Predicting postoperative pain using HOSVD

Abstract

Various embodiments of the present disclosure provide systems and methods for predicting a risk of mild or severe persistent postoperative pain (POP) for a target individual. The risk prediction may be determined based at least in part on a cohort prediction model. The cohort prediction model is initialized with historical multivariate intraoperative vital sign data associated with a surgical type cohort and associated with a binary classification of mild or severe persistent postoperative pain. Phase information for the historical multivariate intraoperative vital sign data is determined using complex high-order singular value decomposition. A relationship between the phase information and mild or severe persistent POP is then determined using discriminant analysis. The phase information for the multivariate intraoperative vital sign data for the target individual is then provided to the cohort prediction model, which classifies the target individual using the determined relationship. The risk prediction then includes a classification.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit under 35 U.S.C. §119(e) of U.S. patent application Ser. No. 63 / 202,374, filed Jun. 8, 2021, which is incorporated by reference in its entirety.

[0002] Government support This invention was made with Government support under Grant No. R01 GM114290 awarded by the National Institutes of Health. The Government has certain rights in this invention.

[0003] SUMMARY OF THE DISCLOSURE Embodiments of the present disclosure generally relate to systems and methods for post-operative pain (POP) risk prediction based on biological and biomedical measurements. [Background technology]

[0004] Long-term post-operative pain conditions and individual responses to pain relief medications are not yet fully understood. Over 100 million patients undergo surgery each year in the United States. Over 60 percent of these patients suffer from acute post-operative pain. Post-operative pain resolution is highly variable, with one-third of patients experiencing stable or even increased daily post-operative pain for at least seven days after surgery.

[0005] Persistent pain after acute postoperative pain (POP) is experienced by 10-50% of individuals after common surgical procedures such as cardiac, thoracic, spinal, or orthopedic surgery. Although even mild levels of persistent postoperative pain (POP) are associated with reduced physical and social activity, 2-10% of patients experiencing this type of pain may develop severe levels of pain, thus delaying recovery and return to normal daily function. Furthermore, persistent POP leads to increased direct healthcare costs through additional resource use. Prediction, identification, and evaluation of persistent POP are significant and under-recognized clinical problems. As a result, recognition of patients at risk for developing this type of pain has remained inadequate.

[0006] It is assumed that POP originates from various interacting factors, including but not limited to biological, psychological, and social factors. For example, psychological factors (depression, psychological vulnerability, stress, and catastrophizing) may be risk factors for the development of persistent POP. As another example, female gender may be a risk factor for the development of persistent POP. More importantly, acute POP, and especially the severity of exercise-induced pain, are major risk factors significantly associated with persistent POP. In such cases, the neuroplastic changes in the central nervous system caused by high-intensity acute POP may be responsible for the development of persistent POP. Summary of the Invention [Means for solving the problem]

[0007] In general, embodiments of the present disclosure provide methods, apparatus, systems, computing devices, computing entities, and / or the like for predicting risk of persistent postoperative pain (POP) for an individual and performing one or more risk prediction-based actions. In various embodiments, multivariate intraoperative vital sign data for a cohort of individuals may be collected and processed. Each individual may be associated with a binary classification indicating whether the individual experienced mild or severe persistent POP. Processing the multivariate intraoperative vital sign data may include performing a complex higher-order singular value decomposition (HOSVD) technique to generate a multi-dimensional representation. For example, the multivariate intraoperative vital sign data may be structured as a three-dimensional tensor (e.g., one dimension representing different vital sign variables, another dimension representing intraoperative time, and yet another dimension representing different individuals), and processing the multivariate intraoperative vital sign data may result in a plurality of first dimension mode data objects, a plurality of second dimension mode data objects, and a plurality of third dimension mode data objects.

[0008] In various embodiments, a cohort prediction model for a cohort may be generated based at least in part on processing the multivariate intraoperative vital sign data for the cohort of individuals. The cohort prediction model may be initialized with the multivariate intraoperative vital sign data for the cohort of individuals to determine a relationship between phase information of the multivariate intraoperative vital sign data. In various embodiments, a risk prediction for persistent POP for an individual of interest may be generated and provided based at least in part on providing the multivariate intraoperative vital sign data for the individual of interest to a cohort prediction model associated with a cohort to which the individual of interest belongs. The multivariate intraoperative vital sign data for the individual of interest may be processed (e.g., via a Hilbert transform technique) to determine phase information, and the cohort prediction model may use the multivariate intraoperative vital sign data and / or the phase information of the multivariate intraoperative vital sign data to determine a binary classification for the individual of mild or severe persistent POP. Thus, in various embodiments, the risk prediction for the individual of interest includes a binary classification of mild or severe persistent POP. In various embodiments, various risk-prediction based actions may then be performed on the individual.

[0009] In some embodiments, a computer-implemented method for predicting a risk of persistent post-operative pain for an individual includes, in part, receiving, by a processor, a prediction input data object including multivariate intra-operative vital sign data for the individual; processing the multivariate intra-operative vital sign data for the individual; providing at least the processed multivariate intra-operative vital sign data to a cohort prediction model associated with a cohort of individuals, the cohort prediction model being initialized with historical data objects associated with the post-operative time points; generating a risk prediction data object including a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time points; and performing one or more risk prediction-based actions for the individual.

[0010] In some embodiments, processing the multivariate intraoperative vital signs data comprises complexing the individual's multivariate intraoperative vital signs data. In some embodiments, complexing the individual's multivariate intraoperative vital signs data comprises augmenting the multivariate intraoperative vital signs data with their Hilbert transform.

[0011] In some embodiments, providing at least the processed multivariate intraoperative vital sign data to the cohort prediction model includes projecting the processed multivariate intraoperative vital sign data onto a three-dimensional manifold of the cohort prediction model and determining phase information of the projection of the processed multivariate intraoperative vital sign data.

[0012] In some embodiments, the cohort prediction model is generated and initialized at least in part based on: receiving historical data objects for each of a cohort including a plurality of individuals, each historical data object associated with a binary classification and including multivariate intra-operative vital sign data for a corresponding individual; processing the plurality of historical data objects to generate a plurality of first dimension mode data objects, a plurality of second dimension mode data objects, and a plurality of third dimension mode data objects; generating a cohort prediction model based at least in part on the plurality of first dimension mode data objects and the plurality of second dimension mode data objects, wherein the plurality of first dimension mode data objects and the plurality of second dimension mode data objects are processed to generate a three-dimensional manifold; and initializing the cohort prediction model with the plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and each binary classification.

[0013] In some embodiments, multiple historical data objects are aggregated and processed together using complex higher-order singular value decomposition (HOSVD), and a three-dimensional manifold is generated based at least in part on the ranks of the components generated by the HOSVD. In some embodiments, the ranks of the components may be determined using a rank feature method based at least in part on the Fisher ranking technique. In some embodiments, the top three ranked components are selected to form the three-dimensional manifold.

[0014] In some embodiments, each of the plurality of first dimensional mode data objects includes a weight for each of the one or more vital sign variable types, each of the plurality of second dimensional mode data objects includes a weight for each of a plurality of intraoperative time points, and each of the plurality of third dimensional mode data objects includes a weight for each of a plurality of individuals.

[0015] In some other embodiments, the plurality of first dimension mode data objects include eigenvectors of a first correlation entropy matrix, the first correlation entropy matrix being generated based at least in part on the plurality of historical data objects, the plurality of second dimension mode data objects include eigenvectors of a second correlation entropy matrix, the second correlation entropy matrix being generated based at least in part on the plurality of historical data objects, and the plurality of third dimension mode data objects include eigenvectors of a third correlation entropy matrix, the third correlation entropy matrix being generated based at least in part on the plurality of historical data objects. In some embodiments, the first correlation entropy matrix is ​​generated by applying a first cross-correlation entropy function to a first moment matrix, the first moment matrix being generated based at least in part on a first mode matrix expansion of a third order tensor, the second correlation entropy matrix is ​​generated by applying a second cross-correlation entropy function to a second moment matrix, the second moment matrix being generated based at least in part on a second mode matrix expansion of a third order tensor, and the third correlation entropy matrix is ​​generated by applying a third cross-correlation entropy function to a third moment matrix, the third moment matrix being generated based at least in part on a third mode matrix expansion of a third order tensor, the third order tensor representing a plurality of historical data objects. In some embodiments, each of the first, second, and third cross-correlation entropy functions is based on a Gaussian function.

[0016] In some embodiments, initializing the cohort predictive model includes determining a relationship between topological information of the projections of the multiple historical data objects onto a three-dimensional manifold and a binary classification.

[0017] In some embodiments, the one or more risk prediction based actions for the individual include displaying the risk prediction data object in a three-dimensional manifold, the three-dimensional manifold being generated based at least in part on the historical data object.

[0018] In some embodiments, an apparatus for predicting a risk of persistent post-operative pain for an individual comprises at least one processor and at least one non-transitory memory including program code, the at least one non-transitory memory and the program code configured, with the at least one processor, to cause the apparatus to at least receive a prediction input data object including multivariate intra-operative vital sign data for the individual, process the multivariate intra-operative vital sign data for the individual, provide at least the processed multivariate intra-operative vital sign data to a cohort prediction model associated with a cohort of individuals, the cohort prediction model initialized with historical data objects associated with post-operative time points, generate a risk prediction data object including a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object associated with the post-operative time points, and perform one or more risk prediction-based actions for the individual.

[0019] In some embodiments, configuring the at least one non-transitory memory and program code with the at least one processor to cause the device to process the multivariate intra-operative vital sign data comprises configuring the at least one non-transitory memory and program code with the at least one processor to cause the device to complex-enhance the individual's multivariate intra-operative vital sign data. In some embodiments, configuring the at least one non-transitory memory and program code with the at least one processor to cause the device to complex-enhance the individual's multivariate intra-operative vital sign data comprises configuring the at least one non-transitory memory and program code with the at least one processor to augment the multivariate intra-operative vital sign data with a Hilbert transform thereof.

[0020] In some embodiments, configuring the at least one non-transitory memory and program code with the at least one processor to cause the device to provide at least the processed multivariate intraoperative vital sign data to a cohort predictive model includes configuring the at least one non-transitory memory and program code with the at least one processor to cause the device to project the processed multivariate intraoperative vital sign data onto a three-dimensional manifold of the cohort predictive model and determine phase information of the projection of the processed multivariate intraoperative vital sign data.

[0021] In some embodiments, a cohort prediction model in which the apparatus is configured to provide at least processed multivariate intra-operative vital sign data is generated and initialized at least in part based on: receiving historical data objects for each of a cohort including a plurality of individuals, each historical data object associated with a binary classification and including multivariate intra-operative vital sign data for a corresponding individual; processing the plurality of historical data objects to generate a plurality of first dimension mode data objects, a plurality of second dimension mode data objects, and a plurality of third dimension mode data objects; generating a cohort prediction model based at least in part on the plurality of first dimension mode data objects and the plurality of second dimension mode data objects, wherein the plurality of first dimension mode data objects and the plurality of second dimension mode data objects are processed to generate a three-dimensional manifold; and initializing the cohort prediction model with the plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and each binary classification.

[0022] In some embodiments, each of the plurality of first dimensional mode data objects includes a weight for each of the one or more vital sign variable types, each of the plurality of second dimensional mode data objects includes a weight for each of a plurality of intraoperative time points, and each of the plurality of third dimensional mode data objects includes a weight for each of a plurality of individuals.

