponding to the smoothed RVP is not valid for further processing.waveform is output to the flow module 38, and RVP features and PAP features and CO LR Output to module 42.

[0035] The flow module 38 is the fourth module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. The flow module 38 is waveform The flow module 38 includes a method in the code to estimate the blood flow waveform from RVP. waveform RVP receives as input. waveform The sensed hemodynamic data corresponding to the RVP is passed to the flow module 38. The flow module 38 receives the smoothed RVP from the validation module 36. waveform The flow module 38 may also receive whether the valid flow rate ("flow_valid") is true or false (i.e., the processed blood flow rate ("flow processed ") is valid or invalid. AE Output to module 40, flow processed A determination indicating whether flow_valid is valid may be output to an output device 102, such as display 24 (shown in FIG. 1). An indication that flow_valid is true (e.g., "flow_valid=true") indicates that the estimated blood flow waveform flow processed But CO AE It means that the estimated blood flow waveform flow is valid for further processing by the module 40. An indication that flow_valid is false (e.g., "flow_valid=false") indicates that the estimated blood flow waveform flow processed The flow module 38 further defines a flow processed CO AE The flow module 38 outputs the flow processed can also be output to the output device 102.

[0036] CO AEModule 40 is the fifth module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. AE Module 40 includes in its code a method for deriving cardiac output from the estimated blood flow waveform. AE Module 40 is a flow processed as input. processed is the CO from the flow module 38 AE It is supplied to module 40. AE Module 40 also determines whether flow_valid is true or false and accordingly AE Receives an indication from flow module 38 indicating whether module 40 is able to proceed. AE Module 40 is an autoencoder for measuring cardiac output ("CO AE ") to CO LR Module 42 and CO filtered Output to module 44. AE Module 40 is a CO AE can also be output to the output device 102.

[0037] CO LR Module 42 is the sixth module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. LR Module 42 is the RVP features and PAP features The code includes a method for estimating cardiac output and change in cardiac output based on the CO LR Module 42 is the RVP features , PAP features , CO AE , CCO, and iCO as inputs. features is transmitted from the RVP feature module 32 or the verification module 36 to the CO LR module 42. Similarly, PAP features The CO is obtained from the PAP feature module 34 or the verification module 36. LR It is supplied to module 42. AE CO AEModule 40 to CO LR The CCO and iCO correspond to cardiac output data of the patient 16 received by the hemodynamic monitor 12 via the catheter 54 using a hot filament and / or thermistor and thermodilution techniques, as described above with reference to FIG. LR Passed to module 42. CO LR Module 42 performs linear regression of cardiac output ("CO LR ”) and the linear regression change in cardiac output (“ΔCO LR ") and CO filtered Output to module 44. LR Module 42 is CO LR and ΔCO LR can also be output to the output device 102.

[0038] CO filtered Module 44 is the seventh module of the flow and cardiac output software code 30 in the hemodynamic monitoring system 10. filtered Module 44 includes in its code a method for filtering estimates of cardiac output or changes in cardiac output via a Kalman filter algorithm. filtered Module 44 is CO AE , CO LR , CCO, iCO, and ΔCO LR It receives as input CO AE CO AE Module 40 to CO filtered It is supplied to module 44. LR and ΔCO LR CO LR Module 42 to CO filtered It is supplied to module 44. LR In the same manner as described above for module 42, the CCO and iCO filtered Also given to Module 44. CO filtered Module 44 measures filtered cardiac output ("CO filtered ") and the change in filtered cardiac output ("ΔCOfiltered ") is output to the output device 102.

[0039] The output device 102 is a device for receiving output from the modules of the flow and cardiac output software code 30. The output device 102 may include the display 24, as shown in FIG. 1. For example, the output device 102 may receive the flow, cardiac output, and / or final estimates of change in cardiac output for display via the display 24. The output device 102 may receive the flow, cardiac output, and / or final estimates of change in cardiac output for display via the display 24. AE Module 40, CO LR Module 42, and CO filtere d module 44. More specifically, the output device 102 can receive an output from the flow processed and CO AE and CO LR and ΔCO LR and CO filtered and ΔCO filtered Each of these outputs can be displayed via display 24 as a corresponding graph showing the value over time.

[0040] Although some modules are shown in Figure 6 as having multiple inputs and outputs, some of the inputs and outputs are optional, and many configurations of hemodynamic monitoring system 10 (shown in Figure 1) are possible. In some examples, PAP features , PAP waveform , CCO, and / or iCO may not be used. In some examples, an estimate of cardiac output for patient 16 (shown in FIG. 1) is calculated using CO AE Module 40, CO LR Module 42, and CO filtered Outputs can be from any one or more of the modules 44. These configurations can depend, for example, on the combination of hemodynamic sensors 14 (shown in FIG. 1) used or the desired output.

[0041] Each module of the hemodynamic monitoring system 10 will be described in more detail below. The RVP feature module 32 will be described with reference to FIG. 7. The PAP feature module 34 will be described with reference to FIG. 8. The verification module 36 will be described with reference to FIGS. 9-11. The flow module 38 will be described with reference to FIGS. 12-16. AE The module 40 will be described with reference to FIG. LR The module 42 will be described with reference to FIGS. 18 and 19. filtered The module 44 will be described with reference to FIGS. 20A to 22. FIG.

[0042] 7 is a graph showing an RVP waveform trace 104 including indices 106, 108, 110, 112, 114, 116, 118, and 120. In FIG. waveform 1 is an example waveform of RVP. waveform corresponds to hemodynamic data sensed by one of the hemodynamic sensors 14A and received by the hemodynamic monitor 12. The RVP waveform trace 104 (represented by the digital hemodynamic data) may include various indices indicative of blood flow and cardiac output of the patient 16. features is generated by the RVP via the RVP feature module 32 as described above with reference to FIG. waveform Extracted from RVP waveform Prior to extracting metrics from the RVP, a beat detection algorithm identifies the start and end of individual beats for each waveform. The RVP beat detection algorithm identifies the start of a beat based on maximum RVP, minimum RVP, maximum or minimum rate of change in RVP, and / or second derivative with respect to time in RVP. waveform After intrabeat identification, various indices of blood flow and cardiac output can be extracted from the waveform on a continuous beat-to-beat basis.

[0043] Index 106 of RVP waveform trace 104 corresponds to minimum diastolic blood pressure. Index 108 of RVP waveform trace 104 corresponds to end-diastolic blood pressure. Index 110 of RVP waveform trace 104 corresponds to maximum systolic blood pressure. Index 112 of RVP waveform trace 104 corresponds to end-systolic blood pressure. Slope S1 is the slope of RVP waveform trace 104 and may also provide an index. While slope S1 is depicted at one location, it is representative of multiple slopes that may be determined at multiple locations along RVP waveform trace 104. For example, index 114, which corresponds to the maximum rate of change of blood pressure with respect to time during systolic upstroke (dP / dt), and index 116, which corresponds to the minimum rate of change of blood pressure with respect to time during post-end-systolic relaxation (dP / dt), are further exemplary indexes. Similarly, the second time derivative of RVP waveform trace 104 may be determined at any location along RVP waveform trace 104 and used as an index. Other exemplary indices include the RVP gradient, or the difference in blood pressure at different times during diastole or systole. For example, index 118 corresponds to the diastolic gradient, or the difference between the minimum diastolic blood pressure (index 106) and the end-diastolic blood pressure (index 108). Index 120 corresponds to the right ventricular pulse pressure, or the blood pressure gradient, which is equal to the difference between the end-diastolic blood pressure (index 108) and the maximum systolic blood pressure (index 110).

[0044] For example, using the RVP feature module 32, additional metrics can be extracted from the RVP waveform trace 104 by the flow and cardiac output software code 30 based on the RVP waveform trace 104 during various time periods. For example, the systolic upstroke (metrics 108-110), systolic downstroke (metrics 110-112), isovolumic relaxation (metrics 112-106), diastole (metrics 106-108), and beat-to-beat interval (metric 106) can be determined by the flow and cardiac output software code 30. Such metrics may include the average RVP during one of the above-referenced intervals.

[0045] 8 is a graph showing a PAP waveform trace 122 including indices 124, 126, 128, 130, and 132 indicative of blood flow and cardiac output. waveform 1 is an example waveform of PAP. waveform corresponds to hemodynamic data sensed by one of the hemodynamic sensors 14A and received by the hemodynamic monitor 12. The PAP waveform trace 122 (represented by the digital hemodynamic data) may include various indices indicative of the blood flow and cardiac output of the patient 16. PAP features , as described above with reference to FIG. 1, via the PAP feature module 34. waveform Extracted from PAP waveform Prior to extracting metrics from the PAP, a beat-to-beat detection algorithm identifies the start and end of individual beats for each waveform. The PAP beat-to-beat detection algorithm identifies the start of a beat based on the maximum PAP, minimum PAP, maximum or minimum rate of change in PAP, and / or the second derivative of PAP with respect to time. waveform After intrabeat identification, various indices of blood flow and cardiac output can be extracted from the waveform on a continuous beat-to-beat basis.

[0046] Indicator 124 of PAP waveform trace 122 corresponds to the start of a heartbeat. Indicator 126 of PAP waveform trace 122 corresponds to the peak systolic pressure, indicating the end of the systolic upstroke. Indicator 128 of PAP waveform trace 122 corresponds to the presence of a dicrotic notch and pressure, indicating the end of the systolic decay. Indicator 130 of PAP waveform trace 122 corresponds to the minimum diastolic pressure of the patient's 16 heartbeat. Mean PAP can also be an indicator. PAP gradient, or the difference in blood pressure between points on PAP waveform trace 122, can also be an indicator. For example, indicator 132 corresponds to pulmonary artery pressure, or the difference between the minimum diastolic blood pressure (indicator 130) and the maximum systolic blood pressure (indicator 126). S2 is the slope of PAP waveform trace 122, which can also provide an indicator. While slope S2 is depicted at one location, it is representative of multiple slopes that may be determined at multiple locations along PAP waveform trace 122. For example, an indicator may include the maximum and / or minimum time derivative of PAP waveform trace 122.

[0047] For example, using the PAP feature module 34, additional metrics can be extracted from the PAP waveform trace 122 by the flow and cardiac output software code 30 based on the PAP waveform trace 122 during various intervals. For example, the interval from peak systolic pressure at indicator 126 to diastole at indicator 128 and the interval from the start of the heartbeat at indicator 124 to diastole at indicator 130 can be extracted from the PAP waveform trace 122. The flow and cardiac output software code 30 may identify additional metrics from the PAP waveform trace 122 during various intervals. For example, the systolic rise (indicators 124-126), the systolic fall (indicators 126-128), the systole (indicators 124-128), the diastole (indicators 128-130), and the beat-to-beat interval (between indicator 124) can be determined by the flow and cardiac output software code 30. Such metrics may include the average PAP during one of the above-referenced intervals. The area under the curve of the PAP waveform trace 122 and the standard deviation of the PAP waveform trace 122 determined for the above-referenced intervals may also serve as additional indicators for the patient 16 .

[0048] Verification module (Figures 9 to 11) FIG. 9 is a schematic block diagram illustrating validation module 36, which includes signal quality detector 134 and decision block 136. FIG. 10 is a graph illustrating RVP waveform trace 137, including indicators 108 and 110 indicative of catheter 54 (shown in FIG. 3) placement and data quality. FIG. 11 is a graph illustrating RVP waveform trace 104 and PAP waveform trace 122, including indicators 110, 126, 138, 140, and 142 indicative of catheter 54 placement and data quality. FIGS. 9, 10, and 11 will be discussed together. Validation module 36 utilizes RVP waveform trace 137 to detect and continuously monitor for catheter 54 placement and signal quality issues that affect data quality. waveform and PAP waveform Using data from

[0049] The validation module 36 filters or cleans the data and validates the RVP waveform , PAP waveform , RVP features , and PAP features are valid and reliable. The RVPs derived from the RVP feature module 32 and the PAP feature module 34, respectively, waveform RVP of features and PAP waveform PAP features is the input to the verification module 36. RVP waveform and PAP waveform The verification module 36 also inputs the RVP in the 10-second segment. waveform and PAP waveform RVP waveform and PAP waveformThe segment may be approximately 10 seconds (e.g., 9.5 to 10.5 seconds), or any other suitable time segment. The validation module 36 may be implemented within a patient care environment, such as an ICU, OR, or another patient care environment. The validation module 36 refines the data collected from the catheter 54 and determines whether the catheter 54 is correctly positioned to provide high quality data via a signal quality detector 134 and a decision block 136. Thus, the validation module 36 determines whether the data from the catheter 54 is properly positioned to provide high quality data via the flow module 38 and the CO LR It is determined whether it is available for further analysis, such as in module 42.

[0050] RVP features and PAP features is input to the signal quality detector 134. The signal quality detector 134 detects the RVP features and PAP features The data generated from the RVP signal quality index (SQI RVP ”) and PAP signal quality index (“SQI PAP ") and RVP, respectively. waveform and PAP waveform The signal quality detector 134 assigns the RVP features and PAP features and the standard deviation of RVP waveform and PAP waveform Artifacts in RVP waveform and PAP waveform and the standard deviation of the feature that represents the difference between

[0051] The signal quality detector 134 detects the RVP features and PAP features and RVP, respectively. features and PAP features Utilizes a signal quality algorithm to analyze data by comparing it to a specified range of values ​​for RVP features and PAP features Consume RVP waveform and PAP waveformWhen both are present, features can be compared between waveforms and features that represent differences between waveforms can be analyzed. The range of values ​​specified is the physiological limits plus system tolerances, or RVP. features , PAP features , and RVP waveform and PAP waveform The signal quality detector 134 calculates the RVP by calculating the standard deviation for each of the features that represent the difference between the RVP and the signal quality detector 134. features and PAP features The data quality is also checked by comparing the RVP with the specified range of values. waveform and PAP waveform It can be used to identify artifacts in the waveform and signal quality issues, such as under- or over-attenuated signals. Data outside of specified values, such as negative pressure, abnormally high pressure, or a peak systolic pressure 126 on the PAP waveform trace 122 that is higher than the peak systolic pressure 110 on the RVP waveform trace 104, represents physiologically impossible data.

