Heart great vessel postoperative pulmonary complication prediction method and device and processing equipment

Abstract

The invention provides a heart great vessel postoperative pulmonary complication prediction method and device and processing equipment, which are used for constructing a heart great vessel postoperative pulmonary complication prediction model on the basis of a machine learning model, so that the heart great vessel postoperative pulmonary complication prediction model has high pertinence to heart great vessel postoperative pulmonary complication and is high in accuracy. And a series of optimal configuration schemes are provided in the aspect of details, so that a good prediction effect can be realized, the high-quality prediction requirement on the high-precision pulmonary complications after the heart great vessel operation in clinical work is met, and high-quality data support is provided for the clinical work.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and specifically to a method, device and processing equipment for predicting pulmonary complications after cardiac and great vascular surgery. Background Art

[0002] Postoperative pulmonary complications (PPCs) include almost all complications that affect the respiratory system after anesthesia and surgery. In recent years, the number of institutions performing cardiac surgery and the number of surgeries have increased year by year, but the incidence of postoperative pulmonary complications after cardiac surgery has not decreased significantly. If the risk of postoperative pulmonary complications in patients can be predicted in advance, it will be of great significance for preventing and treating postoperative pulmonary complications, improving the safety of surgical operations, and improving patient prognosis.

[0003] Regarding the prediction of postoperative pulmonary complications, the Pulmonary Risk Score (PRS) and the Surgical Lung Injury Prediction model (SLIP) are relatively mature tools, but they are general scales and lack operability for practical application in different clinical settings. Studies on postoperative pulmonary complications after surgeries at different sites include neuromuscular scoliosis surgery, liver transplantation, esophageal and kidney transplantation, etc., but there is no research on the prediction model of postoperative pulmonary complications after cardiac and vascular surgery (including heart and great blood vessels). In this regard, the research results of postoperative pulmonary complications after surgeries at other sites cannot be directly used.

[0004] In other words, there is currently no convenient prediction tool for postoperative pulmonary complications of cardiac and vascular surgery. Summary of the invention

[0005] The present application provides a method, device and processing equipment for predicting pulmonary complications after cardiac and great vascular surgery, which are used to construct a prediction model for pulmonary complications after cardiac and great vascular surgery based on a machine learning model, so as to be highly targeted at pulmonary complications after cardiac and great vascular surgery, and provide a series of optimized configuration schemes in details, thereby achieving good prediction results, meeting the high-quality prediction needs for high-precision pulmonary complications after cardiac and great vascular surgery in clinical work, and providing high-quality data support for clinical work.

[0006] In a first aspect, the present application provides a method for predicting pulmonary complications after cardiac and vascular surgery, the method comprising:

[0007] Acquire patient data recorded on the system of a target patient, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data;

[0008] Inputting the patient data into a pre-configured cardiac and great vascular surgery pulmonary complication prediction model, so that the cardiac and great vascular surgery pulmonary complication prediction model carries out corresponding cardiac and great vascular surgery pulmonary complication prediction processing, wherein the cardiac and great vascular surgery pulmonary complication prediction model is obtained by training an initial machine learning model with sample patient data labeled with corresponding cardiac and great vascular surgery pulmonary complication occurrence results;

[0009] Extract the prediction results of pulmonary complications after cardiac and great vascular surgery output by the prediction model for pulmonary complications after cardiac and great vascular surgery.

[0010] In a second aspect, the present application provides a device for predicting pulmonary complications after cardiac and great vascular surgery, the device comprising:

[0011] An acquisition unit, used for acquiring patient data of a target patient recorded on the system, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data;

[0012] A prediction unit, used for inputting patient data into a pre-configured cardiac and great vascular surgery pulmonary complication prediction model, so that the cardiac and great vascular surgery pulmonary complication prediction model carries out corresponding cardiac and great vascular surgery pulmonary complication prediction processing, wherein the cardiac and great vascular surgery pulmonary complication prediction model is obtained by training an initial machine learning model with sample patient data annotated with corresponding cardiac and great vascular surgery pulmonary complication occurrence results;

[0013] The extraction unit is used to extract the prediction results of pulmonary complications after cardiac and great vessel surgery output by the prediction model for pulmonary complications after cardiac and great vessel surgery.

[0014] In a third aspect, the present application provides a processing device, including a processor and a memory, wherein a computer program is stored in the memory, and when the processor calls the computer program in the memory, the method provided in the first aspect of the present application or any possible implementation method of the first aspect of the present application is executed.

[0015] In a fourth aspect, the present application provides a computer-readable storage medium, which stores multiple instructions, and the instructions are suitable for a processor to load to execute the method provided by the first aspect of the present application or any possible implementation of the first aspect of the present application.