[0023] In some other embodiments, the plurality of first dimensional mode data objects include eigenvectors of a first correlation entropy matrix, the first correlation entropy matrix being generated based at least in part on the plurality of historical data objects, the plurality of second dimensional mode data objects include eigenvectors of a second correlation entropy matrix, the second correlation entropy matrix being generated based at least in part on the plurality of historical data objects, and the plurality of third dimensional mode data objects include eigenvectors of a third correlation entropy matrix, the third correlation entropy matrix being generated based at least in part on the plurality of historical data objects. In some embodiments, the first correlation entropy matrix is ​​generated by applying a first cross-correlation entropy function to a first moment matrix, the first moment matrix being generated based at least in part on an expansion of a first mode matrix of a third order tensor, the second correlation entropy matrix is ​​generated by applying a second cross-correlation entropy function to a second moment matrix, the second moment matrix being generated based at least in part on an expansion of a second mode matrix of a third order tensor, and the third correlation entropy matrix is ​​generated by applying a third cross-correlation entropy function to a third moment matrix, the third moment matrix being generated based at least in part on an expansion of a third mode matrix of a third order tensor, the third order tensor representing a plurality of historical data objects. In some embodiments, each of the first, second, and third cross-correlation entropy functions is based on a Gaussian function.

[0024] In some embodiments, to generate and initialize the cohort prediction model, the multiple historical data objects are aggregated and processed together using complex higher-order singular value decomposition (HOSVD), and a three-dimensional manifold is generated based at least in part on the ranks of the components generated by the HOSVD. In some embodiments, the ranks of the components may be determined using a rank feature method based at least in part on the Fisher ranking technique. In some embodiments, the top three ranked components are selected to form the three-dimensional manifold. In some embodiments, initializing the cohort prediction model includes determining a relationship between topological information of the projections of the multiple historical data objects onto the three-dimensional manifold and a binary classification.

[0025] In some embodiments, configuring the at least one non-transitory memory and program code, with the at least one processor, to cause the device to perform one or more risk prediction based actions for the individual includes configuring the at least one non-transitory memory and program code, with the at least one processor, to cause the device to display the risk prediction data objects in a three-dimensional manifold, the three-dimensional manifold being generated based at least in part on the history data objects. The present invention provides, for example, the following: (Item 1) 1. A computer-implemented method for predicting a risk of persistent post-operative pain for an individual, the computer-implemented method comprising: receiving, by a processor, a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with post-operative time points; generating a risk prediction data object including a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; and executing one or more risk-prediction based actions for the individual. (Item 2) 2. The computer-implemented method of claim 1, wherein processing the multivariate intraoperative vital sign data includes complexifying the multivariate intraoperative vital sign data of the individual, and providing at least the processed multivariate intraoperative vital sign data to a cohort predictive model includes projecting the processed multivariate intraoperative vital sign data onto a three-dimensional manifold of the cohort predictive model, and determining phase information of the projection of the processed multivariate intraoperative vital sign data. (Item 3) The cohort prediction model comprises: receiving a historical data object for each of a cohort including a plurality of individuals, each historical data object being associated with a binary classification and including multivariate intraoperative vital signs data for a corresponding individual; processing the plurality of history data objects to generate a plurality of first dimensional mode data objects, a plurality of second dimensional mode data objects, and a plurality of third dimensional mode data objects; generating a cohort prediction model based at least in part on the plurality of first dimension mode data objects and the plurality of second dimension mode data objects, wherein the plurality of first dimension mode data objects and the plurality of second dimension mode data objects are processed to generate a three-dimensional manifold; 2. The computer-implemented method of claim 1, wherein the cohort predictive model is generated and initialized based at least in part on the plurality of third dimension mode data objects and each binary classification, and initializing the cohort predictive model with the plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and each binary classification. (Item 4) 4. The computer-implemented method of claim 3, wherein the multiple historical data objects are aggregated and processed together using a complex higher-order singular value decomposition (HOSVD), and the three-dimensional manifold is generated based at least in part on ranks of the components generated by the HOSVD. (Item 5) each of the plurality of first dimension-mode data objects includes a weight for each of one or more vital sign variable types; each of the plurality of second dimension mode data objects includes a weight for each of a plurality of intraoperative time points; 4. The computer-implemented method of claim 3, wherein each of the plurality of third dimension mode data objects includes a weight for each of the plurality of individuals. (Item 6) 4. The computer-implemented method of claim 3, wherein initializing the cohort predictive model includes determining a relationship between topological information of the projections of the plurality of historical data objects onto the three-dimensional manifold and a binary classification. (Item 7) the plurality of first dimension-mode data objects include eigenvectors of a first correlation entropy matrix, the first correlation entropy matrix being generated based at least in part on the plurality of historical data objects; the plurality of second dimension mode data objects include eigenvectors of a second correlation entropy matrix, the second correlation entropy matrix being generated based at least in part on the plurality of historical data objects; 4. The computer-implemented method of claim 3, wherein the plurality of third dimension mode data objects comprises eigenvectors of a third correlation entropy matrix, the third correlation entropy matrix being generated based at least in part on the plurality of historical data objects. (Item 8) the first correlation entropy matrix is ​​generated by applying a first cross-correlation entropy function to a first moment matrix, the first moment matrix being generated based at least in part on a first mode matrix expansion of a third order tensor; the second correlation entropy matrix is ​​generated by applying a second cross-correlation entropy function to a second moment matrix, the second moment matrix being generated based at least in part on a second mode matrix expansion of the third-order tensor; 8. The computer-implemented method of claim 7, wherein the third correlation entropy matrix is ​​generated by applying a third cross-correlation entropy function to a third moment matrix, the third moment matrix being generated based at least in part on a third mode matrix expansion of the third order tensor, the third order tensor representing the plurality of historical data objects. (Item 9) Item 9. The computer-implemented method of item 8, wherein each of the first, second, and third cross-correlation entropy functions is based on a Gaussian function. (Item 10) 2. The computer-implemented method of claim 1, wherein the one or more risk prediction based actions for the individual include displaying the risk prediction data object in a three-dimensional manifold, the three-dimensional manifold being generated based at least in part on the historical data object. (Item 11) 1. An apparatus for predicting a risk of persistent post-operative pain for an individual, the apparatus comprising at least one processor and at least one non-transitory memory including program code, the at least one non-transitory memory and the program code being configured by the at least one processor to provide the apparatus with at least: receiving a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with post-operative time points; generating a risk prediction data object including a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; and executing one or more risk-prediction-based actions for the individual. (Item 12) 12. The apparatus of claim 11, wherein processing the multivariate intraoperative vital sign data includes complexifying the multivariate intraoperative vital sign data of the individual, and providing at least the processed multivariate intraoperative vital sign data to a cohort predictive model includes projecting the processed multivariate intraoperative vital sign data onto a three-dimensional manifold of the cohort predictive model, and determining phase information of the projection of the processed multivariate intraoperative vital sign data. (Item 13) The cohort prediction model comprises: receiving a historical data object for each of a cohort including a plurality of individuals, each historical data object being associated with a binary classification and including multivariate intraoperative vital signs data for a corresponding individual; processing the plurality of history data objects to generate a plurality of first dimensional mode data objects, a plurality of second dimensional mode data objects, and a plurality of third dimensional mode data objects; generating a cohort prediction model based at least in part on the plurality of first dimension mode data objects and the plurality of second dimension mode data objects, wherein the plurality of first dimension mode data objects and the plurality of second dimension mode data objects are processed to generate a three-dimensional manifold; 12. The apparatus of claim 11, wherein the cohort predictive model is generated and initialized based at least in part on the plurality of third dimension mode data objects and each binary classification, and initializing the cohort predictive model with the plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and each binary classification. (Item 14) Item 14. The apparatus of item 13, wherein the multiple historical data objects are aggregated and processed together using complex higher-order singular value decomposition (HOSVD), and the three-dimensional manifold is generated based at least in part on ranks of components generated by the HOSVD. (Item 15) each of the plurality of first dimension-mode data objects includes a weight for each of one or more vital sign variable types; each of the plurality of second dimension mode data objects includes a weight for each of a plurality of intraoperative time points; Item 14. The apparatus of item 13, wherein each of the plurality of third dimension mode data objects includes a weight for each of the plurality of individuals. (Item 16) 14. The apparatus of claim 13, wherein initializing the cohort prediction model includes determining a relationship between topological information of the projections of the plurality of historical data objects onto the three-dimensional manifold and a binary classification. (Item 17) the plurality of first dimension-mode data objects include eigenvectors of a first correlation entropy matrix, the first correlation entropy matrix being generated based at least in part on the plurality of historical data objects; the plurality of second dimension mode data objects include eigenvectors of a second correlation entropy matrix, the second correlation entropy matrix being generated based at least in part on the plurality of historical data objects; Item 14. The apparatus of item 13, wherein the plurality of third dimension mode data objects include eigenvectors of a third correlation entropy matrix, the third correlation entropy matrix being generated based at least in part on the plurality of historical data objects. (Item 18) the first correlation entropy matrix is ​​generated by applying a first cross-correlation entropy function to a first moment matrix, the first moment matrix being generated based at least in part on a first mode matrix expansion of a third order tensor; the second correlation entropy matrix is ​​generated by applying a second cross-correlation entropy function to a second moment matrix, the second moment matrix being generated based at least in part on a second mode matrix expansion of the third-order tensor; 20. The apparatus of claim 17, wherein the third correlation entropy matrix is ​​generated by applying a third cross-correlation entropy function to a third moment matrix, the third moment matrix being generated based at least in part on a third mode matrix expansion of the third order tensor, the third order tensor representing the plurality of historical data objects. (Item 19) Item 19. The apparatus of item 18, wherein each of the first, second, and third cross-correlation entropy functions is based on a Gaussian function. (Item 20) 12. The apparatus of claim 11, wherein the one or more risk prediction based actions for the individual include displaying the risk prediction data object in a three-dimensional manifold, the three-dimensional manifold being generated based at least in part on the historical data object. [Brief description of the drawings]

[0026] Embodiments of the present disclosure have therefore been described in general terms and with reference to the accompanying drawings, which are not necessarily drawn to scale.

[0027] [Figure 1] 1 provides an example overview of an example system architecture that may be used to practice various embodiments of the present disclosure. [Diagram 2] 1 is a schematic diagram of an exemplary system computing entity in accordance with various embodiments of the present disclosure. [Diagram 3] 1 is a schematic diagram of an exemplary client computing entity in accordance with various embodiments of the present disclosure. [Figure 4] 1 provides a block diagram of an exemplary system computing entity in accordance with various embodiments of the present disclosure. [Figure 5A] 1 provides an exemplary operational process flow for predicting risk of post-operative pain, according to various embodiments of the present disclosure. [Figure 5B] 1 provides an exemplary operational process flow for predicting risk of post-operative pain, according to various embodiments of the present disclosure. [Figure 6] 1 illustrates some of several exemplary cohort predictive models, according to some embodiments of the present disclosure. [Figure 7] 1 provides an illustration of an exemplary process for predicting risk of post-operative pain, according to various embodiments of the present disclosure. [Figure 8] 1 illustrates some exemplary cohort predictive models according to some other embodiments of the present disclosure. [Figure 9A] 13 illustrates the first three temporal factors obtained using two different example sets of kernel widths, according to some embodiments of the present disclosure. [Figure 9B] 13 illustrates the first three temporal factors obtained using two different example sets of kernel widths, according to some embodiments of the present disclosure. [Figure 10] 13 shows the first three temporal factors obtained using an exemplary optimal kernel width, according to some embodiments of the present disclosure. [Figure 11] 13 illustrates an exemplary change in Fisher score values ​​for the top 10 components extracted by applying an exemplary complex HOSVD for an exemplary set of different kernel widths, in accordance with some embodiments of the present disclosure. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0028] Various embodiments of the present disclosure will now be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all, embodiments of the present disclosure are shown. Indeed, the present disclosure may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, but rather, these embodiments are provided so that the present disclosure will satisfy applicable legal requirements. The term "or" (also indicated as " / ") is used herein in both an alternative and conjunctive sense, unless otherwise indicated. The terms "exemplary" and "exemplary" are used to be examples without an indication of quality level. Like numbers refer to like elements throughout.