[0052] Values ​​outside the specified range of values ​​represent erroneous data. The signal quality detector 134 cleans the data by removing the erroneous data and calculates the RVP. waveform and PAP waveform The signal quality detector 134 smooths the RVP based on the amount of error data detected. waveform and PAP waveform SQI for each 10-second segment RVP and SQI PAP The signal quality detector 134 assigns the SQI RVP and SQI PAP and output to decision block 136. waveform and PAP waveformErroneous data (identified by features that fall outside of a specified range of values) in each 10-second segment is removed from the 10-second segment at the heart rate level or any other appropriate level. If RVP_valid is true (e.g., "RVP_valid=true") and / or PAP_valid is true (e.g., "PAP_valid=true"), then RVP waveform The remaining 10-second segments and / or PAP waveform the remaining 10-second segments of the features and / or PAP features are passed on for further analysis, respectively.

[0053] 10 shows an exemplary RVP waveform trace 137 that may be analyzed by signal quality detector 134. Signal quality detector 134 may analyze the RVP waveform trace 137 by features The RVP waveform trace 137 is an RVP waveform similar to the RVP waveform trace 104 shown in FIG. waveform 1 is another exemplary waveform of RVP waveform trace 137. In this example, signal quality detector 134 compares the end-diastolic pressure 108 and peak systolic pressure 110 of each heartbeat of RVP waveform trace 137 to specified value ranges for RVP end-diastolic pressure 108 and RVP peak systolic pressure 110, respectively. Data from each heartbeat, from the RVP end-diastolic pressure 108 to the next RVP end-diastolic pressure 108, or from the RVP peak systolic pressure 110 to the next RVP peak systolic pressure 110, is analyzed against the corresponding specified value range. RVP waveform trace 137 includes end-diastolic pressure exclusion 108E and peak systolic pressure exclusion 110E.

[0054] FIG. 10 illustrates an RVP waveform trace 137 heartbeat extending from an RVP end-diastolic pressure 108 to the next RVP end-diastolic pressure 108, or from an RVP peak systolic pressure 110 to the next peak systolic pressure 110. The signal quality detector 134, via a signal quality algorithm, detects, flags, and removes heartbeats containing data outside of a specified range of values. Thus, the end-diastolic pressure exclusion 108E and peak systolic pressure exclusion 110E are within heartbeats from the segment shown in the RVP waveform trace 137 containing erroneous data, and the RVP waveform 10, the beat from peak systolic pressure 110 to the next peak systolic pressure exclusion 110E contains a blood pressure value outside the specified range of values ​​because the RVP waveform trace 137 contains artifacts, noise, and the preceding end-diastolic pressure exclusion 108E. As a result, the signal quality detector 134 flags and excludes that beat (from peak systolic pressure 110 to peak systolic pressure exclusion 110E in the RVP waveform trace 137), and all data from that beat is removed from the RVP waveform trace 137. waveform 10E. End-diastolic pressure exclusion 108E that falls within the range of that heartbeat is an end-diastolic blood pressure measurement that is excluded from further analysis. As further seen in FIG. 10 , the heartbeat from end-diastolic pressure exclusion 108E in RVP waveform trace 137 to the next end-diastolic blood pressure 108 contains an abrupt drop in blood pressure following peak systolic blood pressure exclusion 110E that is outside the specified range of values. The abrupt drop in blood pressure in RVP waveform trace 137 after peak systolic blood pressure exclusion 110E is outside the specified range of values ​​because such a sudden drop in blood pressure following peak systolic blood pressure 110 is physiologically unexpected. As a result, signal quality detector 134 flags and excludes that heartbeat, and all data from that heartbeat (from end-diastolic pressure exclusion 108E in RVP waveform trace 137 to the next end-diastolic blood pressure 108) is discarded. waveform The peak systolic blood pressure (RVP) measurements within the heartbeat are discarded from further analysis. features RVP including waveformBefore transferring the segment to the subsequent module, the RVP waveform trace 137 is waveform segment, thereby waveform The signal quality detector 134 detects the SQI RVP The excluded beats from the peak systolic pressure 110 to the peak systolic pressure exclusion 110E and from the end-diastolic pressure exclusion 108E to the next end-diastolic pressure 108E, which contain noise and sudden drops in blood pressure, are used in determining the PAP. waveform PAP for cleaning and analyzing features In a similar way to using PAP waveform can be analyzed.

[0055] The range of values ​​specified is RVP waveform and PAP waveform If both and are available, RVP features and PAP features Furthermore, it can also accommodate differences between RVP waveform and PAP waveform Various features can be derived from the difference between the RVP and waveform and PAP waveform depends on whether the is synchronous or asynchronous. Values ​​outside the specified range represent erroneous data.

[0056] As can be seen in FIG. 11, the signal quality detector 134 detects the RVP of the RVP waveform trace 104. features and PAP waveform trace 122 PAP features The RVP waveform trace 104 is synchronized with the PAP waveform trace 122. The RVP waveform trace 104 is synchronized with the PAP waveform trace 122. waveform , and PAP waveform trace 122 is an example waveform of PAP waveform In this example, the signal quality detector 134 analyzes the pulse transit time 138, the systolic slope 140, and the mean blood pressure 142.

[0057] Indicator 110 is determined from the RVP waveform trace 104. Indicator 110 is the peak systolic pressure, or end of systolic upstroke, of the RVP waveform trace 104. Indicator 126 is determined from the PAP waveform trace 122. Indicator 126 is the peak systolic pressure, or end of systolic upstroke, of the PAP waveform trace 122. Indicators 138, 140, and 142 are determined from the difference between the RVP waveform trace 104 and the PAP waveform trace 122. Indicator 138 is the pulse transit time, or the time elapsed between the peak systolic pressure (indicator 110) of the RVP waveform trace 104 and the peak systolic pressure (indicator 126) of the PAP waveform trace 122. Pulse transit time 138 is determined from the RVP waveform trace 104. waveform and PAP waveform delay between the heartbeat and the RVP waveform and PAP waveform and the time difference between when the RVP waveform trace 104 and the PAP waveform trace 122 are detected. Measure 140 is the systolic slope, or the pressure gradient between the peak systolic pressure (measure 110) of the RVP waveform trace 104 and the peak systolic pressure (measure 126) of the PAP waveform trace 122. Measure 142 is the mean blood pressure of the PAP waveform trace 122 and the RVP waveform trace 104.

[0058] The signal quality detector 134 compares the pulse transit time 138, systolic slope 140, and mean blood pressure 142 to specified ranges of values ​​for the pulse transit time 138, systolic slope 140, and mean blood pressure 142, respectively. The signal quality detector 134 determines the difference between the peak systolic blood pressure 110 of the RVP waveform trace 104 and the peak systolic blood pressure 126 of the PAP waveform trace 122, or the RVP features Heart rate and PAP calculated from features Other differences between features of the synchronized RVP waveform trace 104 and PAP wavelength trace 122 can also be checked, such as the difference between the heart rate calculated from the RVP waveform trace 104 and the PAP wavelength trace 122. For example, waveform indicates the presence of five heartbeats and synchronized PAP waveform indicates the presence of 10 heartbeats within the same time window, the signal quality detector 134 determines RVP waveform and / or PAP waveformIf the RVP waveform trace 104 and the PAP waveform trace 122 are asynchronous, the signal quality detector 134 detects that the data from the segment is not valid. features and PAP features can be analyzed.

[0059] The signal quality detector 134 detects and flags any errors or values ​​outside of specified ranges for each metric through a signal quality algorithm. waveform and PAP waveform and RVP features and PAP features and the RVP waveform trace 104 and the PAP waveform trace 122 containing the flagged data are filtered out before being transferred to a subsequent module. waveform and PAP waveform and thereby RVP waveform and PAP waveform and cleaning.

[0060] The signal quality detector 134 detects the SQI RVP and SQI PAP In determining RVP, parameters such as pulse transit time 138, systolic gradient 140, and mean blood pressure 142 are used. features and PAP features The indices 138, 140, and 142, among other parameters determined or derived from the RVP waveform trace 104 and the PAP waveform trace 122, are used by the signal quality detector 134 to determine whether the RVP waveform trace 104 is a PAP waveform trace or not. waveform and PAP waveform Therefore, RVP provides additional data that can be analyzed. waveform and PAP waveform Analyzing both allows for a more comprehensive assessment of data quality, resulting in a more accurate analysis.

[0061] As discussed above, the signal quality detector 134 determines the SQI via a signal quality algorithm. RVP and SQI PAP and a combined RVP and PAP signal quality algorithm ("SQI COMB ") and RVP, respectively. waveform and PAP waveform RVP from 10-second intervals features and PAP features The signal quality detector 134 uses such information in addition to the RVP features and PAP features and RVP waveform and PAP waveform Based on the characteristics that represent the difference between RVP and SQI PAP and SQI COMB and assign it.

[0062] SQI RVP and SQI PAP are the quality of each segment of the waveform, or RVP, respectively. waveform and PAP waveform RVP calculated from 10-second segments of features and PAP features For example, the signal quality detector 134 may assign a signal quality index from zero to five (0 to 5), with zero (0) being the highest quality and 5 being the lowest quality. RVP and SQI PAP can be any suitable range of values.

[0063] The signal quality detector 134 detects the RVP as discussed above. waveform RVP of features and PAP waveform PAP features , respectively. features The specified range of values ​​and PAP features The signal quality detector 134 compares the various RVPs to a specified range of values. features and PAP featuresRVP is calculated based on how close each is to a specified range of values. waveform and PAP waveform SQI for each of RVP and SQI PA Assign P. RVP waveform and PAP waveform RVP features and PAP features If is not within the specified range of values, the SQI will have values ​​from 2 to 5, respectively. RVP and SQI PAP RVP waveform and PAP waveform RVP features and PAP features SQI with values ​​from zero to one (0 to 1) if is within the specified range of values, respectively RVP and SQI PAP The signal quality detector 134 may assign an SQI RVP and SQI PAP and are output to the decision block 136.

[0064] SQI RVP and SQI PAP is provided to a decision block 136 of the verification module 36, which determines whether RVP waveform and PAP waveform The decision block 136 may output the decision to the output device 102 (shown in FIG. 6), which may be, for example, the display 24 (shown in FIG. 1). The decision block 136 may output a binary system of RVPs that are outside a specified range of values, respectively. features and PAP features RVP with waveform and PAP waveform Assign zero (0) to each RVP within the specified range of values. features and PAP features RVP with waveform and PAP waveform For example, assign 1 to RVP features and PAPfeatures If RVP has a signal quality index between 2 and 5, decision block 136 determines whether RVP waveform and PAP waveform Assign zero (0) to RVP features and PAP features If RVP has a signal quality index of zero or one (0 or 1), decision block 136 determines whether RVP waveform and PAP waveform Assign 1 to

[0065] SQI RVP If RVP_valid is assigned a zero (0), decision block 136 determines that RVP_valid is false (e.g., "RVP_valid=false"). This is indicated in FIG. 9 by the arrow labeled "No." The determination that RVP_valid is false can be transmitted from decision block 136 to a display, which may show the indication "INVALID RVP WAVEFORM." SQI RVP If RVP_valid is assigned a value of 1, decision block 136 determines that RVP_valid is true (e.g., "RVP_valid=true"). This is indicated in FIG. 9 by the arrow labeled "Yes." The determination that RVP_valid is true may be transmitted from decision block 136 to a display, which may show the indication "VALID RVP WAVEFORM." The validation module 36 is responsible for verifying the validity of the modules contained within the dashed box shown in FIG. 6 (i.e., flow module 38, CO AE Module 40, CO LR Module 42, and CO filtered It may also output an indication to module 44) indicating whether RVP_valid is true or false. If RVP_valid is true, the subsequent module may waveform If RVP_valid is false, you can continue using the data from the filtered 10 second segment. waveform You cannot continue using data from

[0066] If SQIPAP is assigned a zero (0), decision block 136 determines that PAP_valid is false (e.g., "PAP_valid=false"). This is indicated in FIG. 9 by the arrow labeled "No." The determination that PAP_valid is false can be transmitted from decision block 136 to a display, which may show the indication "INVALID PAP WAVEFORM." The PAP signal quality indicator SQI PAP If PAP_valid is assigned a value of 1, decision block 136 determines that PAP_valid is true (e.g., "PAP_valid=true"). This is indicated in FIG. 9 by the arrow labeled "Yes." The determination that PAP_valid is true may be transmitted from decision block 136 to a display, which may show the display "VALID PAP WAVEFORM." The validation module 36 is connected to the modules contained within the dashed box shown in FIG. 6 (i.e., flow module 38, CO AE Module 40, CO LR Module 42, and CO filtered It may also output an indication to module 44) indicating whether PAP_valid is true or false. If PAP_valid is true, the subsequent module may waveform If PAP_valid is false, you can continue using the data from the filtered 10-second segment. waveform You cannot continue using data from

[0067] SQI COMB RVP waveform and PAP waveform It is generated in the same way using features that represent the difference between SQI COMB RVP waveform 10-second segments of and / or PAP waveform This can be used by decision block 136 to further analyze the effectiveness of the 10 second segment.

[0068] If RVP_valid and / or PAP_valid are true, then the catheter 54 is correctly placed and the RVP features and RVP waveform and PAP features and PAP waveform is of high quality (i.e., has few artifacts and is suitable for further processing and analysis). If RVP_valid is true and PAP_valid is false, then the RVP features and RVP waveform may be less artifactual and still of high quality and still acceptable for further processing and analysis. COMB is assigned a value of 1, and SQI RVP If RVP_valid is assigned a value of 1, RVP_valid is still true and RVP features and RVP waveform is of high quality and acceptable for further analysis. For example, the validation module 36 may waveform can be output to the flow module 38 for determining blood flow. If RVP_valid is false alone or in combination with PAP_valid, then the catheter 54 is not properly placed or there is a signal quality problem, and the catheter 54 should be adjusted. The RVP derived from an improperly placed catheter 54 features and RVP waveform and / or PAP features and PAP waveform are excluded as being of low quality and not acceptable for further analysis.

[0069] The validation module 36 determines whether the signal quality is sufficient and whether the catheter 54 is operating within physiological limits. In some examples, the validation module 36 continuously monitors the placement of the catheter 54 via the signal quality. The validation module 36 determines whether the catheter 54 is correctly placed and therefore a valid RVP from which blood flow and cardiac output measurements can be derived. waveform and PAP waveformRVP to determine whether you are collecting accurate data that will waveform RVP of features and PAP waveform PAP features For example, RVP waveform or PAP waveform If the line is a flat line, the catheter 54 may not be properly connected to the hemodynamic monitor 12, and blood flow and cardiac output cannot be derived. waveform or PAP waveform If the values ​​for , , and are above or below the specified value ranges, the catheter 54 may be improperly positioned and delivering data indicative of central venous pressure rather than RVP and / or PAP, respectively. If the data is above or below the specified value ranges, the data is of low quality and should not be used, and the catheter 54 should be adjusted. If the data is within the specified value ranges, the data is of high quality and may be used for further analysis, such as in additional modules. Therefore, identifying and monitoring placement issues and / or other data quality issues in real time is important to notify clinicians and disable output of error information.