[0016] From the above content, it can be concluded that the present application has the following beneficial effects:

[0017] Focusing on the goal of predicting pulmonary complications after cardiac and great vascular surgery, this application constructs a prediction model for pulmonary complications after cardiac and great vascular surgery based on a machine learning model, which is highly targeted at pulmonary complications after cardiac and great vascular surgery and provides a series of optimization configuration schemes in details, thereby achieving good prediction results, meeting the clinical work's demand for high-quality prediction of high-precision pulmonary complications after cardiac and great vascular surgery, and providing high-quality data support for clinical work. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 A schematic diagram of a process for predicting pulmonary complications after cardiac and vascular surgery in this application;

[0020] Figure 2 A schematic diagram of the structure of a device for predicting pulmonary complications after cardiac and vascular surgery in this application;

[0021] Figure 3 A schematic diagram of the structure of the processing equipment of this application. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0023] The terms "first", "second", etc. in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described here can be implemented in an order other than that illustrated or described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or modules is not necessarily limited to those steps or modules clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. The naming or numbering of steps in this application does not mean that the steps in the method flow must be executed in the time / logical sequence indicated by the naming or numbering. The process steps that have been named or numbered can change the execution order according to the technical purpose to be achieved, as long as the same or similar technical effects can be achieved.

[0024] The division of modules in this application is a logical division. There may be other division methods when it is implemented in actual applications. For example, multiple modules can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection between modules can be electrical or other similar forms, which are not limited in this application. In addition, the modules or submodules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed in multiple circuit modules, and some or all of the modules may be selected according to actual needs to achieve the purpose of the present application.

[0025] Before introducing the method for predicting pulmonary complications after cardiac and great vascular surgery provided by the present application, the background content involved in the present application is first introduced.

[0026] The method, device and computer-readable storage medium for predicting pulmonary complications after cardiac and great vessel surgery provided in the present application can be applied to processing equipment to construct a prediction model for pulmonary complications after cardiac and great vessel surgery based on a machine learning model, thereby being highly targeted at pulmonary complications after cardiac and great vessel surgery and providing a series of optimized configuration schemes in details, thereby achieving good prediction results, meeting the demand for high-quality prediction of high-precision pulmonary complications after cardiac and great vessel surgery in clinical work, and providing high-quality data support for clinical work.

[0027] The method for predicting pulmonary complications after cardiac and great vascular surgery mentioned in the present application can be executed by a device for predicting pulmonary complications after cardiac and great vascular surgery, or a server, physical host, or user equipment (UE) and other different types of processing devices that integrate the device for predicting pulmonary complications after cardiac and great vascular surgery. The device for predicting pulmonary complications after cardiac and great vascular surgery can be implemented in hardware or software, and the UE can be a terminal device such as a smart phone, tablet computer, laptop computer, desktop computer, or personal digital assistant (PDA), and the processing device can be set in the form of a device cluster.

[0028] Among them, it can be understood that the patient data involved in the present application are usually ready-made data, rather than data that need to be actually collected during the implementation of the program. Therefore, the processing equipment that executes the method for predicting pulmonary complications after cardiac and great vascular surgery of the present application, or the processing equipment equipped with the corresponding application service of the method for predicting pulmonary complications after cardiac and great vascular surgery of the present application, usually only needs to meet the required data processing capabilities, and its specific device type or device deployment form is more flexible in actual situations.

[0029] As an example, the processing device may be a device used by doctors in clinical practice, and the present application may be used to predict pulmonary complications after cardiac and vascular surgery.

[0030] Furthermore, corresponding to the situation that in addition to the application of the model, the present application may also involve model training. The processing device may also be divided into a first part responsible for model training and a second part responsible for model application. For example, the background server performs model training, and then sends the model to the front desk (doctor side) through remote service for specific application. This is also possible.

[0031] Next, the method for predicting pulmonary complications after cardiac and great vascular surgery provided by the present application is introduced.

[0032] First, see Figure 1 , Figure 1 A schematic flow chart of the method for predicting pulmonary complications after cardiac and great vessel surgery provided by the present application is shown. The method for predicting pulmonary complications after cardiac and great vessel surgery provided by the present application may specifically include the following steps S101 to S103:

[0033] Step S101, obtaining patient data of a target patient recorded on the system, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data;

[0034] It can be understood that for the convenience of explanation, the current patient is recorded as the target patient.

[0035] For patients who have undergone / performed cardiovascular surgery, there will be corresponding records in the hospital's relevant systems, such as the Hospital Information System (HIS), that is, archives, and the patient data required for this application may be included in these data.