[0029] I. General Overview and Technical Advantages Various embodiments of the present disclosure generate a cohort prediction model for determining a relationship between phase information of multivariate intraoperative vital sign data and mild or severe persistent POP. Various embodiments further apply such determined relationship to predict whether a subject individual is likely to develop mild or severe persistent POP based at least in part on the phase information of multivariate intraoperative vital sign data for the subject individual. In doing so, various embodiments advantageously take into account each individual's unique systematic response to surgical insults in relation to the development of persistent postoperative POP.

[0030] During surgery, different vital sign variable types such as heart rate, blood pressure, and respiration can be used as indicators of an individual's systematic response, since the autonomic nervous system responds continuously to various surgical stimuli. During general anesthesia, when a sufficient dose of anesthetic is administered to prevent a response to skin incision, the hemodynamic response induced by surgical stress is not necessarily attenuated. The sympathetic nervous system essentially changes hemodynamic parameters such as regional blood flow, blood pressure, and heart rate in response to noxious stimuli. Anesthetic drugs interfere with this system at different levels. Among the hemodynamic parameters, heart rate may also include changes in parasympathetic discharge. Therefore, monitoring and analyzing the time series of a patient's hemodynamic response associated with various surgical stimuli and pain imbalance under general anesthesia indirectly characterizes the behavior of the autonomic nervous system to pain stimuli and provides a relationship to the development of persistent POP.

[0031] Various embodiments of the present disclosure use complex HOSVD to investigate dynamic correlations with lead / lag relationships in intraoperative vital signs. In various embodiments, complex vital sign data is generated using a Hilbert transform technique. The multivariate temporal structure of intraoperative vital signs is revealed by quantifying the cross-correlation of the data as a joint function of vital sign variable type and time. Thus, various embodiments advantageously use complex HOSVD to compress the correlation structure into a fairly small number of complex eigenvectors. The complex eigenvectors are used as a new basis to describe the hemodynamic response. After projecting into a subspace with the new basis, the complex correlation between each intraoperative time series and the eigenvectors becomes evident in the magnitude and phase of the correlation. In various embodiments, the phase of the correlation is used to infer the lead / lag relationship in the original intraoperative time series.

[0032] In various embodiments, the multivariate intraoperative vital signs data includes intraoperative time series recorded for various vital signs variable types, such as heart rate, blood oxygen level, end-tidal CO2 level, respiratory tidal volume, systolic blood pressure, diastolic blood pressure, isoflurane concentration, sevoflurane concentration, and / or the like. Various embodiments may use the multivariate intraoperative vital signs data for a cohort of individuals to generate a cohort predictive model. In various embodiments, the multivariate intraoperative vital signs data for a cohort is represented as a three-dimensional tensor

number

[0033] Indeed, various embodiments of the present disclosure provide technical advantages and improvements over various other methods and systems for analyzing multivariate intraoperative vital signs data. For example, cross-spectral analysis is difficult to employ and is not very descriptive of the randomly occurring events and unknown dominant frequencies of the dynamic interactions between coupled biological systems in hemodynamic regulation. Furthermore, various embodiments advantageously determine phase information related to the propagation dynamics of the hemodynamic response as opposed to the orthostatic dynamics. In general, various embodiments of the present disclosure are uniquely and advantageously suited to accurately predict the risk of sustained POP for a target individual based at least in part on an analysis of the individual's intrinsic response to painful stimuli captured in multivariate intraoperative vital signs data.

[0034] II. Exemplary System Architecture 1 is a schematic diagram of an example system architecture 100 for predicting a risk of persistent POP for an individual and performing one or more risk prediction-based actions. The system architecture 100 includes a persistent POP prediction system 101 configured to generate a cohort prediction model, generate and provide a risk prediction data object for the subject individual based at least in part on the cohort prediction model, perform one or more risk prediction-based actions, and / or the like. In various embodiments, the persistent POP prediction system 101 provides a risk prediction data object for the subject individual based at least in part on receiving a prediction input data object from a client computing entity 106.

[0035] In some embodiments, the persistent POP prediction system 101 may communicate with at least one of the client computing entities 106 using one or more communication networks. Examples of communication networks include any wired or wireless communication network, including, for example, a wired or wireless local area network (LAN), a personal area network (PAN), a metropolitan area network (MAN), a wide area network (WAN), etc., as well as any hardware, software, and / or firmware required to implement the same (e.g., network routers and / or the like). In various embodiments, the persistent POP prediction system 101 includes an application programming interface (API) to receive prediction input data objects from the client computing entities 106 via API calls and to provide risk prediction data objects via API responses.

[0036] The persistent POP prediction system 101 may include a system computing entity 102 and a storage subsystem 104. The system computing entity 102 may be configured to generate a cohort prediction model, receive prediction input data objects from one or more client computing entities 106, process the prediction input data objects, and provide a risk prediction data object based at least in part on providing the prediction input data objects to the cohort prediction model. In various embodiments, the system computing entity 102 is a cloud-based computing system, comprising one or more computing devices, each configured to share and allocate computer processing resources and data.

[0037] The storage subsystem 104 may be configured to store data for predicting a risk of persistent POPs for an individual and for performing one or more risk prediction-based actions. For example, a cohort prediction model generated by the system computing entity 102 may be stored in the storage subsystem 104. The storage subsystem 104 may include one or more storage units, such as multiple distributed storage units connected through a computer network. Each storage unit in the storage subsystem 104 may store at least one of the one or more data assets and / or one or more data related to the calculated characteristics of the one or more data assets. Furthermore, each storage unit in the storage subsystem 104 may include one or more non-volatile storage media or memory media, including, but not limited to, a hard disk, a ROM, a PROM, an EPROM, an EEPROM, a flash memory, an MMC, an SD memory card, a memory stick, a CBRAM, a PRAM, a FeRAM, a NVRAM, an MRAM, a RRAM, a SONOS, an FJG RAM, a Millipede memory, a racetrack memory, and / or the like.

[0038] III. Exemplary Computing Entities In general, the terms device, system, computing entity, entity, and / or similar words used interchangeably herein may refer to, for example, one or more computers, computing entities, desktops, mobile phones, tablets, phablets, notebooks, laptops, distributed systems, kiosks, input terminals, servers or server networks, blades, gateways, switches, processing devices, processing entities, set-top boxes, relays, routers, network access points, base stations, etc., and / or any combination of devices or entities adapted to perform the functions, operations, and / or processes described herein. Such functions, operations, and / or processes may include, for example, transmitting, receiving, operating, processing, displaying, storing, determining, creating / generating, monitoring, evaluating, comparing, and / or similar terms used interchangeably herein. In an embodiment, these functions, operations, and / or processes may be performed on data, content, information, and / or similar terms used interchangeably herein.

[0039] 2 provides an illustrative schematic diagram representative of a system computing entity 102 that may be used in conjunction with embodiments of the present disclosure. For example, the system computing entity 102 may be configured to and / or comprise means for generating cohort prediction models, generating and providing persistent POP risk prediction data objects, and performing one or more risk prediction-based actions. As shown in FIG. 2, in one embodiment, the system computing entity 102 may include or communicate with one or more processing elements 205 (also referred to as processors, processing circuits, and / or similar terms used interchangeably herein) that communicate with other elements within the system computing entity 102, for example, via a bus. As will be appreciated, the processing element 205 may be embodied in many different ways.

[0040] For example, processing element 205 may be embodied as one or more complex programmable logic devices (CPLDs), microprocessors, multi-core processors, co-processing entities, application specific instruction set processors (ASIPs), microcontrollers, and / or controllers. Additionally, processing element 205 may be embodied as one or more other processing devices or circuits. The term circuitry may refer to an entirely hardware embodiment or a combination of hardware and computer program products. Thus, processing element 205 may be embodied as an integrated circuit, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a programmable logic array (PLA), a hardware accelerator, other circuits, and / or the like.

[0041] It will thus be understood that processing element 205 may be configured for a particular use or configured to execute instructions stored on a volatile or non-volatile medium or otherwise accessible to processing element 205. Thus, whether configured by hardware or a computer program product, or a combination thereof, processing element 205, when configured accordingly, may be capable of performing steps or operations according to embodiments of the present disclosure.

[0042] In one embodiment, the system computing entity 102 may further include or communicate with non-volatile media (also referred to as non-volatile storage, memory, memory storage, memory circuitry and / or similar terms used interchangeably herein). In one embodiment, the non-volatile storage or memory may include one or more non-volatile storage or memory media 210, including but not limited to a hard disk, ROM, PROM, EPROM, EEPROM, flash memory, MMC, SD memory cards, memory sticks, CBRAM, PRAM, FeRAM, NVRAM, MRAM, RRAM, SONOS, FJG RAM, Millipede memory, racetrack memory, and / or the like.

[0043] As will be appreciated, the non-volatile storage medium or memory medium 210 may store databases, database instances, database management systems, data, applications, programs, program modules, scripts, source code, object code, byte code, compiled code, interpreted code, machine code, executable instructions, and / or the like. The terms database, database instance, database management system, and / or similar terms used interchangeably herein may refer to a collection of records or data stored in a computer-readable storage medium using one or more database models, such as a hierarchical database model, a network model, a relational model, an entity-relationship model, an object model, a document model, a semantic model, a graph model, and / or the like.

[0044] In one embodiment, the system computing entity 102 may further include or be in communication with volatile media (also referred to as volatile storage, memory, memory storage, memory circuitry and / or similar terms used interchangeably herein). In one embodiment, the volatile storage or memory may also include one or more volatile storage media or memory media 215, including but not limited to RAM, DRAM, SRAM, FPM DRAM, EDO DRAM, SDRAM, DDR SDRAM, DDR2 SDRAM, DDR3 SDRAM, RDRAM, TTRAM, T-RAM, Z-RAM, RIMM, DIMM, SIMM, VRAM, cache memory, registered memory, and / or the like.

[0045] As will be appreciated, the volatile storage medium or memory medium 215 may be used to store at least portions of databases, database instances, database management systems, data, applications, programs, program modules, scripts, source code, object code, byte code, compiled code, interpreted code, machine code, executable instructions, and / or the like that are executed, for example, by the processing elements 205. Thus, the databases, database instances, database management systems, data, applications, programs, program modules, scripts, source code, object code, byte code, compiled code, interpreted code, machine code, executable instructions, and / or the like, with the assistance of the processing elements 205 and the operating system, may be used to control certain aspects of the operation of the system computing entity 102.

[0046] As shown, in one embodiment, the system computing entity 102 may also include one or more network interfaces 220 for communicating with various computing entities (e.g., one or more client computing entities 106), such as by communicating data, content, information, and / or like terms used interchangeably herein, which may be transmitted, received, manipulated, processed, displayed, stored, and / or the like. Such communications may be performed using a wired data transmission protocol, such as Fiber Distributed Data Interface (FDDI), Digital Subscriber Line (DSL), Ethernet, Asynchronous Transfer Mode (ATM), Frame Relay, Data over Cable Service Interface Specification (DOCSIS), or any other wired transmission protocol. Similarly, the system computing entity 102 may include one or more network interfaces 220 for communicating with various computing entities (e.g., one or more client computing entities 106), such as one or more network interfaces 220 for communicating with various computing entities (e.g., one or more client computing entities 106), such as one or more network interfaces 220 for communicating with various computing entities (e.g., one or more client computing entities 106), such as one or more client computing entities 106, ... The wireless external communications network may be configured to communicate over a wireless external communications network using any of a variety of protocols, such as 1X (1xRTT), Wideband Code Division Multiple Access (WCDMA), Global System for Mobile Communications (GSM), Enhanced Data Rates for GSM Evolution (EDGE), Time Division Synchronous Code Division Multiple Access (TD-SCDMA), Long Time Evolution (LTE), Evolution Universal Terrestrial Radio Access Network (E-UTRAN), Evolution Data Optimized (EVDO), High Speed ​​Packet Access (HSPA), High Speed ​​Downlink Packet Access (HSDPA), IEEE EE 802.111 (FI), WiFIDI, 802.16 (WIMAX), Ultra Wide Band (UWB), Infrared (IR) protocol, Near Field Communication (NFC) protocol, Wibree, Bluetooth protocol, Wireless Universal Serial Bus (USB) protocol, and / or any other wireless protocol.