[0070] Flow Module (Figures 12 to 16) Figure 12 is a schematic block diagram showing the flow module 38. Figure 13 shows the RVP waveform is supplied via the flow module 38 processed 12 and 13 are discussed together. RVP waveform is input to the flow module 38 to predict hemodynamics. The flow module 38 includes an autoencoder model (or “autoencoder”) 144, a flow filter 146, and a decision block 148.

[0071] The flow module 38 is waveform as input and a processed blood flow waveform estimate of the patient 16, flow processed and the processed blood flow signal quality index (SQIflow The flow module 38 outputs the smoothed or cleaned RVP from the validation module 36, as discussed above with respect to FIGS. 9-11. waveform Also, RVP waveform The absolute value of the RVP is calculated using the flow module 38 (and the CO AE Module 40) and the relative RVP values ​​and RVP waveform 13, the RVP can be normalized for each patient (e.g., patient 16 shown in FIG. 1) so that only the shape of waveform is supplied to the flow module 38 via the flow processed Converted to RVP waveform is the raw blood flow rate (flow raw The waveform of the signal is converted into an estimate of the raw The flow is then passed through a flow filter 146 processed is converted to flow raw reflects physiological prediction of blood flow, and SQI flow The autoencoder model 144 and the flow filter 146 each generate a beat-to-beat flow processed To generate RVPs, waveform and flow raw In some examples, the flow module 38 analyzes the flow processed is output continuously.

[0072] Autoencoder model 144 is RVP waveform Input and flow raw The autoencoder model 144 is a deep learning-based model that performs well when the patient 16 has both low and high cardiac outputs. The autoencoder model 144 generates an output of RVP as discussed below with respect to FIG. waveform flow raw The autoencoder model 144 is trained to convert raw is output to the flow rate filter 146.

[0073] The flow filter 146 extracts the flow raw as input and processed and SQI flow The flow rate filter 146 outputs the flow raw By comparing the characteristics of flow to a range of specified values ​​of blood flow that correspond to physiological limits, raw Thus, the flow filter 146 filters or cleans the flow artifacts or physiological inaccuracies such as negative flow rates, flow values ​​in the thousands, squared flow rates, and inconsistencies such as excessive or sudden increases in flow rate, or excessively different flow rates from beat to beat. processed To remove or exclude from a flow raw As a result, the flow filter 146 produces a smoother, more consistent, and more accurate blood flow waveform, flow processed Therefore, the flow from the flow filter 146 processed flow raw is more accurate and physiologically relevant than

[0074] The flow filter 146 is flow Determine the flow processed Allocate to SQI flow flow processed The flow filter 146 is a value that indicates the quality of the waveform of the flow processed Plus SQI flow To output the flow, as discussed above, raw The flow filter 146 compares the flow to a specified range of values ​​for blood flow. raw Based on the amount of physiological inaccuracy and / or discrepancy or error data detected and removed from the SQI flow flow processed For example, assign to flow processed has an SQI between zero and five (0 and 5) flow is assigned, with zero (0) being the highest quality and five being the lowest quality.flow may be in any suitable range of values. raw The flow filter 146 compares the flow to a specified range of values ​​for blood flow. raw SQI based on how close it is to a specified range of values. flow flow processed Assign to flow raw If is not within or close to all or some of the specified ranges of values, then flow processed has an SQI value ranging from 2 to 5. flow flow can be assigned. raw If is within or close to the range of all specified values, then flow processed is the SQI, which has a value between zero and one (0 to 1). flow The flow rate filter 146 may assign an SQI flow to decision block 148.

[0075] SQI flow flow processed is provided to a decision block 148 in the flow module 38 which determines whether flow is valid. processed The decision block 148 may output a determination of whether the flow is valid or invalid to the output device 102 (shown in FIG. 6), which may be, for example, the display 24 (shown in FIG. 1). The decision block 148 is a binary system and outputs a flow that is not within a specified range of values. processed Assigns zero (0) to the flow within the specified range of values. processed For example, assign 1 to SQI flow If has a value between 2 and 5, decision block 148 determines whether flow processed Assign zero (0) to SQI flow If has a value of zero or one (0 or 1), decision block 148 determines whether flow processed Assign 1 to

[0076] SQI flowIf SQI is assigned a zero (0), then decision block 148 determines that flow_valid is false (e.g., "Flow_valid=false"). This is indicated in FIG. 12 by the arrow labeled "No." The determination that flow_valid is false can be transmitted from decision block 148 to a display, which may show the indication "INVALID BLOOD FLOW." flow If flow_valid is assigned a value of 1, then decision block 148 determines that flow_valid is true (e.g., "Flow_valid=true"). This is indicated in FIG. 12 by the arrow labeled "Yes." The determination that flow_valid is true may be transmitted from decision block 148 to a display, which may show the display "VALID BLOOD FLOW." The flow module 38 may transmit an indication to the CO indicating whether flow_valid is true or false. AE It can also be output to module 40, AE Module 40 is a valid flow processed can be used as input.

[0077] If flow_valid is true, then flow processed is of high quality and acceptable for further processing and analysis. processed can itself be used to improve patient care. If flow_valid is false, flow processed are of low quality and are excluded from further processing and analysis.

[0078] The flow module 38 is waveform flow processed Convert it into a flow processed Determine whether the flow is of high quality. processed can be used alone at the point of patient care or in conjunction with other modules to generate cardiac output.

[0079] 14 is a schematic block diagram illustrating the autoencoder model 144 of the flow module 38. The autoencoder model 144 includes an input 150, a filter 152, a latent space 154, a filter 156, and an output 158. The input 150 and the filter 152 constitute an encoder 160, and the filter 156 and the output 158 ​​constitute a decoder 162.

[0080] The input 150 of the autoencoder model 144 is the RVP waveform By using a 10-second sample, the sample size is large enough for the waveform to be stable and more accurate, but the samples are not continuous RVP. waveform RVP waveform The 10-second samples may be approximately 10-second samples (9.5 to 10.5 seconds) or any other suitable time samples. The autoencoder model 144 may receive the 10-second samples on a rolling basis; for example, sample input 150 may be a 10-second sample but input to the autoencoder model 144 every 2 seconds. The input 150 is encoded via filter 152 of the autoencoder model 144 into a condensed version of the data from the input 150 in a latent space 154. The latent space 154 stores the condensed data. The condensed data in the latent space 154 is then decoded via filter 156 of the autoencoder model 144 to become an output 158. The output 158 ​​of the autoencoder model 144 is referred to as a flow raw The autoencoder model 144 is based on the RVP, as described below with respect to FIG. waveform From flow raw Thus, the encoder 160 of the autoencoder model 144, via the filter 152, is trained to produce the RVP waveform The decoder 162 of the autoencoder model 144 is the part of the autoencoder model 144 that encodes or compresses the input 150 into fewer variables that contain all of the important information of the input 150. rawIt takes such compressed encoded information and expands it through a filter 156 to produce an output 158 ​​of

[0081] Traditionally, blood flow measurements of a patient 16 are unavailable to medical personnel 18 due to the invasive nature of measuring blood flow in humans. Also, conventional autoencoder models are trained to learn a latent space representation of an input, and the compressed input from the latent space is expanded to an output identical to the input. For example, a conventional autoencoder may have an input of an RVP waveform and produce an output of the RVP waveform.

[0082] The autoencoder model 144 is a machine learning model. More specifically, the autoencoder model 144 is a machine learning model based on the RVP model, as discussed below with respect to FIG. waveform flow raw The autoencoder model144 is a deep learning-based model trained to use a neural network architecture to convert RVP waveform flow raw can be converted to a robust flow processed through the flow filter 146 to bring the flow raw Therefore, the autoencoder model 144 provides a practical, minimally invasive method for generating estimates of continuous blood flow measurements in a patient 16. As a result, blood flow measurements in the patient 16 are available for improved patient care. processed CO LR Module 42 and CO filtered As discussed below with respect to module 44, it may also be used to generate cardiac output for patient 16.

[0083] 15 is a schematic diagram illustrating filters 152 and 156, respectively, of encoder 160 and decoder 162 of autoencoder model 144. Filter 152 of encoder 160 comprises multiple filters 152, and filter 156 of encoder 162 comprises multiple filters 156.

[0084] A filter 152 is applied to the input 150 of the autoencoder model 144. The filter 152 is waveform The flow is a mathematical operation that converts the input 150 into compressed data. The data samples of the input 150 are reduced by a given factor by each filter 152 until the fully compressed data arrives in the latent space 154. The filters 156 are raw The filter 156 is applied to the compressed information in the latent space 154 to produce an output 158. The data samples are expanded by a given coefficient by each filter 156 until the output 158 ​​is reached. An algorithm within the autoencoder model 144 automatically generates the architecture of the filters 152 and 156, which is a convolutional neural network with automatically selected convolutions and pooling based on the input 150 and the desired output 158. A complex nonlinear relationship exists between the filtered data in each filtering layer 152 and 156. In the example of FIG. 15, the filters 152 and 156 have an 8-16-32 architecture. The first layer of filters 152 has 8 filters, the second layer of filters 152 has 16 filters, and the third layer of filters 152 has 32 filters. The first layer of filters 156 has 32 filters, the second layer of filters 156 has 16 filters, and the third layer of filters 156 has 8 filters. In the example of Figure 15, the autoencoder model 144 has 12,000 parameters and requires 112 convolution operations. The parameters are determined when the autoencoder model 144 is trained.

[0085] The filter 152 is an RVP waveform The filter 156 allows the compression and encoding of the compressed data flow. raw As a result, the filters 152 and 156 of the trained autoencoder model 144 are waveform flow raw can be converted to

[0086] Figure 16 shows human RVP waveform Based on human flow raw 1 is a flow diagram illustrating a process 164 for training an autoencoder model 144 to predict . The process 164 for training an autoencoder model 144 includes steps 166 through 178. Once the autoencoder model 144 has been trained via process 164, the autoencoder model 144 is ready for use in the flow module 38, as discussed above.

[0087] In step 166, the autoencoder model 144 is trained using the animal's measured RVP waveform and the animal's measured blood flow. Both the animal's RVP waveform and the animal's blood flow are known in step 166. Because directly measuring blood flow in humans is too invasive, information is derived from animal experiments by directly measuring blood flow in the same animal against the animal's known RVP waveform. The autoencoder model 144 is trained using the animal's RVP waveform as input and the animal's measured blood flow as output to obtain the animal's trained autoencoder model 144.

[0088] In step 168, the animal's RVP waveform is input into the animal's trained autoencoder model 144 to generate a predicted animal blood flow. The known animal's RVP waveform is input into the animal's trained autoencoder model 144, and a predicted animal blood flow waveform is output from the autoencoder model 144.

[0089] In step 170, the measured animal blood flow is compared to the predicted animal blood flow to validate the training of the autoencoder model 144. Because both the animal's RVP waveform and the associated waveform of the measured animal blood flow are known, the predicted animal blood flow from the animal's trained autoencoder model 144 can be compared to the known directly measured animal blood flow for various portions of the animal's RVP waveform. Thus, the predicted animal blood flow waveform from the autoencoder model 144 is compared to the directly measured animal blood flow waveform. To determine the performance of the autoencoder model 144, metrics such as mean squared error (MSE), mean absolute error (MAE), correlation, and Bland-Altman analysis are used in the comparison.

[0090] In step 172, it is determined whether the predicted animal blood flow from the animal's trained autoencoder model 144 is valid. If the error between the predicted animal blood flow from the animal's trained autoencoder model 144 and the directly measured animal blood flow is less than a preset threshold, the training of the autoencoder model 144 is determined to be valid. If the training of the autoencoder model 144 is determined to be valid, the process 164 proceeds to step 174. If the error between the predicted animal blood flow from the animal's trained autoencoder model 144 and the directly measured animal blood flow exceeds a preset threshold, the training of the autoencoder model 144 is determined to be invalid. If the training of the autoencoder model 144 is determined to be invalid, the process 164 returns to step 166 and begins training the autoencoder model 144 again. The MSE or MAE metric is iteratively calculated during training to adjust the model parameters until convergence (e.g., minimum MAE or MSE).

[0091] In step 174, the RVP of the human is calculated to generate raw blood flow for the human. waveformis input to the trained autoencoder model144 of the animal. waveform is input to the trained autoencoder model144 of the validated animal, and the raw human blood flow waveform (i.e., “human flow raw ") is output from the autoencoder model 144. Thus, step 174 is used to generate the shape of the human raw blood flow waveform. Since the training of the autoencoder model 144 on animals has been validated, the shape of the human raw blood flow waveform should match the human blood flow (e.g., the human raw blood flow waveform is RVP waveform is greater than zero, indicating the valve is open, and RVP waveform is close to zero, indicating the valve is closed).

[0092] In step 176, the human raw blood flow waveform is converted into a human scaled flow raw Since the shape of the human raw blood flow waveform is accurate, the human raw blood flow waveform is scaled in step 176 to produce flow over each heartbeat. raw is scaled so that the average integral of is the same as the i measured by catheter 54. raw The amplitude of the waveform is adjusted. raw The exact estimated flow raw is generated.

[0093] In step 178, the autoencoder model 144 generates a scaled flow raw The autoencoder model 144 is retrained using the human RVP waveform as input and the human scaled flow to obtain the human trained autoencoder model 144. raw The autoencoder model is trained using the scaled flow of humans as the output. rawWhen retrained using the , the autoencoder model144 can predict the human RVP, which is unknown a priori. waveform From human flow raw can be used to generate

[0094] The animal data allows for training an autoencoder model 144. As a result of training the autoencoder model 144 using the animal data, the autoencoder model 144 is able to measure the RVP. waveform flow, which cannot be measured in humans. raw The availability of blood flow measurements allows for further analysis of the patient's condition and improved patient care.

[0095] CO AE Module (Figure 17) Figure 17 shows the flow processed Autoencoder cardiac output CO AE Convert to CO AE FIG. 4 is a schematic block diagram showing a module 40.