[0047] Although not shown, the system computing entity 102 may include or communicate with one or more input elements, such as a keyboard input, a mouse input, a touch screen / display input, a motion input, a movement input, an audio input, a pointing device input, a joystick input, a keypad input, and / or the like. The system computing entity 102 may also include or communicate with one or more output elements (not shown), such as an audio output, a video output, a screen / display output, a motion output, a movement output, and / or the like.

[0048] As will be appreciated, one or more of the components of the system computing entity 102 may be located remotely from other components, such as in a distributed system. Additionally, one or more of the components may be aggregated and additional components that perform the functions described herein may be included in the system computing entity 102. Thus, the system computing entity 102 can be adapted to accommodate a variety of needs and situations.

[0049] 3 provides a schematic diagram of an example client computing entity 106 that may be used in conjunction with embodiments of the present disclosure. The client computing entity 106 may be operated by various parties, and the system architecture 100 may include one or more client computing entities 106. As shown in FIG. 3, the client computing entity 106 may include an antenna 312, a transmitter 304 (e.g., wireless), a receiver 306 (e.g., wireless), and correspondingly, a processing element 308 (e.g., a CPLD, a microprocessor, a multi-core processor, a co-processing entity, an ASIP, a microcontroller, and / or a controller) that provides signals to and receives signals from the transmitter 304 and the receiver 306.

[0050] The signals provided to and received by the transmitter 304 and receiver 306 may correspondingly include signaling information / data in accordance with the air interface standard of the applicable wireless system. In this regard, the client computing entity 106 may be capable of operating with one or more air interface standards, communication protocols, modulation types, and access types. More specifically, the client computing entity 106 may operate in accordance with any of a number of wireless communication standards and protocols, such as those described above with respect to the system computing entity 102. In a particular embodiment, the client computing entity 106 may operate in accordance with multiple wireless communication standards and protocols, such as UMTS, CDMA 2000, 1xRTT, WCDMA, GSM, EDGE, TD-SCDMA, LTE, E-UTRAN, EVDO, HSPA, HSDPA, Wi-Fi, Wi-Fi Direct, WiMAX, UWB, IR, NFC, Bluetooth, USB, and / or the like. Similarly, the client computing entity 106 may operate according to a number of wired communication standards and protocols, such as those described above with respect to the system computing entity 102 via the network interface 320 .

[0051] Through these communication standards and protocols, the client computing entity 106 can communicate with various other entities (e.g., system computing entity 102, storage subsystem 104) using concepts such as Unstructured Supplementary Service Data (USSD), Short Message Service (SMS), Multimedia Messaging Service (MMS), Dual Tone Multi-Frequency Signaling (DTMF), and / or Subscriber Identity Module Dialer (SIM Dialer). The client computing entity 106 can also download modifications, add-ons, and updates, for example, to its firmware, software (including, for example, executable instructions, applications, program modules), and operating system.

[0052] According to one embodiment, the client computing entity 106 may include location determination aspects, devices, modules, functionality, and / or similar words that are used interchangeably herein. For example, the client computing entity 106 may include an outdoor positioning aspect, such as a position module adapted to obtain, for example, latitude, longitude, elevation, geocode, course, direction, orientation, speed, Universal Coordinated Time (UTC), date, and / or various other information / data. In one embodiment, the position module may obtain data, sometimes known as ephemeris data, by identifying the number of satellites in view and their relative positions (e.g., using the Global Positioning System (GPS)). The satellites may be a variety of different satellites, including a Low Earth Orbit (LEO) satellite system, a Department of Defense (DOD) satellite system, the European Union Galileo Positioning System, the China Compass Navigation System, the Indian Regional Navigation Satellite System, and / or the like. This data can be collected using various coordinate systems, such as decimal degrees (DD), degrees minutes seconds (DMS), Universal Transverse Mercator (UTM), Universal Polar Stereographic (UPS) coordinate system, and / or the like. Alternatively, location information / data can be determined by triangulating the position of the client computing entity 106 in relation to various other systems, including cellular towers, Wi-Fi access points, and / or the like. Similarly, the client computing entity 106 can include indoor positioning aspects, such as, for example, a location module adapted to obtain latitude, longitude, altitude, geocode, course, direction, orientation, speed, time, date, and / or various other information / data. Some of the indoor systems can use various position or location technologies, including RFID tags, indoor beacons or transmitters, Wi-Fi access points, cellular towers, nearby computing devices (e.g., smartphones, laptops), and / or the like. For example, such technologies may include iBeacon®, gimbal proximity beacons, Bluetooth Low Energy (BLE) transmitters, NFC transmitters, and / or the like.These indoor positioning aspects can be used in a variety of settings to determine the location of someone or something within a few inches or centimeters.

[0053] The client computing entity 106 may also comprise a user interface (which may include a display 316 coupled to the processing element 308) and / or a user input interface (coupled to the processing element 308). For example, a user interface may be a user application, browser, user interface, and / or similar words used interchangeably herein that are executed interchangeably on and / or accessible via the client computing entity 106 to interact with and cause the display of information / data from the system computing entity 102, as described herein. The user input interface may comprise any of a number of devices or interfaces that enable the client computing entity 106 to receive data, such as a keypad 318 (hard or soft), a touch display, a voice / speech or motion interface, or other input device. In embodiments that include a keypad 318, the keypad 318 may include (or display) conventional numeric (0-9) and associated keys (#, *), as well as other keys used to operate the client computing entity 106, and may include a full set of alphabetical keys or a set of keys that may be activated to provide a full set of alphabetical keys. In addition to providing input, the user input interface can be used to activate or deactivate certain features, such as, for example, a screen saver and / or a sleep mode.

[0054] The client computing entity 106 may also include volatile storage or memory 322 and / or non-volatile storage or memory 324, which may be embedded and / or removable. For example, the non-volatile memory may be ROM, PROM, EPROM, EEPROM, flash memory, MMC, SD memory cards, memory sticks, CBRAM, PRAM, FeRAM, NVRAM, MRAM, RRAM, SONOS, FJG RAM, Millipede memory, racetrack memory, and / or the like. The volatile memory may be RAM, DRAM, SRAM, FPM DRAM, EDO DRAM, SDRAM, DDR SDRAM, DDR2 SDRAM, DDR3 SDRAM, RDRAM, TTRAM, T-RAM, Z-RAM, RIMM, DIMM, SIMM, VRAM, cache memory, registered memory, and / or the like. The volatile and non-volatile storage or memory may store databases, database instances, database management systems, data, applications, programs, program modules, scripts, source code, object code, byte code, compiled code, interpreted code, machine code, executable instructions, and / or the like to implement the functionality of the client computing entity 106. As shown, this may include user applications resident on the entity or accessible through a browser or other user interface to communicate with the system computing entity 102, various other computing entities, and / or the storage subsystem 104.

[0055] In another embodiment, the client computing entity 106 may include one or more components or functionality that are the same or similar to those of the system computing entity 102, as described in more detail above. It should be recognized that these architectures and descriptions are provided for illustrative purposes only and are not limited to the various embodiments.

[0056] In various embodiments, the client computing entity 106 may be embodied as an artificial intelligence (AI) computing entity of an Amazon Echo, an Amazon Echo Dot, an Amazon Show, a Google Home, and / or the like. As such, the client computing entity 106 may be configured to provide and / or receive information / data from a user via input / output mechanisms of a display, a camera, a speaker, a voice-activated input, and / or the like. In certain embodiments, the AI ​​computing entity may include one or more predefined executable program algorithms stored within an on-board memory storage module and / or accessible over a network. In various embodiments, the AI ​​computing entity may be configured to retrieve and / or execute one or more of the predefined program algorithms in response to the occurrence of a predefined trigger event.

[0057] IV. Exemplary System Operation Model Generation Module FIG. 4 provides a block diagram of an exemplary system computing entity 102. In various embodiments, the system computing entity 102 includes a model generation module 410. The model generation module 410 may be configured to generate a cohort predictive model based at least in part on historical data objects for a cohort of individuals (e.g., a surgical type cohort) that have undergone similar surgical procedures or procedures. In various embodiments, the model generation module 410 may also be configured to initialize or train the cohort predictive model. All surgical procedures consist of physical interventions on a particular body system. Thus, the type of procedure specifies the organs, organ systems, or tissues involved, as well as the degree of invasiveness. The impact of the type of surgery on the development of chronic or persistent POP is well understood by those skilled in the art. Longer and more complex surgeries are often associated with a higher risk of chronic pain development, but the pattern is irregular and also related to the type of tissue involved in the surgery. It may thus be appreciated that the predictive models may be specific to a surgical cohort, and thus the system computing entity 102 (e.g., the model generation module 410) may be configured to generate cohort-specific predictive models, or cohort predictive models. In particular, the model generation module 410 may be configured to generate cohort predictive models based at least in part on historical data objects, each of which includes multivariate intraoperative vital signs data and is associated with a binary classification of mild or severe persistent POP.

[0058] 5A thus provides a process 500 for generating and initializing a cohort predictive model. In various embodiments, the operations of process 500 may be performed by a system computing entity 102 and / or a model generation module 410, where the system computing entity 102 may comprise a processing element 205, memory 210, 215, a network interface 220, and / or the like for performing the operations of process 500.

[0059] As illustrated in FIG. 5A, process 500 includes operation 501. In various embodiments, process 500 begins with operation 501. Operation 501 includes receiving a historical data object for each of a cohort including a plurality of individuals. Each historical data object is associated with a binary classification and includes multivariate intraoperative vital sign data for the individuals of the cohort. In various embodiments, the binary classification is a classification of whether a corresponding individual of the cohort experienced mild or severe persistent POP. The binary classification may correspond to a particular postoperative period, time frame, time point, and / or the like. For example, a binary classification may be a classification of whether a corresponding individual experienced mild or severe persistent POP 30 days after surgery, while another binary classification may be for 90 days after surgery. In various embodiments, each historical data object may be associated with one or more binary classifications, each corresponding to a different postoperative period, time frame, time point, and / or the like.

[0060] In various embodiments, the multivariate intraoperative vital sign data includes data collected on a variety of different vital sign variable types (e.g., heart rate, respiratory tidal volume, blood pressure, blood oxygen, and / or the like) throughout the intraoperative period. In various embodiments, the multivariate intraoperative vital sign data includes hemodynamic data. The dynamic interplay between surgical perturbations to circulatory function and the sympathetic / parasympathetic responses under general anesthesia to compensate for them is reflected in the fluctuations of hemodynamic parameters during surgery. Heart rate can be used to characterize the autonomic nervous system, as it drives the function of the heart by increasing or decreasing the heart rate. Arterial blood pressure can be used as an imperfect estimate of the adequacy of tissue perfusion. Peripheral capillary oxygen saturation (SpO2), which measures the amount of oxygen in the blood, also contains relevant information regarding the state of circulation during surgery, and therefore the autonomic state (state of the autonomic nervous system). Respiration slows the cyclical fluctuations in baseline heart rate, which also affects blood pressure. Therefore, respiratory-related parameters such as respiratory tidal volume and end-tidal CO2 provide additional information regarding autonomic status and provide insight into the risk of persistent POP.

[0061] Respiration couples with heart rate fluctuations through centrally mediated mechanisms, simultaneously mechanically perturbing aortic pressure, venous return, and pulmonary vasculature. Respiration-induced cyclical variations in blood pressure affect heart rate through the autonomically mediated baroreceptor reflex. Variations in peripheral vascular resistance are another source of perturbation to homeostasis, as vascular beds regulate local blood flow to balance demand and supply. These variations perturb blood pressure and lead to compensatory variations in heart rate.

[0062] The frequency components of the fluctuations of hemodynamic parameters, which indirectly reflect the frequency bands of sympathetic and parasympathetic activity that compensate for short-term fluctuations in heart rate and other hemodynamic parameters, are concentrated in three fundamental spectral peaks: low-, mid-, and high-frequency peaks. The high-frequency peaks (0.3-0.5 Hz) represent the respiratory frequency and shift with variations in respiratory rate. The mid-frequency peaks (0.09-0.15 Hz) represent blood pressure oscillations that occur at frequencies lower than the respiratory frequency and are related to the frequency response of the baroreceptor reflex. The low-frequency peaks (0.02-0.09 Hz) are associated with fluctuations in vasomotor tone.

[0063] The discussed spectral characteristics of the fluctuations in hemodynamic parameters are mainly associated with the activities of the sympathetic and parasympathetic nervous systems and the renin-angiotensin system to control cardiovascular response, and specifically, the hemodynamic fluctuations occurring at high frequencies (above about 0.1 Hz) are associated with the activities of the parasympathetic nervous system. On the other hand, the hemodynamic fluctuations at lower frequencies may reflect the integrated activities of the sympathetic and / or parasympathetic nervous systems. The renin-angiotensin system is a hormonal system that regulates blood pressure, and fluid and electrolyte balance, as well as systemic vascular resistance. Blocking this system has been shown to dramatically increase the amplitude of low frequencies.

[0064] Thus, to provide a comprehensive view of the autonomic nervous system status during surgery, the multivariate intraoperative vital sign data includes hemodynamic parameter data collected at a high sampling rate. For example, in some embodiments, the multivariate intraoperative vital sign data is collected at a rate of one sample per second. The multivariate intraoperative vital sign data for an individual may include periodic measurements of heart rate, blood oxygen level, end-tidal CO2, respiratory tidal volume, systolic blood pressure, diastolic blood pressure, isoflurane concentration, sevoflurane concentration, and / or the like. In some other embodiments, the multivariate intraoperative vital sign data is collected at a rate of one sample per minute. This sampling rate limits the analysis to a narrow band at low frequencies. This narrow band at lower frequencies remains useful for developing cohort predictive models, since observable changes in hemodynamic parameters and corresponding surgical decisions occur at intervals of one minute or more.

[0065] Process 500 further includes operation 502. In various embodiments, operation 502 may follow operation 501. Operation 502 includes processing the plurality of data objects to generate a plurality of first dimension mode data objects, a plurality of second dimension mode data objects, and a plurality of third dimension mode data objects.

[0066] In various embodiments, the multiple historical data objects may be, include, aggregate, and / or be third-order tensors. For example, the multiple data objects may include recordings of I1 intraoperative vital sign variable types (e.g., heart rate, respiratory tidal volume) across I3 different patients in a surgical type cohort, with intraoperative vital signs recorded at I2 time points for each patient. Intraoperative vital sign recordings across a number of different time points may be cut to a common time window (e.g., I2 time points) to fit this constraint. Thus, the multivariate intraoperative vital sign data for multiple patients may be represented as an I1 x I2 x I3 array of vital signs,

number

number

[0067] In various embodiments, processing the multiple data objects includes processing the multivariate intraoperative vital sign data for each patient individually. For example, for one patient, a matrix AI holds values ​​for each vital sign i1 and time point i2. 1X I2 may be obtained and then processed. In such an embodiment, a matrix AI 1X Processing I2 is the matrix AI 1X This involves performing a singular value decomposition (SVD) technique on I2.

[0068] Equation 1 uses the matrix AI to approximate the original data matrix. 1X The SVD of I2 is applied to the R components.

number

[0069] In various embodiments, the multivariate intraoperative vital sign data in matrix AI1I2 for each of a cohort of patients may be concatenated into an I1×I2I3 matrix, whereby the SVD technique is performed on the larger matrix. A first dimension mode data object and a second dimension mode data object are then generated, where the second dimension mode data object (e.g., a temporal mode data object) may have length I2I3. However, the second dimension mode data object may not capture common temporal dynamics across patients.

[0070] To capture common temporal dynamics across patients, in various embodiments, a high-order singular value decomposition (HOSVD) technique is used to decompose the original data tensor

number

number

[0071] Analogous to SVD, the first dimension mode data object U (1) can be a prototypical pattern across intraoperative vital sign variable types (e.g., multivariate mode data objects), and U (2) can be temporal dynamics across intraoperative time points (e.g., temporal mode data objects). These multivariate and temporal mode data objects represent dynamics that are common among all patients in the cohort. A third dimensional mode data object U (3) may represent patient-specific variability or patient factors for multivariate temporal dynamics.

[0072] To capture the propagation dynamics, real-valued multivariate intraoperative vital sign data were used.

number

[0073] In various embodiments, the first dimension mode data object, the second dimension mode data object, and the third dimension mode data object may differ significantly across cohorts and correlate within cohorts. For example, the evolutionary dynamics of the multivariate intraoperative vital signs data may have at least one temporal mode that differs significantly between cohorts, or more specifically, performing a complex HOSVD technique on the multivariate intraoperative vital signs data may result in at least one second dimension mode data object that differs significantly between different cohorts.

[0074] In some embodiments, operation 502 includes creating a correlation entropy matrix based at least in part on a complex-valued third-order tensor X and performing a complex HOSVD technique on the correlation entropy matrix, which is referred to herein as a robust complex HOSVD technique or robust complex HOSVD.

[0075] In some embodiments, creating the correlation entropy matrix is ​​a complex-valued third-order tensor

number

[0076] Two stochastic processes

number

number

[0077] In Equation 3, E[·] is the stochastic process x t and y tWe denote the mathematical expectation over k(·,·) by a positive definite kernel function that satisfies the Mercer condition. By using a kernel function in the argument of the expectation operator, the kernel space induced by the correlation entropy contains the statistics of the data mapped to the new RKHS. By this means, the inner product in the new RKHS defines a metric that depends on the data statistics of the space spanned by the data, but responds to the overall data statistics, similar to the Mahalanobis distance, where the Mercer kernel space is used instead. In correlation entropy, the data statistics enter into the definition of the inner product. By choosing a symmetric positive definite kernel function, Equation 3 becomes a symmetric positive definite function, giving a translation-invariant similarity measure. Additionally, according to the Moore-Aronszajn theorem, there exists a unique RKHS associated with the correlation entropy function. Given that conventional correlation functions are not necessarily positive definite, such a RKHS associated with a correlation function does not exist. Thus, a substantial advantage of the cross-correlation entropy function is that it uses the structure of a unique RKHS in the definition of the similarity measure. One common kernel function used in the correlation entropy function is the Gaussian kernel given by Equation 4:

number

[0078] In Equation 4, σ represents the variance of the data, which is called the kernel width parameter or kernel size. The kernel width controls the influence of higher-order moments in the similarity assessment of Equation 3. By increasing the kernel width σ, the higher-order moments decay rapidly, and eventually the second-order moments become dominant. The cross-correlation entropy function then reduces to a conventional correlation. In contrast, when σ is too small, the data points are only similar to themselves. At this point, the kernel function approximates a Dirac delta function, and the cross-correlation entropy function no longer characterizes the statistics of the data. With an appropriate kernel width, the cross-correlation entropy function is l p -norm is estimated by weighting the higher order moments.

[0079] The cross-correlation entropy function shares with it the fact that it quantifies the similarity between pairs of lags in a time series. Considering that the time-varying content of intraoperative vital signs represents a multivariate stochastic process, a robust complex HOSVD technique can be built based on the cross-correlation entropy function.

[0080] In some embodiments, the moment matrix H (1) is the first mode matrix expansion of X (I1×I2I3)-matrix X (1) The cross-correlation entropy function is then generated based at least in part on the (I1 × I1) correlation entropy matrix V (1) To generate the moment matrix H (1) Similarly, the (I2 × I2) correlation entropy matrix V (2) and (I3×I3) correlation entropy matrix V (3) are the moment matrices H (2) and H (3) can be generated by applying the cross-correlation entropy function to the random processes contained in (2) and H (3) are generated based at least in part on a second mode matrix expansion of X and a third mode matrix expansion of X, respectively.

number

[0081] In some embodiments, performing the complex HOSVD technique on the correlation entropy matrix includes eigendecomposition of the correlation entropy matrix. The correlation entropy matrix is ​​similar to the covariance matrix in RKHS. Therefore, based on the spectral theory, there exists a set of orthonormal bases and a set of positive real eigenvalues, so that the correlation entropy matrix is ​​diagonal in this set of bases. In some embodiments, the eigendirections can be extracted through singular value decomposition of the correlation entropy matrix.

[0082] In order to apply the SVD procedure to the correlation entropy matrix, the data must be zero-mean in feature space. In some embodiments, the data can be centered by subtracting the information cross potential from the entries of the correlation entropy matrix. In some other embodiments, the widely used approach of the kernel method of removing the mean from the entries of the Gram matrix can be used and modified to center the correlation entropy matrix. For example,

number

number

[0083] In some embodiments, the eigenvectors can be obtained by singular value decomposition of a centered version of the correlation entropy matrix. For example, a first set of eigenvectors

number

number

number

[0084] In some embodiments, the first set of eigenvectors can be stored in a first dimension mode data object, the second set of eigenvectors can be stored in a second dimension mode data object, and the third set of eigenvectors can be stored in a third dimension mode data object.

[0085] Process 500 further includes operation 503. In various embodiments, operation 503 may follow operation 502. Operation 503 includes generating a cohort predictive model based at least in part on the plurality of first dimension mode data objects (e.g., multivariate mode data objects) and the plurality of second dimension mode data objects (e.g., temporal mode data objects).

[0086] In some embodiments, the first and second dimensional mode data objects extracted through applying complex HOSVD on the complex-valued tensor X are used to describe physiological dynamic correlations and provide insight into any lead-lag relationships between individual responses expressed in instantaneous phases of the complex vital signs in the cohort prediction model. For example, the first and second dimensional mode data objects may be combined (e.g., by cross product) to form various components, as described above. To obtain the most prominent multivariate and temporal factors for generating the cohort prediction model, a rank feature method based at least in part on the Fisher ranking technique may be used to select a number of top-ranked components. In various embodiments, the top three ranked components are selected. The cohort prediction model is then generated using the selected components. In various embodiments, the cohort prediction model includes an n-dimensional data manifold or structure, where n corresponds to the number of selected components. For example, the cohort prediction model includes a three-dimensional data manifold, where the third dimension of the data manifold is based at least in part on the three selected components. It may be appreciated that each cohort predictive model may be based at least in part on different dimensions. For example, a cohort predictive model for an orthopedic surgery cohort may have a dimension that strongly weights activation of blood oxygen level later in the intraoperative period, another dimension that strongly weights activation of respiratory tidal volume, and another dimension that strongly weights activation of a combination of heart rate and blood pressure early and late in the intraoperative period, while a cohort predictive model for a thoracic surgery cohort may have a dimension that strongly weights heart rate and another dimension that strongly weights blood pressure.