[0096] CO AE Module 40 is a flow processed Input of CO AE The output of flow is processed is the CO from the flow module 38 AE module 40, and the flow module 38 outputs the flow raw and passes through the flow filter 146 raw flow processed Filter to flow processed CO AE Before sending to module 40, flow processed The flow module 38 determines that the CO AE Module 40 (flow processed can continue (in response to an indication that the flow is valid) or processedThe CO also provides an indication of whether flow_valid is true or false so that the CO cannot continue (in response to an indication that flow_valid is invalid). AE Transmit to module 40. CO AE The module 40 measures the CO AE To derive flow processed Use.

[0097] CO AE Module 40 is a CO AE To generate the flow processed Using a portion of the waveform, flow is measured at the heart rate level. processed Calculate the average value of the waveform of flow processed The waveform of CO AE is integrated to produce CO AE flow processed is a single value that represents an estimate of cardiac output for a 10-second portion of the flow processed The waveform is integrated every 10 seconds, and the CO AE Samples are taken on a rolling basis and more frequently than every 10 seconds. AE The sample can be input to module 40, in which case the sample is processed 10-second sample, but CO AE can be input to module 40. processed By utilizing a 10 sec sample, the sample size is AE The waveform is large enough to be stable and more accurate, but the sample is AE It is small enough to represent the continuous waveform of CO AE It has a robust continuous waveform of CO AE The delay in CO2 should not be too long to capture any changes in cardiac output in the patient. AE It is important to frequently update the data, which improves patient care.

[0098] Also, CO AE Module 40 is a CO AE is the relative RVP and RVP waveformIt can be generally described as "data independent" with respect to absolute pressure values ​​(e.g., absolute values ​​of RVP) because it can be based on the shape or form of the device and does not directly depend on absolute pressure values. This is important because, in practice, absolute pressure values ​​are not always accurate. For example, healthcare professionals may introduce some measurement error into RVP readings when leveling a pressure transducer for a patient. As another example, raising or lowering a patient's bed for surgery or in the ICU can also change absolute pressure values. AE Module 40 does not directly depend on absolute pressure values, so CO AE Furthermore, CO AE CO can accurately calculate not only low stroke volumes, but also high stroke volumes not previously possible, further improving patient care. AE Module 40 is a CO AE CO LR Module 42 and CO filtered The output may be to module 44 as well as to output device 102 (shown in FIG. 6).

[0099] CO LR Module (Figures 18-19) Figure 18 shows the CO LR 19 is a schematic block diagram showing the CO LR 18-19 are a schematic block diagram showing the reference features and current features used as input to the regression model 180 of module 42. RVP features , PAP features , demographic information, and SvO2, RVP waveform and PAP waveform to estimate cardiac output and / or changes in cardiac output based on the characteristics of CO LR 18-19 show a CO2 regression model including a regression model 180 and a cardiac output estimator 182. LR Module 42 is shown.

[0100] Regression model 180 is COLR It is the first submodule of module 42. The regression model 180 is features , PAP features , demographic information, and SvO2 as inputs. The regression model 180 is configured to calculate ΔCO, which is an estimate of the change in cardiac output over time for the patient 16 (shown in FIG. 1). LR The regression model 180 is a conventional machine learning model that works well when the available data is well-distributed rather than multimodal. For example, changes in patient cardiac output data tend to be more homogeneous data sets compared to patient cardiac output data, so conventional machine learning techniques such as the regression model 180 can be used to successfully train across populations.

[0101] Regression model 180 is trained using a training data set in which changes in cardiac output and corresponding changes in RVP and PAP waveform characteristics are known for a patient. In examples in which demographic information and / or SvO2 are used, regression model 180 can be trained using a training data set that also includes demographic information and / or SvO2 corresponding to known changes in cardiac output information.

[0102] Once the regression model 180 is trained, it is validated by the RVP feature module 32 (or validation module 36) as shown in FIG. waveform RVP derived from features The regression model 180 uses the PAP feature module 34 to generate the PAP waveform PAP derived from features You can also consume RVP. features and PAP features are aggregated over a defined time window (or interval). For example, RVP features and PAP features can be aggregated over a 10 second time window that includes multiple heartbeats. features and PAPfeatures can be aggregated over longer or shorter time windows. The RVP feature module 32 and PAP feature module 34 (shown in FIGS. 1 and 6) extract features for each heartbeat within the time window and then calculate the average of each feature over the entire time window. The averaged features over the time window are used to generate the RVP feature that is consumed by the regression model 180. features and PAP features In some examples, regression model 180 may be in the form of an averaged RVP obtained from good (i.e., valid or non-discarded) beats within a time window determined by validation module 36 (shown in FIG. 9 ). features and PAP features Consume only.

[0103] Regression model 180 is RVP features and PAP features In one example, the regression model 180 uses the RVP feature model 32 to generate the RVP. waveform All RVPs derived from features and PAP feature module 34 waveform All PAPs derived from features In another example, the regression model 180 can use the derived RVP features and PAP features In yet another example, the regression model 180 may use only a portion of the RVP features Use only PAP features Generally, the regression model 180 does not use the individual RVPs. features and PAP features Any number or combination of can be used. An exemplary RVP used by regression model 180 featuresincludes the following: pulse rate (i.e., a pressure-based heart rate calculation, although in other examples heart rate referenced from an electrocardiogram or other means can also be used), max dP / dt, min dP / dt, systolic time, systolic blood pressure, end systolic blood pressure, end diastolic blood pressure, pulse pressure, and mean blood pressure. PAP used by regression model 180 features Also included are the following: pulse rate, maximum dP / dt, minimum dP / dt, systolic time, systolic blood pressure, end-systolic blood pressure, end-diastolic blood pressure, pulse pressure, and mean blood pressure. Other RVPs not specifically listed herein, such as other pressure or time-based parameters, may also be used. features and / or PAP features can also be used by the regression model 180. Features such as pulse rate and systolic time are, for physiological reasons, related to RVP. waveform or PAP waveform , or the RVP. features and PAP features The physiological similarities between RVP and waveform RVP derived from features and PAP waveform PAP derived from features Only one source can be used, rather than both, and the regression model 180 can predict other features that are expected to be physiologically different, such as maximum dP / dt and minimum dP / dt, based on RVP. features and PAP features Both can be used.

[0104] In some examples, the regression model 180 can also consume demographic information about the patient 16. Such demographic information can include age, weight, height, sex (or gender), body mass index (BMI), health status, etc. The regression model 180 can receive the demographic information about the patient 16, for example, from either the system memory 22 (shown in FIG. 1 ) where the demographic information about the patient 16 is stored (e.g., in a patient information database) or when entered by the medical professional 18 at the user interface 46 (shown in FIG. 1 ). In some examples, the regression model 180 can also consume SvO2 data received via the hemodynamic sensor 14B (shown in FIGS. 1 and 5 ). Overall, the RVP of the patient 16 can be calculated using the regression model 180's SvO2 data. features , PAP features Several variations of regression model 180 (i.e., more complex or less complex variations of regression model 180) using different combinations of , demographic information, and / or SvO2 are possible without changing the quality of regression model 180.

[0105] As shown in FIG. 19, the regression model 180 is i The corresponding reference RVP feature ("RVP features_ref ") and the current time window W i+n (where "n" represents any integer value) features_current ") and RVP features_ref are individual RVPs features Configure the first set of RVP features_current are individual RVPs features Similarly, the regression model 180 constructs a second set of i The corresponding reference PAP feature ("PAP features_ref ") and the current time window W i+n The current PAP feature ("PAP features_current ") and PAP features_ref individual PAP features Configure the first set of PAP features_currentindividual PAP features This constitutes the second set of

[0106] Reference time window W i is ΔCO LR is the initial time interval calculated by the regression model 180. The reference time window W i is ΔCO LR To calculate i+n ) in some examples. i may be the time interval corresponding to when the patient 16 is first admitted to the ICU, OR, or other patient care environment. i may be a time interval corresponding to a first known cardiac output measurement, such as a measurement of iCO of the patient 16. In yet another example, the reference time window W i RVP waveform and / or PAP waveform may be any suitable time interval over which .times. ...

[0107] CO LR Module 42 determines the reference time window W i can be configured to update the reference time window W to a new time interval. i may be updated periodically. In some examples, the reference time window W i Then RVP features and PAP features may be updated after a set period of time when it is considered too old. i RVP waveform and / or PAP waveform can be continuously updated at each time interval measured. That is, in such an example, the reference time window W i is the current time window W i+n Since the time interval immediately before i is the reference time window W(i+n)-1 In another example, the reference time window W i can be updated at any time either manually by the medical professional 18 via the user interface 46 or automatically as instructed in the flow and cardiac output software code 30. In yet another example, the reference time window W i is not updated.

[0108] As shown in Figure 19, the reference time window W i is a 10-second interval, and RVP waveform The reference time window W represents the 10-second portion of i The length (in hours) of CO AE The reference time window W can be determined based on the minimum requirements for proper functioning of module 40 or other modules represented by flow and cardiac output software code 30. i Other time lengths are possible for the reference time window W, such as longer or shorter than a 10 second interval. i and the current time window W i+n The current time window W i+n The length of the reference time window W i Therefore, the length of the current time window W i+n The RVP can also be set to a 10-second interval. waveform can represent different 10-second portions of

[0109] Current Time Window W i+n is ΔCO LR is the current time interval calculated by the regression model 180. The current time window W i+n is the reference time window W i The current time window W can be any time interval after i+nmay correspond to real-time or near real-time measurements from the patient 16, thus enabling the regression model 18 to generate real-time or near real-time estimates of changes in cardiac output of the patient 16.

[0110] RVP features_ref and PAP features_ref Each individual one of features_current or PAP features_current corresponds to each individual one of the RVPs. features_ref and PAP features_ref are the RVPs measured from the corresponding time windows, respectively. features_current and PAP features_current For example, the RVP used by the regression model 180 features One of them is RVP, as shown in Figure 7. waveform In such an example, the regression model 180 may be the end-diastolic blood pressure 108 derived from the reference time window W i RVP in waveform End-diastolic blood pressure measurements derived from (i.e., end-diastolic blood pressure w_i ) and the current time window W i+n RVP during waveform End-diastolic blood pressure measurements derived from (i.e., end-diastolic blood pressure w_i+n ) and the same example, w_i RVP features_ref One of the end-diastolic blood pressure w_i+n RVP features_current It is one of them.

[0111] Regression model 180 is ΔCO LR RVP of patient 16 is multiplied by a coefficient to estimate features , PAP features , demographic information, and / or S02-based variables. More specifically, the regression model 180 is a linear model constructed to include variables based on the reference time window W i to the current time window W i+n ΔCO LRTo determine the reference time window W i RVP between features_ref and PAP features_ref and the individual features of the current time window W i+n RVP between features_current and PAP features_current The variation between the corresponding individual features of

[0112] The variables used in the regression model 180 can take several different forms. The variables used in the regression model 180 are RVP features_ref , PAP features_ref , RVP features_current , PAP features_current , demographic information, and / or SvO2. The variables used in the regression model 180 may include one or more measured or received values ​​of RVP, features_current or PAP features_current One of them and RVP features_ref or PAP features_ref The regression model 180 may also include a difference variable that represents the difference between a corresponding one of the current feature and the reference feature. The difference variable may also be expressed as a ratio of the difference between the current feature and the reference feature divided by the reference feature, resulting in a normalized variable. For example, one difference variable in regression model 180 may be (HeartRate w_i+n -HeartRate w_i ) / HeartRate w_i It can be expressed as:

[0113] The example difference variable shown above is the current time window W i+n heart rate measurements and the reference time window W i The difference between the heart rate measurements in the reference time window W i RVP divided by the heart rate measurement. features_current or PAP features_current One of them and RVP features_ref or PAP features_refSimilar difference variables can be generated for any pair of the corresponding ones of the regression model 180. The variables used in the regression model 180 are RVP features_ref , PAP features_ref , RVP features_current , PAP features_current , demographic information, and one or more measured or received values ​​of SvO2, and / or one or more difference variables. For example, one combination variable in regression model 180 may be:

[0114]

number

[0115] It can be expressed as:

[0116] The exemplary combination variables shown above are the reference pulse pressure and the current pulse pressure, the current time window W i+n measurements, as well as the current time window W i+n Includes a difference variable using a measure of systolic time of RVP. features_ref , PAP features_ref , RVP features_current , PAP features_current Similar combination variables can be generated for any combination of one or more measured or received values ​​of demographic information, and SvO2, and / or one or more difference variables.

[0117] The terms in the regression model 180 include or are formed from the variables described above multiplied by predetermined coefficients. For example, the coefficients may be determined during a training process for the regression model 180. The regression model 180 may include any number of terms from 1 to n. The regression model 180 may be configured to calculate the ΔCO LR The general form of the regression model 180 is: ΔCO LR=(a1×del_feat1)+(a2×del_feat2)+(a3×del_feat3)+...+(a n ×del_feat n ) (Formula 1) where:

[0118]

number

[0119] is.

[0120] In the above formula 1, the variables del_feat1 to del_feat 11 RVP features In another example, the variables in Equation 1 are based on the PAP features or RVP features , PAP features , and / or demographic information.

[0121] ΔCO represented in Equation 1 above LR is the reference time window W i The initial cardiac output corresponding to the current time window W i+n The variables del_feat1 through del_feat 11 is a normalized variable, so ΔCO LR represents the percent change rather than the magnitude of the change. LR ΔCO may be particularly useful to healthcare professionals 18 in this format, for example, when the percent change is more relevant compared to the magnitude of the change for a particular application, or when the magnitude of the change in cardiac output may vary widely within a patient population, but the percent change is more consistent. LR is the reference time window W i The corresponding iCO measurement or CO AE CO from module 40 AE From the known cardiac output values, such as ΔCO LRThis can be converted into the magnitude of the change by multiplying it by .

[0122] Alternatively, the regression model 180 may be features_ref or PAP features_ref By including non-normalized variables in ΔCO LR For example, one unnormalized difference variable in regression model 180 may be configured to calculate (EndSystolic Pressure w_i+n -EndSystolicPressure w_i ) and one non-normalized combination variable in the regression model 180 is (HeartRate w_i+n -HeartRate w_i )×(HeartRate w_i+n ×SystoleTime w_i+n ) It can be expressed as:

[0123] In yet another example, the regression model 180 calculates ΔCO LR A combination of normalized and denormalized variables can be used to calculate . For example, one such combination is

[0124]

number

[0125] It can be expressed as:

[0126] ΔCO LR is the output of the regression model 180. LR is the reference time window W for patient 16. i and the current time window W i+n This shows how much cardiac output has changed between ΔCO and ΔCO. LRis independent and important information regarding the hemodynamic status of the patient 16. The healthcare professional 18 may monitor the CO2 to determine whether the patient 16 is increasing or decreasing cardiac output. LR ΔCO from module 42 LR For example, a healthcare professional 18 may be performing an intervention on a patient 16, such as administering an inotropic substance, and may want to know how much cardiac output is changing, but may not need to know the actual value of cardiac output at that time.