[0087] In some other embodiments, the first and second dimensional mode data objects extracted through applying complex HOSVD on the correlation entropy matrix are used to describe physiological dynamic correlations and provide insight into how intraoperative vital sign dynamics are associated with long-term postoperative pain development using a cohort prediction model. In some embodiments, a rank feature method based at least in part on the Fisher ranking technique may be used to select a number of top-ranked components from the extracted first and / or second dimensional mode data objects to obtain the most prominent multivariate and temporal factors for generating the cohort prediction model. In some embodiments, the top three ranked components providing the highest Fisher scores are selected to form a three-dimensional data manifold. Then, using the selected components, a cohort prediction model is generated. In various embodiments, the cohort prediction model includes an n-dimensional data manifold or structure, where n corresponds to the number of selected components. For example, the cohort prediction model includes a three-dimensional data manifold, where the third dimension of the data manifold is based at least in part on the three selected components. It may be appreciated that each cohort predictive model may be based at least in part on different dimensions.

[0088] As mentioned above, creating the correlation entropy matrix includes applying a cross-correlation entropy function to a random process included in one or more of the moment matrices, and the kernel width controls the influence of higher-order moments in the similarity assessment. Some embodiments show the relationship between the sparseness of the temporal factors and the value of the Fisher score obtained for the most prominent eigendirections (or the most prominent components from the extracted first dimension mode data object and / or second dimension mode data object). For example, as shown in FIG. 11, the Fisher score decreases for very small and very large kernel widths, which represents how the value of the Fisher score changes in the top 10 extracted components for different sets of kernel widths. FIGS. 9A and 9B show the first three temporal factors obtained using two different sets of kernel widths σ1=7.82, σ2=0.96 and σ1=782.12, σ2=96.65, respectively, where σ1 and σ2 are the eigenvalues ​​of the moment matrix V (1)c and V (2)c is the kernel width parameter associated with σ1=78.21, σ2=9.66. Figure 10 shows the same temporal factors obtained using the optimal kernel width σ1=78.21, σ2=9.66. The temporal factors achieved using kernel width σ1=78.21, σ2=9.66 are sparser than those obtained using kernel width σ1=7.82, σ2=0.96 and are denser than those obtained using kernel width σ1=782.12, σ2=96.65. In addition, as shown in Figure 11, for very small and very large kernel widths, the Fisher scores are distributed into different components, which is undesirable. On the other hand, for the optimal set of kernel widths, the top three components contain the highest Fisher scores and therefore show good performance for modeling the differences between categories of data.

[0089] 5A further illustrates process 500 including operation 504. In various embodiments, operation 504 may follow operation 503. Operation 504 includes generating and initializing a cohort predictive model with a plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and the respective binary classifications. As discussed above, each historical data object may be associated with a binary classification. The binary classification of the historical data object may be determined at least in part based on corresponding individuals of the cohort reporting average pain intensity on a numerical scale for a particular postoperative period, time frame, time point, and / or the like.

[0090] In some embodiments, initializing the cohort predictive model includes projecting each historical data object into the n-dimensional manifold of the cohort predictive model. As discussed above, each complex HOSVD component identifies sub-hemodynamic parameters (multivariate factors) that are instructionally activated across the individuals of the cohort and have common intraoperative temporal dynamics (temporal factors). Overall, the complex HOSVD model reveals a reasonable portrait of surgical dynamics (population dynamics) in which different subsets of hemodynamic parameters are active at different times during surgery, and the variation across the individuals of the cohort encodes the individual dynamic variables.

[0091] In some embodiments, when a complex HOSVD technique is used in operation 502, initializing the cohort prediction model further includes modifying the phase information of each historical data object based at least in part on a plurality of third dimension mode data objects or individual patient factors. In some embodiments, for a better representation of the dynamics, it may be beneficial to associate each principal component (as one basis of a subspace) with each dynamic mode of the individual responses of the cohort individuals encoded in the patient factors (e.g., third dimension mode data objects). In some embodiments, the coordinate systems provided by the common multivariate temporal factors and the multivariate temporal dynamics of the cohort individuals are not necessarily the same or precisely aligned. Given that all factors of the complex HOSVD are complex-valued factors, the patient-specific variation of the identified multivariate temporal dynamics includes scaling and rotation adjustments that appear in the cross product of the multivariate temporal dynamics with the patient factors.

[0092] It is essential to have a common coordinate system for all individuals of the cohort in order to compare the complex correlation between each hemodynamic response and the extracted multivariate temporal dynamics. At the same time, to take into account the dynamic variation between patients, in some embodiments, instead of rotating the dynamics, the complex conjugate of the elements given by the patient factors can be used to scale and rotate the hemodynamic response (e.g., phase information) before projecting it into the n-dimensional manifold of the cohort predictive model. The process can be performed separately for each complex HOSVD component. From a geometric point of view, this process can be considered as an active transformation in which the position of the points changes in the coordinate system, as opposed to a passive transformation that changes the coordinate system in which the points are described.

[0093] Thus, in some embodiments, the phase information for each historical data object is rectified, rotated, translated, and / or the like based at least in part on the patient factors represented in the multiple third dimension mode data objects and then projected onto an n-dimensional manifold (e.g., a three-dimensional manifold) of the cohort predictive model.

[0094] In some embodiments, the initialization of the cohort prediction model then includes training the cohort prediction model with a binary classification. For example, a linear discriminant analysis (LDA) can be performed to distinguish between historical data objects having a binary classification of mild persistent POP and historical data objects having a binary classification of severe persistent POP within the three-dimensional manifold. Thus, a relationship between phase information of multivariate intraoperative vital signs data representing the hemodynamic response of individuals of the cohort and mild or severe persistent POP can be determined.

[0095] As previously mentioned, the binary classification may be associated with a particular postoperative period, time frame, time point, and / or the like. For example, a first binary classification may be associated with mild or severe persistent POP at 30 days after surgery (e.g., post-op), and a second binary classification may be associated with mild or severe persistent POP at 90 days after surgery (e.g., post-op). Thus, a cohort prediction model may be initialized to determine a relationship between phase information and mild or severe persistent POP at a particular postoperative period, time frame, time point, and / or the like. In various embodiments, one or more cohort prediction models may be generated for a cohort, with each cohort prediction model associated with a particular postoperative period, time frame, time point, and / or the like. In various embodiments, one cohort prediction model may determine and store a relationship between phase information and various postoperative periods, time frames, time points, and / or the like. Thus, through process 500, the cohort predictive model may exploit the link between the dynamics of an individual's response to surgical stimuli and the development of long-term post-operative pain.

[0096] FIG. 6 illustrates a portion of six exemplary cohort prediction models with complex HOSVD applied in operation 502. Specifically, FIG. 6 illustrates various three-dimensional manifolds 600 (e.g., 600A-F), each initialized with topological information of historical data objects for a corresponding cohort. The three-dimensional manifolds 600 are extracted by applying complex HOSVD to a complex-valued tensor X. As previously mentioned, the cohorts may be surgical type cohorts. For example, three-dimensional manifold 600A corresponds to a thoracic surgery cohort, three-dimensional manifold 600B corresponds to an orthopedic surgery cohort, three-dimensional manifold 600C corresponds to a pancreatic / biliary surgery cohort, three-dimensional manifold 600D corresponds to a transplant surgery cohort, three-dimensional manifold 600E corresponds to a urological surgery cohort, and three-dimensional manifold 600F corresponds to a colorectal surgery cohort. It will be appreciated that in various embodiments, cohort predictive models may be generated and initialized for different surgical type cohorts and may also be associated with other cohorts, such as demographic cohorts.

[0097] The topological information of the various historical data objects after being projected onto each of the three-dimensional manifolds 600 is shown in FIG. 6. Each historical data object is also associated with either mild or severe persistent POP. Thus, using discriminant analysis techniques such as LDA, a relationship or correlation can be determined between the topological information and mild or severe persistent POP. For example, in the three-dimensional manifold 600D for the transplant surgery cohort, historical data objects having a binary classification of severe persistent POP have topological information that is negative in the first dimension of the manifold, positive in the second dimension of the manifold, and negative in the third dimension of the manifold, while historical data objects having a binary classification of mild persistent POP have topological information that is projected as positive in the first dimension of the manifold.

[0098] Thus, a relationship between phase information projected onto and / or associated with the dimensions of the three-dimensional manifold 600 and a binary classification of mild or severe POP may be determined. In some embodiments, each historical data object is associated with a non-binary classification indicative of persistent POP. For example, the non-binary classification may be a numerical value within a range of persistent POP representative values. In such embodiments, the cohort predictive model may be initialized with a multi-way discriminant analysis to determine a relationship between the phase information and each non-binary classification.

[0099] 8 illustrates some of six example cohort prediction models with robust complex HOSVD applied in operation 502. Specifically, FIG. 8 illustrates various three-dimensional manifolds 800 (e.g., 800A-F), each initialized with phase information of historical data objects for a corresponding cohort, where the three-dimensional manifolds 800 are extracted by applying complex HOSVD on a correlation entropy matrix generated from a complex-valued tensor X. As previously mentioned, the cohorts may be surgical type cohorts. For example, three-dimensional manifold 800A corresponds to a thoracic surgery cohort, three-dimensional manifold 800B corresponds to an orthopedic surgery cohort, three-dimensional manifold 800C corresponds to a pancreatic / biliary surgery cohort, three-dimensional manifold 800D corresponds to a transplant surgery cohort, three-dimensional manifold 800E corresponds to a urological surgery cohort, and three-dimensional manifold 800F corresponds to a colorectal surgery cohort. As described above, in various embodiments, it will be understood that cohort prediction models can be generated and initialized for different surgical type cohorts and can also be associated with other cohorts, such as demographic cohorts. After being projected onto each of the three-dimensional manifolds 800, the topological information of the various historical data objects is shown in FIG. 8. Each historical data object is also associated with either mild or severe persistent POP. Thus, using discriminant analysis techniques such as LDA, a relationship or correlation can be determined between the topological information and mild or severe persistent POP. Thus, a relationship between phase information projected onto and / or associated with the dimensions of the three-dimensional manifold 800 and a binary classification of mild or severe POP may be determined. In some embodiments, each historical data object is associated with a non-binary classification indicative of persistent POP. For example, the non-binary classification may be a numerical value within a range of persistent POP representative values. In such embodiments, the cohort predictive model may be initialized with a multi-way discriminant analysis to determine a relationship between the phase information and each non-binary classification.

[0100] Prediction Module Referring back to FIG. 4 , the system computing entity 102 may, in various embodiments, comprise a prediction module 420. The prediction module 420 may be configured to generate a risk prediction data object for the individual of interest. The risk prediction data object generated by the prediction module 420 may indicate at least a likelihood and / or classification of whether the individual of interest will experience mild or severe persistent POP. Thus, in various embodiments, the risk prediction data object includes a binary classification of mild or severe persistent POP. In other embodiments, the risk prediction data object includes a non-binary classification indicating the degree of persistent POP. In various embodiments, the risk prediction data object is associated with a particular post-operative time frame, time point, period, and / or the like (e.g., 30 days post-operative, 90 days post-operative). In various embodiments, the risk prediction data object includes a confidence score.

[0101] In various embodiments, the prediction module 420 may be configured to generate a risk prediction data object based at least in part on the cohort prediction model generated by the model generation module 410. For example, the prediction module 420 may communicate with the model generation module 410 to provide multivariate intraoperative vital sign data and / or phase information of the multivariate intraoperative vital sign data for the individual of interest, receive a classification (e.g., a binary classification of mild or severe persistent POP, a non-binary classification of the extent of persistent POP) from the cohort prediction model, etc. In an exemplary embodiment, the prediction module 420 may communicate with the model generation module 410 via a model API.