[0127] In some instances, CO LR The module 42 outputs ΔCO from the regression model 180 to the output device 102 as shown in FIG. LR For example, the output device 102 may output a ΔCO LR A graph showing the values ​​of CO over time can be displayed. LR The module 42 monitors the hemodynamic status of the patient 16 by measuring ΔCO LR to the output device 102. In some examples, the regression model 180 outputs CO LR The cardiac output estimator 182 in module 42 calculates ΔCO LR can be passed.

[0128] The cardiac output estimator 182 calculates the CO LR It is the second sub-module of module 42. The cardiac output estimator 182 calculates ΔCO LR and CO AE The cardiac output estimator 182 receives as inputs the CCO, CCO, and iCO. LR The cardiac output estimator 182 outputs a reference time window W i By using the cardiac output value corresponding to ΔCO (i.e., the "reference cardiac output"), LR From CO LR Calculate.

[0129] The cardiac output estimator 182 calculates the ΔCO from the regression model 180. LRAs explained above, ΔCO LR is the reference time window W i to the current time window W i+n The cardiac output estimator 182 represents the change in cardiac output from CO AE Module 40 to CO AE Specifically, the cardiac output estimator 182 also receives a reference time window W i CO corresponding to AE The cardiac output estimator 182 is also configured to receive the CCO measurement and the iCO measurement. Specifically, the cardiac output estimator 182 uses a value of i Use the CCO or iCO measurements corresponding to the reference time window W i CO corresponding to AE The value of iCO, the measured CCO, and the measured iCO are each a reference cardiac output. One of the reference cardiac outputs is used to initialize or calibrate the cardiac output estimator 182 (i.e., CO LR is used to provide a starting point for calculating

[0130] The cardiac output estimator 182 has the following general form: CO LR =RefCO w_i +(ΔCO LR ×RefCO w_i ) (Formula 2) According to CO LR Calculate where: RefCO w_i is the CCO measurement, the iCO measurement, or the reference time window W i CO corresponding to AE is the value of (i.e., reference cardiac output), ΔCO LR is the output from the regression model 180, ΔCO LR is the value.

[0131] In the above formula 2, RefCO w_i CO AE value of, or the reference time window W ican be, but not necessarily, CCO or iCO measurements corresponding to a reference time window W i If a corresponding iCO measurement is available, CO LR Module 42 can be configured to use measurements of iCO. LR Module 42 uses CO instead of CCO or iCO measurements. AE can be configured to use

[0132] As shown in Equation 2, the cardiac output estimator 182 calculates the CO LR The magnitude of the change in cardiac output is added to the reference cardiac output to form CO LR is the output of the cardiac output estimator 182. LR is the current time window W i+n is the value of cardiac output of the patient 16 corresponding to CO LR Module 42 outputs the CO 2 signal from cardiac output estimator 182 to output device 102, as shown in FIG. LR For example, the output device 102 may output a CO LR A graph showing the values ​​of CO over time can be displayed. LR The module 42 may include a CO LR to the output device 102. In some examples, the cardiac output estimator 182 outputs the CO filtered CO in module 44 LR can be passed, which is explained in more detail below.

[0133] CO LR Module 42 is the RVP waveform Features (and optionally PAP waveform The cardiac output is estimated from the autoencoder model 144 shown in Figures 12 to 15 and the CO AEThis is a different method of estimating cardiac output for the patient 16 compared to the method described above using module 40. This adds a level of redundancy within the hemodynamic monitoring system 10 for estimating cardiac output. LR Module 42 measures cardiac output (CO LR ) in Figs. 20A to 22 filtered As will be described in more detail below with reference to module 44, CO AE plus CO LR The CO described above can be used. AE Similar to module 40, another advantage here is that CO LR The aim is to enable module 42 to determine continuous cardiac output. LR CO is a continuous cardiac output estimate as it can be updated more frequently on a rolling basis such as every 10 seconds, or every 2 seconds. LR is a more continuous estimate compared to iCO, which is only measured intermittently. LR A continuous cardiac output estimate such as can capture transient or sudden changes in cardiac output of the patient 16, which improves patient care.

[0134] In addition, CO AE Similar to module 40, CO AE Module 40 is a CO AE CO calculated using LR Therefore, the relative RVP and the RVP waveform The CO LRThe time-based (e.g., heart rate, systolic time, etc.) and normalized pressure-based (e.g., systolic blood pressure, pulse pressure, etc.) variables used in module 42 are not affected by errors or biases in absolute pressure values. This is important because, in practice, absolute pressure values ​​are not always accurate. For example, medical personnel may introduce some measurement error in RVP measurements when leveling a pressure transducer for a patient. As another example, raising or lowering a patient's bed for surgery or in the ICU can also change absolute pressure values. LR The module 42 can be configured to not directly depend on absolute pressure values, so that CO LR Furthermore, CO LR Module 42 measures ΔCO, which may be independent and useful information about the patient 16. LR can also be calculated.

[0135] CO filtered Module (Figures 20A to 22) 20A and 20B show CO filtered 20A and 20B show variations of the Kalman filter algorithm of module 44. A first variation of the Kalman filter algorithm is configured to output a filtered estimate of cardiac output, as shown in FIG. 20A. A second variation of the Kalman filter algorithm is configured to output a filtered estimate of change in cardiac output, rather than cardiac output itself, as shown in FIG. 20B.

[0136] FIG. 20A illustrates a CO filtered 20A is a schematic block diagram of the module 44. As shown in FIG. filtered The module 44 includes a filter sub-module 184A including a prediction block 186A and an update block 188A, and a ΔCO AE Sub-module 189A and CO AE , CO LR , CCO, iCO, and ΔCOLR CO filtered To estimate CO filtered are input to module 44. More specifically, AE is ΔCO AE Input to submodule 189A, ΔCO AE Sub-module 189A measures the autoencoder change in cardiac output ("ΔCO AE ") to output the linear regression change in cardiac output, ΔCO LR and / or ΔCO AE is input to the prediction block 186A of the filter sub-module 184A. AE , CO LR , CCO, and iCO are input to the update block 188A of the filter sub-module 184A. filtered is output from the filter sub-module 184A.

[0137] The filter sub-module 184A is filtered This is a first variation of the Kalman filter algorithm of module 44. filtered (A second variation of the Kalman filter algorithm of module 44 is described below with reference to FIG. 20B.) Filter sub-module 184A includes instructions in code for implementing a Kalman filter algorithm that combines (or filters) one or more estimates of cardiac output into a single, more robust filtered estimate of cardiac output.

[0138] In general, a Kalman filter algorithm predicts the current state of input variables and can then update that prediction with additional information, such as from measurement inputs, to output a filtered estimate. Filter sub-module 184A is configured to predict the current state of cardiac output of patient 16. Filter sub-module 184A also consumes measurement inputs over time. As shown in FIG. 20A, the measurement inputs that can be consumed by filter sub-module 184A include CO2, an estimate of cardiac output estimated by flow and cardiac output software code 30, and the CO2. AE and CO LR and CCO and iCO, which are measurements obtained, for example, via catheter 54. Each of the measurement inputs may contain statistical noise and other inaccuracies, which can be represented as corresponding inaccuracies in the Kalman filter algorithm of filter sub-module 184A. Filter sub-module 184A filters CO filtered Output.

[0139] ΔCO AE Sub-module 189A is a CO filter associated with filter sub-module 184A. filtered It is an additional submodule of module 44. ΔCO AE Sub-module 189A generates a ΔCO for input to filter sub-module 184A. AE Calculate ΔCO AE Submodule 189A is a CO AE CO as input from module 40 AE Then, ΔCO AE Submodule 189A is a CO AE Outputs CO AE is consumed by the filter sub-module 184A in the prediction block 186A. For example, ΔCO AE Submodule 189A measures CO AE Then, ΔCO AE Submodule 189A is ΔCO AE To determine CO AEThe slope between the values ​​can be calculated. AE is calculated in different ways, CO LR ΔCO from module 42 LR may be different.

[0140] The Kalman filter algorithm of the filter sub-module 184A can be conceptualized as two distinct phases or steps, including a prediction phase and an update phase. The prediction block 186A represents the prediction phase of the filter sub-module 184A. As shown in FIG. 20A, ΔCO LR and / or ΔCO AE is input to the prediction block 186A. The prediction phase of the filter sub-module 184A is also described in more detail below with reference to FIG. 21A. The update block 188A represents the update phase of the filter sub-module 184A. As shown in FIG. 20A, CO AE and CO LR Measurement inputs including , CCO, and iCO are input to update block 188A. The update phase of filter sub-module 184A is also described in more detail below with reference to Figure 21B.

[0141] In prediction block 186A (i.e., in the prediction phase), the filter sub-module 184A filters past filtered estimates of cardiac output (e.g., CO , as described in more detail below) from a Kalman filter algorithm. filtered past values ​​of ) and ΔCO LR and / or ΔCO AE(or another prediction model, as described in more detail below) to predict cardiac output for the current time step, along with the corresponding prediction uncertainty, resulting in a predicted estimate of cardiac output. The predicted estimate of cardiac output generally does not directly include information from the current measurement input. However, the prediction phase of the Kalman filter algorithm of the filter submodule 184A can be initialized using one of the measurement inputs for the first iteration of the algorithm. For example, the prediction phase can be initialized using iCO measurements. After initialization, the prediction phase is performed using the previous filtered estimate, as described in more detail below. The Kalman filter algorithm can also be re-initialized based on predefined rules and triggers, such as the time elapsed since the previous initialization, the time elapsed between the two most recent consecutive measurements, etc.

[0142] In update block 188A (i.e., in the update phase), the filter sub-module 184A filters one or more of the measurement inputs (e.g., CO AE , CO LR , CCO, and iCO) and the corresponding measurement (or estimate) uncertainty for the current time step. As the filter sub-module 184A receives the measurement inputs, the predicted estimate of cardiac output is updated using a weighted average of the predicted and estimated values. Estimates with greater certainty and less uncertainty are given greater weight. The weighted average results in a filtered estimate of cardiac output that lies between the predicted estimate of cardiac output and the measurement input and has a better estimated uncertainty than either alone. This process is repeated at every time step (i.e., every iteration of the filter sub-module 184A's Kalman filter algorithm), and the filtered estimate of cardiac output and its uncertainty from the previous time step inform the prediction phase at the next iteration (i.e., the current time step).

[0143] Typically, prediction and update phases alternate, with the predicted estimate of cardiac output advancing at each time step until the next measurement input is received, after which the update phase incorporates the measurement input. That is, the Kalman filter algorithm of filter submodule 184A is configured such that each iteration of the Kalman filter algorithm includes a prediction phase and an update phase. However, if a measurement input is not available at a particular time step, the corresponding update phase can be skipped, and multiple prediction phases can be performed consecutively. Similarly, if multiple measurement inputs are received at the same time step (e.g., CO AE , CO LR , CCO, and iCO), multiple update phases can be performed sequentially to further refine the filtered estimate of cardiac output. Filter sub-module 184A can include update phases for any one source or any combination of sources of measurement inputs. In other words, filter sub-module 184A is configured such that the total number of update phases per iteration of the Kalman filter algorithm (which may be zero in some cases) depends on which of the measurement inputs are available during that iteration. As described in further detail below with reference to FIG. 21C, filter sub-module 184A generates a filtered estimate of cardiac output that is generally more accurate than any single one of the measurement inputs.

[0144] The Kalman filter algorithm of filter sub-module 184A is recursive. That is, filter sub-module 184A uses only the filtered estimate of cardiac output from the previous time step, the uncertainty associated with the previous filtered estimate of cardiac output, and the current measurement input to calculate the filtered estimate of cardiac output for the current time step; no other past information is required. For each time step, the previous filtered estimate of cardiac output is fed back to prediction block 186A to inform the next prediction of filter sub-module 184A. In this way, information about the previous measurement inputs and their respective uncertainties is still carried over to the next prediction, as the previous filtered estimate of cardiac output has been refined based on the measurement input from that time step.

[0145] The final filtered estimate of cardiac output produced by filter sub-module 184A at each time step is CO filtered The value of CO filtered is a filtered representation of right ventricular cardiac output ("RVCO"). filtered is the output of the filter sub-module 184A, and therefore CO filtered is one possible output of module 44. In some examples, CO filtered Module 44 transmits CO from filter sub-module 184A to output device 102, as shown in FIG. filtered For example, the output device 102 may output a CO filtered In some instances, a graph showing the values ​​of CO filtered The module 44 may include a CO filtered is continuously output to the output device 102.

[0146] FIG. 20B illustrates a CO filtered 20B is a schematic block diagram of the module 44. As shown in FIG.filtered The module 44 includes a filter sub-module 184B including a prediction block 186B and an update block 188B, and a ΔCO AE Sub-module 189B and CO AE and ΔCO LR is ΔCO filtered To estimate CO filtered are input to module 44. More specifically, AE is ΔCO AE is input to submodule 189B, and ΔCO AE Submodule 189B is ΔCO AE Output ΔCO LR and ΔCO AE is input to the update block 188B of the filter sub-module 184B. filtered is output from the filter sub-module 184B.

[0147] The filter sub-module 184B is filtered A second variation of the Kalman filter algorithm of module 44. Filter sub-module 184B includes instructions in code for implementing a Kalman filter algorithm that combines (or filters) one or more estimates of change in cardiac output into a single, more robust filtered estimate of change in cardiac output.

[0148] Filter sub-module 184B is generally similar in function to filter sub-module 184A described above with reference to Figure 20A, except that filter sub-module 184B is configured to filter multiple changes in cardiac output values ​​instead of multiple cardiac output values. For brevity, features and functions of filter sub-module 184B that are generally the same as filter sub-module 184A (shown in Figure 20A) will not be fully repeated in this section, except to the extent necessary to explain differences.

[0149] The filter sub-module 184B is configured to predict the current state of change in cardiac output of the patient 16. The filter sub-module 184B also consumes measurement inputs over time. As shown in FIG. 20B, the measurement inputs that may be consumed by the filter sub-module 184B are the flow and cardiac output, ΔCO, and the respective estimates of the change in cardiac output estimated by the cardiac output software code 30. AE and ΔCO LR Each of the measurement inputs may contain statistical noise and other inaccuracies, which can be represented as corresponding inaccuracies in the Kalman filter algorithm of the filter sub-module 184B. The filter sub-module 184B calculates ΔCO filtered Output.

[0150] ΔCO AE Sub-module 189B is a CO filter associated with filter sub-module 184B. filtered It is an additional sub-module of module 44. ΔCO AE As with submodule 189A, ΔCO AE Sub-module 189B generates a ΔCO for input to filter sub-module 184B. AE Calculate ΔCO AE Submodule 189B is a CO AE CO as input from module 40 AE Then, ΔCO AE Submodule 189B is ΔCO AE and ΔCO AE is consumed by the filter sub-module 184B in the update block 188B. For example, ΔCO AE Submodule 189B measures CO AE Then, the value ΔCO AE Submodule 189B is ΔCO AE To determine CO AE The slope between the values ​​can be calculated. AE is calculated in different ways, CO LRΔCO from module 42 LR may be different.

[0151] Prediction block 186B represents the prediction phase of filter sub-module 184B. In prediction block 186B (i.e., in the prediction phase), filter sub-module 184B uses a past filtered estimate of change in cardiac output from a Kalman filter algorithm (e.g., ΔCO, as described in more detail below). filtered The filter sub-module 184B predicts the change in cardiac output for the current time step based on the past values ​​of ΔCO, along with the corresponding prediction uncertainty, resulting in a predicted estimate of the change in cardiac output. The predicted estimate of the change in cardiac output generally does not directly include information from the current measurement inputs. However, the prediction phase of the Kalman filter algorithm of the filter sub-module 184B can be initialized using one of the measurement inputs for the first iteration of the algorithm. For example, the prediction phase can be initialized using one of the measurement inputs for the first iteration of the algorithm, such as ΔCO. AE or ΔCO LR After initialization, the prediction phase is performed using the previous filtered estimates, as described in more detail below. The Kalman filter algorithm can also be re-initialized based on predefined rules and triggers, such as the time elapsed since the previous initialization, the time elapsed between the last two consecutive measurements, etc.

[0152] The update block 188B represents the update phase of the filter sub-module 184B. As shown in FIG. AE and ΔCO LR The measurement inputs, including ΔCO and ΔCO, are input to an update block 188B. In the update block 188B (i.e., in the update phase), the filter sub-module 184B filters one or more of the measurement inputs (e.g., ΔCO AE and ΔCO LR) and the corresponding uncertainty of the measurement (or estimate) for the current time step. As filter sub-module 184B receives measurement inputs, the predicted estimate of the change in cardiac output is updated using a weighted average of the predicted and estimated values. Estimates with greater certainty and less uncertainty are given greater weight. The weighted average results in a filtered estimate of the change in cardiac output that lies between the predicted estimate of the change in cardiac output and the measurement input and has a better estimated uncertainty than either alone. This process is repeated at every time step (i.e., every iteration of the Kalman filter algorithm of filter sub-module 184B), and the filtered estimate of the change in cardiac output and its uncertainty from the previous time step inform the prediction phase at the next iteration (i.e., the current time step). Similar to filter sub-module 184A described above with reference to FIG. 20A, the prediction and update phases of filter sub-module 184B can alternate, or in other examples, the update phase can be skipped or repeated depending on the availability of measurement inputs. The filter sub-module 184B produces a filtered estimate of the change in cardiac output that is generally more accurate than any one of the measurement inputs.

[0153] The Kalman filter algorithm of filter sub-module 184B is recursive. That is, filter sub-module 184B uses only the filtered estimate of change in cardiac output from the previous time step, the uncertainty associated with the previous filtered estimate of change in cardiac output, and the current measurement input to calculate a filtered estimate of change in cardiac output for the current time step. For each time step, the previous filtered estimate of change in cardiac output is fed back to prediction block 186B to inform the next prediction of filter sub-module 184B. In this way, information about the previous measurement inputs and their respective uncertainties is still carried over to the next prediction, as the previous filtered estimate of change in cardiac output has been refined based on the measurement input from that time step.

[0154] The final filtered estimate of the change in cardiac output produced by the filter sub-module 184B at each time step is ΔCO filtered is the value of the filtered change in cardiac output, ΔCO filtered is the output of the filter sub-module 184B, and therefore CO filtered is another possible output of module 44. In some examples, CO filtered Module 44 transmits ΔCO from filter sub-module 184B to output device 102 as shown in FIG. filtered For example, the output device 102 may output a ΔCO filtered In some instances, a graph showing the values ​​of CO filtered The module 44 monitors the hemodynamic status of the patient 16 by measuring ΔCO filtered is continuously output to the output device 102.

[0155] As shown as separate examples in FIGS. 20A-20B, CO filteredModule 44 can be configured to include either or both of filter sub-module 184A and filter sub-module 184B. filtered Module 44 includes only filter sub-module 184A and is therefore configured to filter cardiac output values. filtered Module 44 includes only filter sub-module 184B and is therefore configured to filter changes in cardiac output values. filtered Module 44 includes a filter sub-module 184A and a filter sub-module 184B and is therefore configured to filter both cardiac output values ​​and changes in cardiac output.

[0156] 20A and 20B together, the Kalman filter algorithm of filter sub-module 184A and filter sub-module 184B is a powerful estimator when there are multiple noisy estimates and / or measurements of variables such as cardiac output or change in cardiac output. The properties of the Kalman filter algorithm, including a prediction phase and an update phase, allow filter sub-module 184A and filter sub-module 184B to effectively handle uncertainty due to noisy measurement inputs and provide a more accurate CO filtered and ΔCO filtered Additionally, filter sub-module 184A and filter sub-module 184B may be configured to filter and / or measure cardiac output and changes in cardiac output received from various sources to generate more accurate CO2 estimates and / or measurements. filtered and ΔCO filtered The method may operate on these estimates and / or measurements to integrate the

[0157] Filter sub-module 184A and filter sub-module 184B may also tolerate receiving asynchronous or sporadic estimates and / or measurements of cardiac output or changes in cardiac output from various sources. That is, filter sub-module 184A and filter sub-module 184B may not require measurement input from all possible sources at each time step, and may even tolerate not receiving any measurement input at all for a particular time step. For example, filter sub-module 184A may be configured to receive a CO AE Module 40 and CO LR 12) is no longer producing good values, in which case the flow module 38 may determine that the autoencoder model 144 (shown in FIG. 12) is no longer producing good values. AE is not calculated for that time step and is not provided to the filter sub-module 184A. In such an example, the Kalman filter algorithm of the filter sub-module 184A automatically adapts to this change and CO LR Only measurement input from module 42 is used to measure CO filtered The Kalman filter algorithm can continue to generate CO AE Measurement input from module 40 and CO LR The hemodynamic monitoring system 10 may also be automatically re-adapted to use both the measured input from the hemodynamic monitoring module 10 and the measured input from the hemodynamic monitoring module 10. Similarly, if additional sources of estimates and / or measurements of cardiac output or changes in cardiac output (e.g., measurements from additional or different devices such as echocardiograms, or estimates from other models) are added to the hemodynamic monitoring system 10, these can be easily integrated into the Kalman filter algorithm as additional update phases.

[0158] The use of the Kalman filter algorithm in filter sub-module 184A and filter sub-module 184B is very computationally efficient because the Kalman filter algorithm is recursive. Similarly, the recursive nature of the Kalman filter algorithm means that it does not integrate information in isolation. Instead, the Kalman filter algorithm allows filter sub-module 184A and filter sub-module 184B to intelligently combine information using the context of what the filtered estimate of cardiac output or change in cardiac output was at the previous time step. In this way, the CO filtered and ΔCO filtered can be a more accurate estimate of the cardiac output of the patient 16. Furthermore, the CO AE Module 40 and CO LR Similar to module 42, another advantage here is that the CO filtered The module 44 allows for the determination of continuous cardiac output. filtered is a continuous cardiac output estimate, as it can be updated more frequently on a rolling basis, such as every 10 seconds, or every 2 seconds. filtered is a more continuous estimate compared to iCO, which is only measured intermittently. filtered A continuous cardiac output estimate such as can capture transient or sudden changes in cardiac output of the patient 16, which improves patient care.

[0159] In addition, CO AE Module 40 and CO LR Similar to module 42, CO filtered Module 44 is CO AE CO based on filtered Therefore, the relative RVP and the RVP waveform It can also be based on the form or shape of CO LR CO based on filteredIt can be described as "data independent" with respect to absolute pressure values ​​(e.g., absolute values ​​of RVP) because the absolute pressure values ​​may similarly be unaffected by the absolute pressure values. This is important because, in practice, absolute pressure values ​​are not always accurate. For example, healthcare professionals may introduce some measurement error in RVP measurements when leveling a pressure transducer for a patient, and as another example, raising or lowering a patient's bed for surgery or in an ICU may also change the absolute pressure values. filtered The module 44 can be configured to not depend directly on absolute pressure values, so that CO filtered can be generated accurately.

[0160] Figure 21A shows CO filtered 21B is a graph showing the predicted value p of module 44. filtered 21C is a graph showing the measured value m of the module 44. filtered 21A-21C are graphs illustrating filtered values ​​f over time from filter sub-module 184A (shown in FIG. 20A ), each showing cardiac output values ​​over time. However, it should be understood that the concepts described with respect to FIGS. 21A-21C are also generally applicable to the function of filter sub-module 184B (shown in FIG. 20B ), and that the graphs in FIGS. 21A-21C could instead represent changes in cardiac output values ​​over time.

[0161] Each of the graphs in FIGS. 21A-21C illustrates the same arbitrary time steps n, n-1, n-2, n-3, and n-4 along the x-axis. Each of time steps n, n-1, n-2, n-3, and n-4 corresponds to a respective iteration of the Kalman filter algorithm of the filter sub-module 184A. That is, time step n represents the current iteration of the filter sub-module 184A. Time steps n-1, n-2, n-3, and n-4 represent time steps preceding time step n. For example, time step n-1 represents the iteration immediately preceding time step n, time step n-2 represents the iteration immediately preceding time step n-1 (and two iterations before time step n), and so on. As described in more detail below, each of the graphs in FIGS. 21A-21C also illustrates cardiac output values ​​corresponding to time steps n, n-1, n-2, n-3, and n-4. The graphed cardiac output values ​​in FIGS. 21A-21C are shown in units of liters per minute (L / min), but in other examples can be expressed in other rate units.

[0162] 21A is a graph corresponding to the function of prediction block 186A shown in FIG. 20A. In FIG. 21A, cardiac output values ​​are shown for each of time steps n, n-1, n-2, n-3, and n-4. Predicted value p is the cardiac output value corresponding to the current time step n. Predicted value p is an exemplary value of a predicted estimate of cardiac output predicted in prediction block 186A. The cardiac output values ​​corresponding to time steps n, n-1, n-2, n-3, and n-4 can be used at the current time step n (or used at a previous time step) by prediction block 186A to generate predicted value p corresponding to the current time step n. filtered, can be the previous value of p. If any of time steps n, n-1, n-2, n-3, and n-4 represents the first time step at which the predicted value p was calculated, the cardiac output value at the previous time step relative to the first time step can instead be the measured input value used for initialization. The predicted value p at the current time step n is shown connected to the previous value by a dashed line in FIG. 21A because the predicted value p is a predicted estimate of cardiac output used internally within filter submodule 184A. Unlike the other cardiac output values ​​shown in FIG. 21A for previous time steps, the predicted value p is generally not the final filtered output of filter submodule 184A.

[0163] As shown in Figure 21A, the predicted value p is expressed as a function of the corresponding uncertainty e p The uncertainty e p represents the uncertainty associated with the predicted value p as an estimate of the standard deviation. The uncertainty in the predicted value p may be configurable or predetermined for each application of the filter sub-module 184A and may depend on the type of prediction used. The arrow extending from the predicted value p is proportional to the magnitude of the uncertainty in the predicted value p. In one non-limiting example, the predicted value p has an associated uncertainty e of ±0.4 L / min. p Therefore, if the predicted value p is 4.0 L / min, the uncertainty e p The range indicated by the arrows is 3.6 to 4.4 L / min, indicating that there is approximately a 68% probability that the true value falls within this range.

[0164] The prediction phase of the Kalman filter algorithm of the filter sub-module 184A can be configured as one of several different prediction models for generating the predicted value p. In one example, the prediction phase can use an extrapolation method to calculate the predicted value p. For example, the prediction phase can use a linear extrapolation method based on the cardiac output values ​​at time steps n-1 and n-2 to determine the value of the predicted value p at the current time step n. Using linear extrapolation in the prediction phase means that the change (i.e., slope) in cardiac output between two previous time steps is carried over to the current time step n. That is, the slope between the cardiac output values ​​between time steps n-2 and n-1 is considered to be the same between time steps n-1 and n. In other examples, the prediction phase can use extrapolation techniques by using other types of fit functions, such as a cubic fit, and incorporating additional past cardiac output values ​​(e.g., from time steps n-3, n-4, etc.). For example, the prediction block 186A can use a CO filtered In yet another example, such as an example where the system is stable, the prediction block 186A can predict the CO filtered The predicted value p at the current time n is calculated using the value of CO at time step n-1. filtered can be extrapolated from the value of

[0165] Alternatively, the prediction phase may use ΔCO to calculate the predicted value p. LR and / or ΔCO AE For example, the prediction block 186A can use the CO filtered Starting with the value of ΔCO for the current time step n, we use LR or ΔCO AEIn another example, the prediction block 186A may add the value of CO at time step n−1 to determine the predicted value p at the current time step n. filtered to the value of ΔCO for the current time step n. LR and ΔCO AE the average (e.g., ΔCO LR and ΔCO AE In yet another example, the prediction block 186A may add a weighted average based on specified or predetermined weights, such as 0.5 and 0.5 for CO at time step n-1. filtered the value of ΔCO at the current time step n LR Determine a first predicted value by adding the value of CO at time step n-1 filtered the value of ΔCO at the current time step n AE A second predicted value may be determined by adding the values ​​of ΔCO and then determining a predicted value p as the average (e.g., weighted average) of the first and second predicted values. Further details and examples of weighting the inputs are described below with reference to FIG. 21B and are also applicable here. In one non-limiting example, ΔCO LR and ΔCO AE The specified or predetermined weights associated with ΔCO can be based on the relative accuracy (or uncertainty) associated with the two models (i.e., the regression model 180 shown in FIGS. 18-19 and the autoencoder model 144 shown in FIGS. 12-15). LR and ΔCO AE The weights associated with may vary over time depending on the relative signal quality metrics associated with such methods. Compared to extrapolation methods, these predictive models use the CO of the immediately preceding time step (e.g., time step n-1) rather than one or more previous time steps. filtered Instead of extrapolation, we only need the value of ΔCO LR and / or ΔCO AEThe trade-off of using ΔCO is that the quality of the prediction depends to some extent on the capabilities of the regression model 180 and the autoencoder model 144. For example, as the reference features used in the regression model 180 become outdated, ΔCO LR The quality of the RVP may be reduced. features_ref and PAP features_ref If ΔCO is updated more recently, LR Since prediction block 186B does not receive any change in cardiac output value as input, prediction using ΔCO LR and / or ΔCO AE It should be noted that the prediction model using ΔCO is generally only used in the prediction block 186A of the filter sub-module 184A (shown in FIG. 20A), and not in the prediction block 186B of the filter sub-module 184B (shown in FIG. 20B). That is, the prediction block 186B makes predictions using extrapolation, while the prediction block 186A either uses extrapolation or uses ΔCO. LR and / or ΔCO AE can be used to make predictions.