[0102] Thus, the system computing entity 102 (e.g., prediction module 420) is configured to perform operations for determining and predicting an individual's risk of developing persistent POP, such as those provided in Figure 5B. Figure 5B illustrates an exemplary process 510 for generating and determining a risk prediction data object for an individual indicating the likelihood and / or classification of whether the individual will experience persistent POP. In various embodiments, the system computing entity 102 comprises a processing element 205, memory 210, 215, network interface 220, and / or the like for performing each operation of the process 510.

[0103] 5B, process 510 includes operation 511. In various embodiments, process 510 may begin with operation 511. Operation 511 includes receiving a predictive input data object for an individual, the predictive input data object including multivariate intraoperative vital signs data associated with the individual. For example, the predictive input data object may be received via network interface 220 from another computing entity. As another example, the predictive input data object may be received via a user interface. In various embodiments, the predictive input data object may be received via an API call or query.

[0104] As previously discussed, multivariate intraoperative vital sign data includes data across multiple intraoperative time points for different vital sign variable types. For example, multivariate intraoperative vital sign data for an individual may include periodic measurements of heart rate, blood oxygen level, end-tidal CO2, respiratory tidal volume, systolic blood pressure, diastolic blood pressure, isoflurane concentration, sevoflurane concentration, and / or the like.

[0105] Process 510 further includes operation 512. In various embodiments, operation 512 may follow operation 511. Operation 512 includes processing the multivariate intraoperative vital signs data for the individual. In various embodiments, processing the prediction input data object includes complexifying the multivariate intraoperative vital signs data (e.g., by performing a Hilbert transform technique). Because only one individual is represented in the multivariate intraoperative vital signs data of the prediction input data object, higher order techniques (e.g., complex HOSVD) are not necessary because individual or patient dimensions are irrelevant. However, in various embodiments, persistent POP risk predictions for one or more individuals may be determined simultaneously by determining phase information using complex HOSVD.

[0106] Process 510 further includes operation 513. In various embodiments, operation 513 may follow operation 512. Operation 513 includes providing the processed (e.g., complex-valued) multivariate intraoperative vital sign data to a cohort predictive model associated with the cohort. In various embodiments, the processed (e.g., complex-valued) multivariate intraoperative vital sign data is provided to the cohort predictive model based at least in part on associating the predictive input data object with the cohort. In various embodiments, the cohort is a surgical type cohort. For example, the predictive input data object may be associated with one of: (i) a thoracic surgery cohort, (ii) an orthopedic surgery cohort, (iii) a urological surgery cohort, (iv) a colorectal surgery cohort, (v) a transplant surgery cohort, and (vi) a pancreatic / biliary surgery cohort.

[0107] In various embodiments, the predictive input data object may include additional data indicative of the cohort to which the predictive input data object should be associated, and by extension, the cohort prediction model to which the predictive input data object should be provided. For example, the predictive input data object may be associated with a particular surgical type cohort based at least in part on the medical records and / or instructions for a particular surgical type included in the predictive input data object. In various embodiments, the predictive input data object may be associated with a cohort based at least in part on an analysis of multivariate intraoperative vital sign data. It may be appreciated that various vital sign data patterns may exist specific to some surgical types, and thus, for example, a surgical type cohort may be determined based at least in part on the multivariate intraoperative vital sign data. In various embodiments, the predictive input data object may be associated with and / or classified as a particular surgical type cohort based at least in part on performing a supervised machine learning method.

[0108] Thus, the predictive input data object is provided to a cohort prediction model associated with the cohort associated with the predictive input data object or to which the individual of interest belongs. In various embodiments, a cohort may be associated with one or more cohort prediction models, each associated with a particular postoperative period, time frame, time point, and / or the like, and the predictive input data object is provided to each of the one or more cohort prediction models to generate one or more risk prediction data objects for different postoperative times. In other embodiments, a cohort may be associated with one cohort prediction model configured to provide a classification of persistent POP for different postoperative times, and the predictive input data object is provided to the cohort prediction model.

[0109] As illustrated in FIG. 5B, process 510 further includes operation 514. In various embodiments, operation 514 may follow operation 513. Operation 514 includes generating a risk prediction data object based at least in part on the cohort prediction model. In various embodiments, the cohort prediction model has been initialized and a relationship between phase information and a classification (e.g., binary, non-binary) of persistent POP has been determined. Thus, based at least in part on the phase information of the processed (e.g., complex-valued) multivariate intraoperative vital signs data of the prediction data object, a classification of a predicted risk of persistent POP for the individual of interest may be determined and generated. In various embodiments, the risk prediction data object includes a classification of a predicted risk of persistent POP for the individual.

[0110] Specifically, as discussed above, the cohort prediction model may include an n-dimensional manifold onto which the complexed multivariate intraoperative vital sign data of the individual of interest may be projected. In various embodiments, the cohort prediction model may be initialized with historical data objects (e.g., in operation 504) such that a classification for the individual of interest may be determined based at least in part on the projection of the complexed multivariate intraoperative vital sign data of the individual of interest. In some embodiments, a classification for the individual may be determined based at least in part on topological information of the projection of the complexed multivariate intraoperative vital sign data onto the n-dimensional manifold. In some embodiments, an axis within the n-dimensional manifold (e.g., the three-dimensional manifold 600) may be determined based at least in part on discriminant analysis (e.g., LDA), and a classification for the individual may be determined based at least in part on topological information of the projection of the complexed multivariate intraoperative vital sign data of the individual of interest onto the axis within the n-dimensional manifold. In various embodiments, a binary classification of mild or severe persistent POP for the individual may be determined. In various embodiments, a non-binary classification of the extent of persistent POP may be determined.

[0111] Additionally, the classification of the predicted risk of persistent POP for the subject individual may be associated with a particular post-operative period, time frame, time point, and / or the like. For example, the cohort prediction model may determine a relationship between phase information and persistent POP for 30 days post-operatively, and may use that relationship to determine a classification of the predicted risk of persistent POP at 30 days post-operatively for the subject individual.

[0112] Thus, the cohort prediction model may provide a classification and a risk prediction data-object may be generated that includes the classification. In various embodiments, the risk prediction data-object includes one or more classifications each associated with a different post-operative time, and thus the risk prediction data-object provides a predicted risk over the post-operative period. In various embodiments, the risk prediction data-object includes a confidence score in the classification or prediction. In various embodiments, the risk prediction data-object includes a selected n-dimension of the cohort prediction model.

[0113] 5B, process 510 further includes operation 515. In various embodiments, operation 515 may follow operation 514. Operation 515 includes performing one or more risk prediction-based actions for the individual. In various embodiments, the one or more risk prediction-based actions include displaying the risk prediction data object, a binary classification of whether the individual will develop persistent POP, and / or a binary classification of whether the individual will develop mild or severe persistent POP. In various embodiments, the first dimension mode data object and the second dimension mode data object may also be displayed. In various embodiments, the one or more risk prediction-based actions include sending the risk prediction data object to a client computing entity 106 associated with the individual. For example, the risk prediction data object may be provided in an API response in response to an API call.

[0114] 7, a diagram 700 for a general overview of predicting risk of persistent POP for an individual of interest is provided. As illustrated in diagram 700, various factors during a surgical procedure 702 may affect a patient's autonomic state 704, manifested as multivariate intraoperative vital signs data 706. For example, surgical stimuli and inputs, anesthesia inputs, and physiological support may all affect the patient's autonomic state 704. The patient's autonomic state 704 is reflected in the multivariate intraoperative vital signs data 706, or the observed acute physiological response to the surgical procedure 702.

[0115] As illustrated, at operation 710, a complex HOSVD technique may be performed on the multivariate intraoperative vital signs data 706 to determine and extract phase information from the multivariate intraoperative vital signs data 706. Such phase information may be visualized and / or displayed at operation 712 along with additional data, such as dimensional mode data objects. Meanwhile, the phase information determined from performing the complex HOSVD technique at operation 710 may be used to determine and predict postoperative outcomes at operation 714. That is, a prediction of whether an individual may develop persistent POP and / or whether an individual may develop mild or severe persistent POP may be determined at operation 714 based at least in part on the phase information determined from the complex HOSVD technique. The prediction of postoperative outcomes may be further processed or applied, such as to determine postoperative opioid requirements, or other medication requirements.

[0116] V. Computer Program Products The embodiments of the present disclosure may be implemented in a variety of ways, including computer program products constituting an article of manufacture. Such computer program products may include, for example, one or more software components including software objects, methods, data structures, and / or the like. The software components may be coded in any of a variety of programming languages. An exemplary programming language may be a low-level programming language, such as assembly language, associated with a particular hardware architecture and / or operating system platform. A software component including assembly language instructions may require translation by an assembler into executable machine code prior to execution by the hardware architecture and / or platform. Another exemplary programming language may be a high-level programming language that may be portable across multiple architectures. A software component including high-level programming language instructions may require translation into an intermediate representation by an interpreter or compiler prior to execution.

[0117] Other examples of programming languages ​​include, but are not limited to, macro languages, shell or command languages, job control languages, scripting languages, database query or search languages, and / or report writing languages. In one or more exemplary embodiments, a software component that includes instructions in one of the foregoing examples of programming languages ​​may be executed directly by an operating system or other software components without first being converted into another format. The software components may be stored as files or other data storage structures. Software components of similar type or functionally related may be stored together, for example, in a particular directory, folder, or library. The software components may be static (e.g., pre-established or fixed) or dynamic (e.g., created or changed at run time).

[0118] A computer program product may include a non-transitory computer readable storage medium that stores applications, programs, program modules, scripts, source code, program code, object code, byte code, compiled code, interpreted code, machine code, executable instructions, and / or the like (also referred to herein as executable instructions, instructions for execution, computer program products, program code, and / or similar terms used interchangeably herein). Such non-transitory computer readable storage media include all computer readable media (including volatile and non-volatile media).

[0119] In one embodiment, the non-volatile computer readable storage medium may include a floppy disk, a flexible disk, a hard disk, a solid state storage device (SSS) (e.g., a solid state drive (SSD), a solid state card (SSC), a solid state module (SSM), an enterprise flash drive, a magnetic tape, or any other non-transitory magnetic medium, and / or the like. The non-volatile computer readable storage medium may also include a punch card, a paper tape, an optical mark sheet (or any other physical medium having a pattern of holes or other optically recognizable indicia), a compact disk read only memory (CD-ROM), a compact disk rewriteable (CD-RW), a digital versatile disk (DVD), a Blu-ray disk (BD), any other non-transitory optical medium, and / or the like. Such a non-volatile computer readable storage medium may also include a read only memory (ROM), a programmable read only memory (ROM), a programmable read only memory (PG), a programmable write only memory (PG), a programmable write only memory (RW ... The non-volatile computer readable storage medium may include a programming and / or read only memory (PROM), an erasable programmable read only memory (EPROM), an electrically erasable programmable read only memory (EEPROM), a flash memory (e.g., serial, NAND, NOR, and / or the like), a multimedia memory card (MMC), a secure digital (SD) memory card, a smart media card, a compact flash (CF) card, a memory stick, and / or the like. Furthermore, the non-volatile computer readable storage medium may also include a conductive bridging random access memory (CBRAM), a phase change random access memory (PRAM), a ferroelectric random access memory (FeRAM), a non-volatile random access memory (NVRAM), a magnetoresistive random access memory (MRAM), a resistive random access memory (RRAM), a silicon oxide nitride oxide silicon memory (SONOS), a floating junction gate random access memory (FJG RAM), a Millipede memory, a racetrack memory, and / or the like.