[0166] Each of the predictive models described above uses extrapolation or ΔCO to make predictions regarding the current value of cardiac output of the patient 16. LR Furthermore, each of the predictive models is based on the fundamental assumption that the change in cardiac output within a measurement time window (e.g., a 10-second interval) can be represented by a single aggregate or average value.

[0167] FIG. 21B is a graph corresponding to the function of update block 188A shown in FIG. 20A. In FIG. 21B, cardiac output values ​​are shown for each of time steps n, n-1, n-2, n-3, and n-4. Measurement m is the cardiac output value corresponding to the current time step n. Measurement m is an exemplary value of the measurement input received by filter sub-module 184A in update block 188A. As explained above, the measurement input may be CO AE , CO LR, CCO, or iCO. The measurement m is used by update block 188A to update the predicted value p for the current time step n. The cardiac output values ​​corresponding to time steps n, n-1, n-2, n-3, and n-4 are previous measurement input values ​​used at previous time steps by update block 188A to update the previous predicted estimate of cardiac output for the respective time step.

[0168] Figure 21B shows CO AE Module 40, CO LR It represents cardiac output values ​​from a single source, such as one of module 42, CCO, or iCO. That is, measurement m represents the CO AE , CO LR , CCO, or iCO. In this example, the Kalman filter algorithm of filter sub-module 184A has only one update phase that uses measurement m. In other examples where cardiac output values ​​are received from multiple sources, there are multiple measurements m at the current time step n.

[0169] As shown in Figure 21B, a measurement m is expressed as a function of the corresponding uncertainty e m The uncertainty e m represents the uncertainty in measurement m as an estimate of the standard deviation. The uncertainty in measurement m may be configurable or predetermined for each application of the filter sub-module 184A and may depend on the type of measurement input used. The arrow extending from measurement m is proportional to the magnitude of the uncertainty in measurement m. In one non-limiting example, CO AE CO from module 40 AE is the associated uncertainty e of 0.5 L / min m Since there is a possibility that AE If the measured value m is 5.0 L / min, the uncertainty e mThe range indicated by the arrows for is 4.5 to 5.5 L / min, indicating that there is approximately a 68% probability that the true value lies within this range. m depends on how each uncertainty is determined. p It may be the same as or different from.

[0170] In the update phase of the Kalman filter algorithm of the filter submodule 184A, the predicted values ​​p and the measured values ​​m are dynamically or adaptively weighted to produce a filtered value f, as shown in Figure 21C. This weighting is based on the uncertainty e p or uncertainty e m This can result in an uncertainty associated with the filtered value f at the current time step n being predicted to be equal to or less than either

[0171] When the Kalman filter algorithm of the filter submodule 184A is initialized, the uncertainty e p and uncertainty e m Multiple sources of measurement input (e.g., CO AE Module 40, CO LR If modules 42, CCO, and / or iCO are present, each source will have a different associated uncertainty e m The uncertainty e of the predicted value p p is the uncertainty associated with the prediction model and the CO used to determine the predicted value p filtered and any uncertainty associated with past values ​​of p. Thus, after the filter sub-module 184A determines a predicted value p for the current time step n and receives a measurement m, the filter sub-module 184A may filter the corresponding uncertainty e p and uncertainty e m The predicted value p and the measured value m each have information about their associated uncertainty e p and e mThe predicted value p and the measured value m with greater uncertainty are given less weight, and the predicted value p and the measured value m with less uncertainty are given more weight. Using the example above to illustrate, the uncertainty e p is 0.4 L / min, and the uncertainty e m was 0.5 L / min, the predicted value p is given more weight and the measured value m is given less weight.

[0172] Uncertainty p and uncertainty e m Each of the uncertainties e can be fixed or modulated over time. m is the SQI as explained above with reference to Figure 9. RVP or SQI PAP To illustrate, the SQI RVP If the value of is between zero and one (indicating signal quality within the specified range, 0 to 1), the flow and cardiac output software code 30 (shown in Figure 1) calculates the corresponding RVP waveform Use data to CO AE and CO LR However, the COAE and COLR associated with a signal quality index of zero and the COAE and COLR associated with a signal quality index of one can be estimated. AE and CO LR and (SQI RVP =0 and SQI RVP =1), both are valid but may have different uncertainties. For example, SQI RVP = 0, the uncertainty in the measurement may be low, so the uncertainty e associated with the measurement m m can be multiplied by 1 (or left unchanged). In the same example, SQI RVP = 1, the uncertainty in the measurement may be high, so to increase the uncertainty, we use the uncertainty e associated with the measurement m m can be multiplied by some factor, such as 2 or more. In another example, the uncertainty ep and / or uncertainty m is SQI RVP and SQI PAP It can be scaled based on a factor other than .

[0173] FIG. 21C is a graph corresponding to the functioning of filter sub-module 184A using information from prediction block 186A and update block 188A, as shown in FIG. 20A and described above with reference to FIGS. 21A-21B. In FIG. 21C, cardiac output values ​​are shown for each of time steps n, n-1, n-2, n-3, and n-4. The filtered value f is the cardiac output value corresponding to the current time step n. The filtered value f is the CO2 output by filter sub-module 184A for the current time step n. filtered The filtered value f is also used by prediction block 186A in prediction for the next time step. The cardiac output values ​​corresponding to time steps n, n-1, n-2, n-3, and n-4 are filtered by the previous CO output values ​​output by filter sub-module 184A in the previous time steps. filtered value.

[0174] The filtered value f, and its associated uncertainty e f is the predicted value p, the measured value m, and the associated uncertainty e p and e m and are calculated according to the following general form:

[0175]

number

[0176] where: e f ≦e m , e f ≦e p , and 1 / e f 2 =(1 / ep 2 )+(1 / e m 2 ) (Equation 4) Furthermore, Equation 3 can be simplified and rewritten to explicitly show the update of the prediction p using the measurement m (i.e., the measurement input in the Kalman filter algorithm).

[0177]

number

[0178] Equation 5 is f=p+(K×(mp)) (Equation 6) It can also be expressed as: where K is

[0179]

number

[0180] and e f ≦e m , and e f ≦e p Two different measurement input sources at once (e.g. CO LR and CO AE To extend Equation 3 to show two simultaneous measurements m from

[0181]

number

[0182] can be rewritten as, where m_1 is CO AE , CO LR a first measurement from a first source of measurement input, such as one of: m_2 is CO AE , CO LRa second measurement from a second source of measurement input, such as a different one of e f ≦e m_1 , e f ≦e m_2 , e f ≦e p , and 1 / e f 2 =(1 / e p 2 )+(1 / e m_1 2 ) +(1 / e m_2 2 ) (Formula 8) is.

[0183] As shown in Equation 3, the predicted value p is weighted by the respective contribution of the measured value m to the overall noise, and the measured value m is weighted by the respective contribution of the predicted value p to the overall noise. In other words, if either the predicted value p or the measured value m has a larger uncertainty, the other value is weighted more heavily. To illustrate, let us consider the case where the predicted value p has an uncertainty e p = 0 (no uncertainty), Equation 3 assumes that the predicted value p is 1(e m / e m ), and the measurement m is weighted by zero (0 / e m ) is simplified to be weighted by ∑ m = ∑ p = ...

[0184] Equation 7 shows how the filtered value f is calculated when there are two simultaneous measurements m, such as from two different sources of measurement input. Equation 7 can be further extended to include any number of measurements m (e.g., m_1, m_2, ... m_n). For example, the filtered value f can be calculated by AE and CO LRIt can be calculated using measurements m from each of the sources of measurement input available in the system, including , CCO, and iCO.

[0185] As shown in Figure 21C, the filtered value f is calculated using the corresponding uncertainty e f The uncertainty e f represents the uncertainty in the filtered value f. The arrows extending from the filtered value f are proportional to the magnitude of the uncertainty in the filtered value f. In an example involving a single measurement m, the uncertainty e f can be determined using Equation 4 above. In an example involving multiple measurements m, the uncertainty e f can be determined using Equation 8 above. As explained above, the uncertainty e f is ideally the individual uncertainties e p and uncertainty e m Therefore, the uncertainty e f The arrows indicating the uncertainty e p and uncertainty e m The arrows indicating uncertainty e f is the uncertainty e p and uncertainty e m Since the filtered value f is smaller than the predicted value p or the measured value m alone, CO filtered Module 44 calculates the CO for the current time step n. AE or CO LR CO may be more accurate than either filtered can be output.

[0186] Figure 22 shows the CO AE Module 40, CO LR Module 42, and CO filtered 1 is a graph comparing exemplary cardiac output values ​​from module 44 with exemplary iCO values ​​over time. AE Module 40, CO LR Module 42, COfiltered Module 44 and iCO are each different sources of cardiac output within the hemodynamic monitoring system 10 shown in FIG. LR CO from module 42 LR values ​​(shown as open squares), CO AE CO from module 40 AE values ​​(shown as black circles), CO filtered CO from module 44 filtered values ​​(shown as solid lines), and iCO values ​​(shown as black asterisks) are shown over time.

[0187] As shown in FIG. 22 , the graph of exemplary cardiac output values ​​is divided into three sections, including a first section 190, a second section 192, and a third section 194. The first, second, and third sections 190, 192, and 194 are not connected to one another because they are separated by an arbitrary amount of time, represented by the space along the x-axis between the first section 190 and the second section 192, and between the second section 192 and the third section 194. The first section 190 corresponds to the pre-bypass condition of the patient 16. The first section 190 includes exemplary cardiac output values ​​measured or obtained for the patient 16 before bypass surgery. The second section 192 corresponds to the post-bypass condition of the patient 16. The second section 192 includes exemplary cardiac output values ​​measured or obtained for the patient 16 after bypass surgery. The third section 194 corresponds to the ICU condition of the patient 16. The third portion 194 includes exemplary cardiac output values ​​measured or obtained for the patient 16 during the period the patient 16 was hospitalized in an ICU setting.

[0188] Figure 22 shows the CO filtered CO from module 44 filtered The value is CO LR CO from module 42 LR Value, CO AE CO from module 40 AE values ​​and iCO values, which are more continuous and less noisy. LR , CO AEThe values ​​of , and iCO may be synchronous, asynchronous, tend to be in the same direction, and / or tend to be in different directions at each point in time.

[0189] For example, in one region slightly before the midpoint along second portion 192 (viewed from left to right), LR and CO AE The values ​​of CO tend to diverge or are in opposite directions. Over the remainder of second portion 192 (as well as most of first portion 190 and third portion 194). LR and CO AE As shown in Figure 22, CO filtered The value of CO LR and CO AE This is between the values ​​that deviate from CO filtered Module 44 intelligently weights the estimate of cardiac output and CO filtered This is because we use the Kalman filter algorithm to integrate CO filtered The values ​​are less noisy than the other cardiac output values ​​shown in FIG.

[0190] As another example, CO LR , CO AE , and there are some time points where no value for iCO is received. LR , CO AE 22 where one or more of iCO, iCO, and iCO are missing represent asynchronous cardiac output estimation activity from various sources in hemodynamic monitoring system 10. However, filtered At these times, CO filtered Still available from module 44. filtered Module 44 is generally always CO filtered CO filtered The graph of is more continuous than for the other cardiac output sources shown in FIG.

[0191] Exemplary Implementations and Alternatives (Figure 23) Many different configurations of hemodynamic monitoring system 10 are possible using combinations of the components and modules described above with reference to FIGS. 1-22. That is, hemodynamic monitoring system 10 according to the techniques described herein can be implemented in many different ways, including using various combinations of hemodynamic sensors 14 to acquire hemodynamic data about patient 16 (shown in FIG. 1). Accordingly, some of the inputs and outputs shown in FIG. 6 are optional. Depending on the hemodynamic data available from patient 16 or the desired output, as shown in FIG. 1, RVP feature module 32, PAP feature module 34, verification module 36, flow module 38, CO2 feature module 39, and other modules may be implemented to estimate the patient's blood flow, cardiac output, and / or changes in cardiac output. AE Module 40, CO LR Module 42, and CO filtered Various combinations of modules 44 can be implemented together.

[0192] For example, PAP waveform One or more of PAP, CCO, and iCO may not be available to the patient 16 or may be ignored for other reasons. waveform In instances where PAP is not available or is not used, the PAP feature module 34 is also not used and PAP features is the verification module 36 or CO LR Instead, only the RVP feature module 32 is used, and the verification module 36 and CO LR Only module 42 is RVP features In instances where CCO and iCO are not available or used, the CCO and iCO measurements are received. LR Module 42 or CO filtered It is not passed to module 44. Instead, LR Module 42 is CO AE CO from module 40 AEUse only CO filtered Module 44 is CO AE , CO LR , and ΔCO LR Use some combination of:

[0193] Furthermore, the estimated cardiac output for patient 16 was CO AE Module 40, CO LR Module 42, and CO filtered The estimated flow may be output from any one or more of the modules 44. processed CO AE To estimate CO AE In addition to or instead of being passed to module 40, the estimated change in cardiac output can be output from flow module 38. Similarly, the estimated change in cardiac output can be calculated using the ΔCO filtered To estimate CO filtered In addition to or instead of being passed to module 44, LR It can be output from module 42.