[0120] In one embodiment, the volatile computer readable storage medium may include random access memory (RAM), dynamic random access memory (DRAM), static random access memory (SRAM), fast page mode dynamic random access memory (FPM DRAM), enhanced data output dynamic random access memory (EDO DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), double data rate type 2 synchronous dynamic random access memory (DDR2 SDRAM), double data rate type 3 synchronous dynamic random access memory (DDR3 SDRAM), Rambus dummy random access memory (RDRAM), twin transistor RAM (TTRAM), thyristor RAM (T-RAM), zero capacitor (Z-RAM), Rambus in-line memory module (RIMM), dual in-line memory module (DIMM), single in-line memory module (SIMM), video random access memory (VRAM), cache memory (including various levels), flash memory, register memory, and / or the like. Where embodiments are described as using a computer-readable storage medium, it will be understood that other types of computer-readable storage media may be used instead of or in addition to the computer-readable storage media described above.

[0121] As should be understood, various embodiments of the present disclosure may also be implemented as methods, apparatus, systems, computing devices, computing entities, and / or the like. Thus, embodiments of the present disclosure may take the form of data structures, apparatus, systems, computing devices, computing entities, and / or the like that execute instructions stored on a computer-readable storage medium to perform certain steps or operations. Thus, embodiments of the present disclosure may also take the form of entirely hardware embodiments, entirely computer program product embodiments, and / or embodiments that include a combination of a computer program product and hardware that performs certain steps or operations.

[0122] The embodiments of the present disclosure are described above with reference to block diagrams and flowchart diagrams. It should therefore be understood that each block of the block diagrams and flowchart diagrams may be implemented in the form of a computer program product, a complete hardware embodiment, a combination of hardware and computer program products, and / or an apparatus, system, computing device, computing entity, and / or the like that executes instructions, operations, steps, and similar terms used interchangeably (e.g., executable instructions, instructions for execution, program code, and / or the like) on a computer-readable storage medium for execution. For example, the acquisition, loading, and execution of code may be performed sequentially, such that one instruction is acquired, loaded, and executed at a time. In some exemplary embodiments, the acquisition, loading, and / or execution may be performed in parallel, such that multiple instructions are acquired, loaded, and / or executed together. Thus, such embodiments may produce a specifically configured machine that executes the steps or operations specified in the block diagrams and flowchart diagrams. Thus, the block diagrams and flowchart diagrams support various combinations of the embodiments for executing the specified instructions, operations, or steps.

[0123] VI. Conclusion It should be understood that the examples and embodiments described herein are for illustrative purposes only, and various modifications or changes in light of them may be suggested to those skilled in the art and are within the spirit and scope of the present application. Although this disclosure is deemed complete and comprehensive, additional context and insight may be gleaned from the appendices attached herewith (which generally describe systems, devices, and methods according to embodiments of the present specification). It should be understood that the examples and embodiments in Appendices A and B are also for illustrative purposes and are non-limiting in nature. The contents of Appendices A and B are incorporated herein by reference in their entirety.

[0124] Many modifications and other embodiments of the disclosure described herein will come to mind to one skilled in the art to which this disclosure pertains having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. It is to be understood, therefore, that the disclosure is not limited to the specific embodiments disclosed, and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.

Claims

1. 1. A computer-implemented method for predicting a risk of persistent post-operative pain for an individual, the computer-implemented method comprising: receiving, by a processor, a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with postoperative time points, the cohort predictive model being: receiving a historical data object for each of a cohort including a plurality of individuals, each historical data object being associated with a binary classification, each historical data object including multivariate intraoperative vital signs data for a corresponding individual; processing the plurality of historical data objects to generate a plurality of first dimension mode data objects, a plurality of second dimension mode data objects, and a plurality of third dimension mode data objects; generating a cohort prediction model based at least in part on the plurality of first dimension mode data objects and the plurality of second dimension mode data objects, wherein the plurality of first dimension mode data objects and the plurality of second dimension mode data objects are processed to generate a three-dimensional manifold; initializing the cohort predictive model with the plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and each binary classification; and generating a risk prediction data object comprising a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; performing one or more risk prediction-based actions for the individual; and 4. A computer-implemented method comprising:

2. A computer-implemented method for predicting a risk of persistent postoperative pain for an individual, the computer-implemented method comprising: receiving, by a processor, a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with post-operative time points; generating a risk prediction data object comprising a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; performing one or more risk prediction-based actions for the individual; and Including, 22. The computer-implemented method of claim 21, wherein processing the multivariate intraoperative vital sign data includes complexifying the multivariate intraoperative vital sign data for the individual, and providing at least the processed multivariate intraoperative vital sign data to a cohort predictive model includes projecting the processed multivariate intraoperative vital sign data onto a three-dimensional manifold of the cohort predictive model, and determining phase information of the projection of the processed multivariate intraoperative vital sign data.

3. 2. The computer-implemented method of claim 1, wherein the multiple historical data objects are aggregated and processed together using a complex higher-order singular value decomposition (HOSVD), and the three-dimensional manifold is generated based at least in part on ranks of components generated by the HOSVD.

4. each of the plurality of first dimension-mode data objects includes a weight for each of one or more vital sign variable types; each of the plurality of second dimension mode data objects includes a weight for each of a plurality of intraoperative time points; The computer-implemented method of claim 1 , wherein each of the plurality of third dimension mode data objects includes a weight for each of the plurality of individuals.

5. 2. The computer-implemented method of claim 1, wherein initializing the cohort predictive model comprises determining a relationship between topological information of projections of the plurality of historical data objects onto the three-dimensional manifold and a binary classification.

6. the plurality of first dimension-mode data objects include eigenvectors of a first correlation entropy matrix, the first correlation entropy matrix being generated based at least in part on the plurality of historical data objects; the plurality of second dimension mode data objects include eigenvectors of a second correlation entropy matrix, the second correlation entropy matrix being generated based at least in part on the plurality of historical data objects; 2. The computer-implemented method of claim 1, wherein the plurality of third dimension mode data objects comprises eigenvectors of a third correlation entropy matrix, the third correlation entropy matrix being generated based at least in part on the plurality of historical data objects.

7. the first correlation entropy matrix is ​​generated by applying a first cross-correlation entropy function to a first moment matrix, the first moment matrix being generated based at least in part on a first mode matrix expansion of a third order tensor; the second correlation entropy matrix is ​​generated by applying a second cross-correlation entropy function to a second moment matrix, the second moment matrix being generated based at least in part on a second mode matrix expansion of the third-order tensor; 7. The computer-implemented method of claim 6, wherein the third correlation entropy matrix is ​​generated by applying a third cross-correlation entropy function to a third moment matrix, the third moment matrix being generated based at least in part on a third mode matrix expansion of the third order tensor, the third order tensor representing the plurality of historical data objects.

8. The computer-implemented method of claim 7 , wherein each of the first cross-correlation entropy function, the second cross-correlation entropy function, and the third cross-correlation entropy function is based on a Gaussian function.

9. A computer-implemented method for predicting a risk of persistent postoperative pain for an individual, the computer-implemented method comprising: receiving, by a processor, a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with post-operative time points; generating a risk prediction data object comprising a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; performing one or more risk prediction-based actions for the individual; and Including, 20. The computer-implemented method of claim 19, wherein the one or more risk prediction based actions for the individual include displaying the risk prediction data object in a three-dimensional manifold, the three-dimensional manifold being generated based at least in part on the historical data object.

10. 1. A device for predicting the risk of persistent post-operative pain for an individual, comprising: The apparatus comprises at least one processor and at least one non-transitory memory containing program code; The at least one non-transitory memory and the program code are configured to, in the at least one processor, at least receiving a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with postoperative time points, the cohort predictive model being: receiving a historical data object for each of a cohort including a plurality of individuals, each historical data object being associated with a binary classification, each historical data object including multivariate intraoperative vital signs data for a corresponding individual; processing the plurality of historical data objects to generate a plurality of first dimension mode data objects, a plurality of second dimension mode data objects, and a plurality of third dimension mode data objects; generating a cohort prediction model based at least in part on the plurality of first dimension mode data objects and the plurality of second dimension mode data objects, wherein the plurality of first dimension mode data objects and the plurality of second dimension mode data objects are processed to generate a three-dimensional manifold; initializing the cohort predictive model with the plurality of historical data objects based at least in part on the plurality of third dimension mode data objects and each binary classification; and generating a risk prediction data object comprising a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; performing one or more risk prediction-based actions for the individual; and The apparatus is configured to cause the apparatus to perform the steps of:

11. An apparatus for predicting risk of persistent postoperative pain for an individual, comprising: The apparatus comprises at least one processor and at least one non-transitory memory containing program code; The at least one non-transitory memory and the program code are configured to, in the at least one processor, at least receiving a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with post-operative time points; generating a risk prediction data object including a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; performing one or more risk prediction-based actions for the individual; and and configured to cause the device to The apparatus, wherein processing the multivariate intraoperative vital sign data includes complexifying the multivariate intraoperative vital sign data of the individual, and providing at least the processed multivariate intraoperative vital sign data to a cohort predictive model includes projecting the processed multivariate intraoperative vital sign data onto a three-dimensional manifold of the cohort predictive model, and determining phase information of the projection of the processed multivariate intraoperative vital sign data.

12. 11. The apparatus of claim 10, wherein the multiple historical data objects are aggregated and processed together using complex higher-order singular value decomposition (HOSVD), and the three-dimensional manifold is generated based at least in part on ranks of components generated by the HOSVD.

13. each of the plurality of first dimension-mode data objects includes a weight for each of one or more vital sign variable types; each of the plurality of second dimension mode data objects includes a weight for each of a plurality of intraoperative time points; The apparatus of claim 10 , wherein each of the plurality of third dimension mode data objects includes a weight for each of the plurality of individuals.

14. The apparatus of claim 10 , wherein initializing the cohort predictive model comprises determining a relationship between topological information of projections of the plurality of historical data objects onto the three-dimensional manifold and a binary classification.

15. the plurality of first dimension-mode data objects include eigenvectors of a first correlation entropy matrix, the first correlation entropy matrix being generated based at least in part on the plurality of historical data objects; the plurality of second dimension mode data objects include eigenvectors of a second correlation entropy matrix, the second correlation entropy matrix being generated based at least in part on the plurality of historical data objects; The apparatus of claim 10 , wherein the plurality of third dimension mode data objects comprises eigenvectors of a third correlation entropy matrix, the third correlation entropy matrix being generated based at least in part on the plurality of historical data objects.

16. the first correlation entropy matrix is ​​generated by applying a first cross-correlation entropy function to a first moment matrix, the first moment matrix being generated based at least in part on a first mode matrix expansion of a third order tensor; the second correlation entropy matrix is ​​generated by applying a second cross-correlation entropy function to a second moment matrix, the second moment matrix being generated based at least in part on a second mode matrix expansion of the third-order tensor; 16. The apparatus of claim 15, wherein the third correlation entropy matrix is ​​generated by applying a third cross-correlation entropy function to a third moment matrix, the third moment matrix being generated based at least in part on a third mode matrix expansion of the third order tensor, the third order tensor representing the plurality of historical data objects.

17. The apparatus of claim 16 , wherein each of the first, second and third cross-correlation entropy functions is based on a Gaussian function.

18. An apparatus for predicting risk of persistent postoperative pain for an individual, comprising: The apparatus comprises at least one processor and at least one non-transitory memory containing program code; The at least one non-transitory memory and the program code are configured to, in the at least one processor, at least receiving a predicted input data object including multivariate intraoperative vital signs data for the individual; processing the multivariate intraoperative vital signs data for the individual; and providing at least the processed multivariate intraoperative vital signs data to a cohort predictive model associated with the cohort of individuals, the cohort predictive model being initialized with historical data objects associated with post-operative time points; generating a risk prediction data object including a classification of phase information determined based at least in part on the cohort prediction model, the risk prediction data object being associated with the post-operative time point; performing one or more risk prediction-based actions for the individual; and and configured to cause the device to The one or more risk prediction based actions for the individual include displaying the risk prediction data object in a three-dimensional manifold, the three-dimensional manifold being generated based at least in part on the historical data object.

Images