[0194] 23 is a schematic block diagram illustrating the inputs and outputs for each module of an exemplary implementation of the hemodynamic monitoring system 10. FIG. 23 illustrates the RVP feature module 202, the validation module 204, the flow module 206, and the CO AE Module 208 and CO LR Module 210 and CO filtered The RVP feature module 202, the verification module 204, the flow rate module 206, the CO AE Module 208, CO LR Module 210, CO filtered Each of the modules 212 and output devices 214 is generally similar to the components or modules of the same name described above with reference to Figures 1-22, and one possible configuration of these components and modules is described here.

[0195] As shown in FIG. 23, the RVP feature module 202waveform The RVP feature module 202 receives as input the RVP waveform From RVP features The RVP feature module 202 extracts the RVP features to the verification module 204. The RVP feature module 202 features CO LR It is also output to module 210.

[0196] The verification module 204 is waveform and RVP features and RVP as input. features is provided from the RVP feature module 202 to the validation module 204. The validation module 204 determines whether RVP_valid is true or negative (i.e., RVP waveform The RVP_valid parameter is set to true (e.g., RVP_valid=true) and outputs an indication to the flow module 206 indicating whether the RVP_valid parameter is valid or not. waveform means that the hemodynamic data corresponding to RVP_valid is valid for further processing by the flow module 206. An indication that RVP_valid is false (e.g., RVP_valid=false) indicates that RVP waveform This means that the hemodynamic data corresponding to the smoothed RVP is not valid for further processing. waveform to the flow module 206.

[0197] The flow module 206 receives the smoothed RVP from the validation module 204. waveform as input. The flow module 206 determines whether flow_valid is true or negative (i.e., flow processed CO AE The flow_valid flag is output to the module 208. An indication that flow_valid is true (e.g., flow_valid=true) indicates that the estimated blood flow waveform flow processed CO AEIt means that the estimated blood flow waveform flow is valid for further processing by the module 208. An indication that flow_valid is false (e.g., flow_valid=false) indicates that the estimated blood flow waveform flow processed This means that the flow is not available for further processing. processed CO AE The flow module 206 also outputs to the flow processed to the processing device 214.

[0198] CO AE The module 208 is a flow processed as input. processed is the CO from the flow module 206 AE The CO is supplied to module 208. AE The module 208 also determines whether flow_valid is true or false and accordingly AE The flow module 206 receives an indication of whether the flow module 208 can proceed. AE Module 208 is a CO AE CO LR Module 210 and CO filtered Output to module 212.

[0199] CO LR The module 210 is an RVP features and CO AE and RVP as input. features RVP feature module 202 to CO LR It is supplied to module 210. AE CO AE Module 208 to CO LR It is supplied to module 210. LR Module 210 is a CO LR and ΔCO LR and CO filtered Output to module 212.

[0200] CO filtered Module 212 is a CO AE, CO LR , and ΔCO LR It receives as input CO AE CO AE Module 208 to CO filtered The CO LR and ΔCO LR CO LR Module 210 to CO filtered The CO filtered Module 212 is a CO filtered is output to the output device 214.

[0201] As shown in FIG. 23, the output device 214 includes a flow module 206 and a CO filtered 2. More specifically, the output device 214 receives the output from the flow module 212. processed and CO filtered Each of these outputs can be displayed (e.g., via display 24 as shown in FIG. 1) as a corresponding graph showing values ​​over time.

[0202] Any of the various systems, devices, apparatus, etc. in the present disclosure can be sterilized (e.g., with heat, radiation, ethylene oxide, hydrogen peroxide, etc.) to ensure they are safe for patient use, and the methods herein can include sterilization (e.g., with heat, radiation, ethylene oxide, hydrogen peroxide, etc.) of the associated systems, devices, apparatus, etc.

[0203] The treatment techniques, methods, steps, etc. described or suggested in this specification or the references incorporated herein may be performed on living animals or non-living simulations such as cadavers, cadaver hearts, anthropomorphic ghosts, simulators (e.g., in which body parts, tissues, etc. are simulated), etc.

[0204] Although the present invention has been described with reference to illustrative examples, it will be understood by those skilled in the art that various modifications can be made and equivalents can be substituted for elements of the present invention without departing from the scope of the invention. In addition, many modifications can be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope of the invention. Therefore, it is not intended that the invention be limited to the particular examples disclosed, but rather that the invention will include all examples falling within the scope of the appended claims. [Explanation of symbols]

[0205] 10 Hemodynamic Monitoring System 12 Hemodynamic monitor 14 Hemodynamic Sensor 14A hemodynamic sensor, second pressure transducer, invasive hemodynamic sensor, pressure sensing hemodynamic sensor, pressure transducer sensor, first pressure transducer 14B Hemodynamic sensor, module, oxygen saturation measurement module, invasive hemodynamic sensor 14C Hemodynamic Sensor, Implantable Hemodynamic Sensor 14D Hemodynamic Sensor, Implantable Hemodynamic Sensor 16 patients 18. Healthcare workers 20 System Processor 22 System Memory, Computer-Readable System Memory 24 displays 26 Analog-to-Digital Converter (ADC), ADC 28 Digital-to-Analog Converter (DAC), DAC 30 Flow and Cardiac Output Software Code 32 RVP feature modules 34 PAP feature modules 36 Verification Module 38 Flow Module 40 CO AE Module 42 CO LR Module 44 COfiltered Module 46 User Interface 48 Control Elements 50 Sensory Alarms 52 Input and / or Output (I / O) Connector, I / O Connector 54 Catheter 56 Sheath 58 lumens 60 Fluid Connector 60A Distal Port Connector 60B Proximal Infusion Connector 60C Right Ventricular Pacing Connector 60D Balloon Connector 62 Optical Connector 64 Thermistor Connector 66 Thermal Filament Connector 68 ports 68A port, distal port 68B Proximal Port, Proximal Infusion Port 68C Right ventricular port 70 Balloon 72 Tip 74 Syringe 76 Housing 78 Fluid Input Port 80 Catheter fluid port 82 I / O cable 84 I / O connectors 86 Housing 88 Protective Door 90 Cable 92 Connectors 102 Output Devices 104 RVP waveform trace 106 indicators 108 Index, End-diastolic Blood Pressure 108E End-diastolic blood pressure excluded 110 Index, Maximum Systolic Blood Pressure 110E Maximum systolic blood pressure excluded 112 indicators 114 indicators 116 indicators 118 indicators 120 indicators 122 PAP waveform traces 124 indicators 126 Index, Maximum Systolic Blood Pressure 128 indicators 130 indicators 132 indicators 134 Signal Quality Detector 136 Decision Block 137 RVP waveform trace 138 Index, Pulse Transit Time 140 Index, Systolic Slope 142 Index, mean blood pressure 144 Autoencoder Model, Animal-Trained Autoencoder Model 146 Flow Filter 148 Decision Block 150 inputs 152 Filter, filtering layer 154 Latent space 156 Filter, Filtering Layer 158 Output 160 Encoder 162 decoder 180 Regression Models 182 Cardiac Output Estimator 184A Filter Submodule 184B Filter Submodule 186A Prediction Block 186B Prediction Block 188A Update Block 188B Update Block 189A ΔCO AE Submodules 189B ΔCO AE Submodules 190 First Part 192 Second Part 194 Third Part 202 RVP feature module 204 Verification Module 206 Flow Module 208 COAE Module 210 CO LR Module 212 CO filtered Module 214 Output Devices

Claims

1. 1. A system for determining catheter placement in a patient, comprising: a first hemodynamic sensor that continuously generates a first hemodynamic sensor signal representative of a right ventricular pressure waveform of the patient, the catheter connected to the first hemodynamic sensor; The display and one or more processors; When executed by the one or more processors, the system: receiving the first hemodynamic sensor signal representative of the right ventricular pressure waveform of the patient; extracting features from the right ventricular pressure waveform of the patient; comparing the features extracted from the patient's right ventricular pressure waveform with the one or more specified value ranges of the features to determine that the catheter is correctly placed if the value of the features extracted from the patient's right ventricular pressure waveform is within one or more specified value ranges, or to determine that the catheter is not correctly placed in the patient or properly connected to the system, or that a signal quality problem is occurring, if the value of the features extracted from the patient's right ventricular pressure waveform is not within the one or more specified value ranges; causing the display or module to output an indication of whether the catheter is properly positioned within the patient based on whether the value of the feature extracted from the right ventricular pressure waveform of the patient is within the one or more specified value ranges. a computer-readable memory encoded with instructions; A system comprising:

2. 2. The system of claim 1, wherein the one or more specified value ranges of the features correspond to physiological limits for each of the features extracted from the right ventricular pressure waveform of the patient plus a standard deviation for each feature.

3. The system of claim 1 , wherein the specified value range is used to identify noise, an under-damped signal, or an over-damped signal in the right ventricular pressure waveform of the patient.

4. The system of claim 1 , wherein the system continuously monitors the placement of the catheter.

5. The system of claim 1 , wherein a determination of correct catheter placement indicates that the right ventricular pressure waveform is of high quality.

6. 2. The system of claim 1, wherein the instructions, when executed by the one or more processors, further cause the system to filter data from the right ventricular pressure waveform of the patient by excluding data from a portion of the right ventricular pressure waveform that includes the features extracted from the right ventricular pressure waveform of the patient that are not within the one or more specified value ranges.

7. 2. The system of claim 1, wherein the features extracted from the right ventricular pressure waveform of the patient include one or more of end-diastolic blood pressure, maximum systolic blood pressure, minimum diastolic blood pressure, end-systolic blood pressure, maximum rate of change of blood pressure over time, minimum rate of change of blood pressure over time, diastolic gradient, and right ventricular pulse pressure.

8. 2. The system of claim 1, wherein the instructions, when executed by the one or more processors, further cause the system to assign a signal quality index to the right ventricular pressure waveform based on whether the value of the feature extracted from the right ventricular pressure waveform of the patient is within the one or more specified value ranges, the signal quality index indicating a quality of the right ventricular pressure waveform.

9. 9. The system of claim 8, wherein the signal quality index is assigned to the right ventricular pressure waveform based on the values ​​of the features from the right ventricular pressure waveform at 10 second intervals.

10. 9. The system of claim 8, wherein the signal quality index ranges from zero to five, with zero being the highest quality and five being the lowest quality, and the signal quality index is based on how close the features extracted from the patient's right ventricular pressure waveform are to the one or more specified value ranges.

11. 9. The system of claim 8, wherein the instructions, when executed by the one or more processors, further cause the system to determine whether the right ventricular pressure waveform of the patient is valid based on the signal quality indicator of the right ventricular pressure waveform of the patient, wherein a valid right ventricular pressure waveform indicates that the catheter is correctly positioned in the patient, and an invalid right ventricular pressure waveform indicates that the catheter is not correctly positioned in the patient or is not properly connected to the system, or that a signal quality problem is occurring.

12. 10. The system of claim 1, wherein the features extracted from the right ventricular pressure waveform of the patient determined to be within the one or more specified value ranges are available for further analysis.

13. a second hemodynamic sensor that continuously generates a second hemodynamic sensor signal representative of the patient's pulmonary artery pressure waveform, the catheter connected to the second hemodynamic sensor; The instructions, when executed by the one or more processors, further provide the system with: receiving the second hemodynamic sensor signal representative of the pulmonary artery pressure waveform of the patient; extracting features from the pulmonary artery pressure waveform of the patient; comparing the features extracted from the patient's pulmonary artery pressure waveform with one or more specified value ranges for the features to determine whether the values ​​of the features are within the one or more specified value ranges for the features; causing the display or the module to output an indication of whether the catheter is properly positioned within the patient based on whether the value of the feature extracted from the patient's pulmonary artery pressure waveform is within the one or more specified value ranges. The system of claim 1.

14. 14. The system of claim 13, wherein the features extracted from the right ventricular pressure waveform and the pulmonary artery pressure waveform of the patient comprise features determined by comparing the right ventricular pressure waveform and the pulmonary artery pressure waveform of the patient.

15. 15. The system of claim 14, wherein the feature determined by comparing the right ventricular pressure waveform and the pulmonary artery pressure waveform of the patient depends on whether the right ventricular pressure waveform and the pulmonary artery pressure waveform are synchronous or asynchronous.

16. 16. The system of claim 15, wherein the features determined by comparing the right ventricular pressure waveform and the pulmonary artery pressure waveform of the patient when the right ventricular pressure waveform and the pulmonary artery pressure waveform are synchronized include one or more of pulse transit time, systolic slope, and mean blood pressure.

17. 16. The system of claim 15, wherein the feature determined by comparing the right ventricular pressure waveform and the pulmonary artery pressure waveform of the patient when the right ventricular pressure waveform and the pulmonary artery pressure waveform are asynchronous comprises a change in pulse wave transit time.

18. The system of claim 1 , wherein the first hemodynamic sensor is connected to a hemodynamic monitor including the display, the one or more processors, and the computer-readable memory.

19. 1. A method for determining catheter placement in a patient, comprising: receiving, by a hemodynamic monitoring system, sensed hemodynamic data representative of a right ventricular pressure waveform of the patient; performing, by the hemodynamic monitoring system, waveform analysis of the hemodynamic data to extract features from the right ventricular pressure waveform of the patient; comparing, by the hemodynamic monitoring system, the features extracted from the patient's right ventricular pressure waveform with one or more specified value ranges for the features to determine whether the values ​​of the features extracted from the patient's right ventricular pressure waveform are within the one or more specified value ranges, or whether the values ​​of the features extracted from the patient's right ventricular pressure waveform are not within the one or more specified value ranges; A method comprising:

20. 1. A system for determining data quality from a catheter in a patient, comprising: a first hemodynamic sensor that continuously generates a first hemodynamic sensor signal representative of the patient's right ventricular pressure waveform, the catheter being connected to the first hemodynamic sensor; The display and one or more processors; When executed by the one or more processors, the system: receiving the first hemodynamic sensor signal representative of the right ventricular pressure waveform of the patient; extracting features from the right ventricular pressure waveform of the patient; comparing the features extracted from the patient's right ventricular pressure waveform with the one or more specified value ranges of the features to determine that the data from the catheter is of high quality if the value of the features extracted from the patient's right ventricular pressure waveform is within the one or more specified value ranges, or to determine that the data from the catheter is not of high quality if the value of the features extracted from the patient's right ventricular pressure waveform is not within the one or more specified value ranges; causing the display or module to output an indication of whether the data from the catheter is of high quality based on whether the value of the feature extracted from the right ventricular pressure waveform of the patient is within the one or more specified value ranges. a computer-readable memory encoded with instructions; A system comprising:

Images