[0036] Among them, cardiac and vascular surgery itself is an existing clinical surgery, such as coronary artery bypass surgery, valve replacement surgery and extracorporeal circulation surgery.

[0037] It can be understood that the patient data obtained here usually does not involve patient privacy, or the patient privacy can be removed again after the data is obtained.

[0038] For this patient's data, it can be seen that it involves four major aspects: general information, preoperative factors, intraoperative factors, and postoperative factors.

[0039] As an example, for general information, it may be related to:

[0040] Age, gender, height, weight, smoking, drinking and chronic comorbidities, length of hospital stay, length of stay in the intensive care unit (ICU) and discharge diagnosis, etc.

[0041] Preoperative factors may include respiratory tract infection in the past month, antibiotic use, laboratory results (white blood cell count, neutrophil count, hemoglobin, blood creatinine, serum albumin and fasting blood glucose) and ejection fraction.

[0042] Intraoperative factors may include:

[0043] Whether it is an emergency surgery, the method of anesthesia, the surgery (type, site and duration), extracorporeal circulation time, blood loss, blood transfusion, preoperative antibiotic use, additional antibiotics during surgery, nasogastric tube and urinary catheter, etc.

[0044] Postoperative factors may include:

[0045] The occurrence and types of postoperative pulmonary complications, endotracheal intubation, tracheotomy, nebulization inhalation, sedation time, intraaortic balloon pumping, dialysis, body temperature, blood gas analysis results from the day after surgery to the first day (blood glucose, arterial oxygen partial pressure (PaO2), fractional oxygen concentration in inspired air (FiO2), blood lactate (Lac), blood oxygen saturation (SpO2), total hemoglobin (THbc)) and laboratory results from the day after surgery to the first day (white blood cell count, neutrophil count, hemoglobin, blood creatinine, serum albumin), occurrence of nosocomial infection, death and outcome, etc.

[0046] Step S102, inputting patient data into a pre-configured cardiac and great vascular surgery pulmonary complication prediction model, so that the cardiac and great vascular surgery pulmonary complication prediction model carries out corresponding cardiac and great vascular surgery pulmonary complication prediction processing, wherein the cardiac and great vascular surgery pulmonary complication prediction model is obtained by training an initial machine learning model with sample patient data annotated with corresponding cardiac and great vascular surgery pulmonary complication occurrence results;

[0047] It is easy to see that, in response to the demand for predicting pulmonary complications after cardiac and great vascular surgery, the present application configures a model for predicting pulmonary complications after cardiac and great vascular surgery. The model for predicting pulmonary complications after cardiac and great vascular surgery is a machine learning model, which is obtained by training an initial machine learning model with sample patient data that is labeled with corresponding results of pulmonary complications after cardiac and great vascular surgery (the annotations are used as true values ​​to guide the model for prediction).

[0048] Among them, for the model architecture of the prediction model for pulmonary complications after cardiac and vascular surgery, it can adopt the existing model architecture, or it can be a model architecture obtained by further improving and optimizing the existing model architecture, or it can be a self-developed, novel model architecture, all of which are possible in actual situations. Similarly, the model training architecture and loss function used in the model training process also have this feature.

[0049] Generally speaking, model training involves the following steps:

[0050] In each round of model training, the training samples are input into the model so that the model can carry out corresponding prediction processing and realize forward propagation. Then, the loss function is calculated based on the prediction results output by the model, and the model parameters are optimized according to the loss function calculation results to realize back propagation. In this way, when the preset model training requirements such as training time, number of trainings or prediction accuracy are met, the model training can be completed. At this time, the model is a prediction model for pulmonary complications after cardiac and vascular surgery that can be put into actual use.

[0051] In this case, the patient data of the current target patient can be input into a pre-configured / trained model for predicting pulmonary complications after cardiac and great vessel surgery, so that it can carry out the corresponding prediction processing for pulmonary complications after cardiac and great vessel surgery and obtain the prediction results for pulmonary complications after cardiac and great vessel surgery.

[0052] Step S103, extracting the prediction result of pulmonary complications after cardiac and great vessel surgery output by the prediction model of pulmonary complications after cardiac and great vessel surgery.

[0053] It is easy to understand that after completing the prediction process of pulmonary complications after cardiac and great vessel surgery, the prediction model for pulmonary complications after cardiac and great vessel surgery will output the corresponding prediction results of pulmonary complications after cardiac and great vessel surgery. At this time, the prediction results of pulmonary complications after cardiac and great vessel surgery output by the model can be extracted.

[0054] Among them, the prediction results of pulmonary complications after cardiac and major vascular surgery can be expressed in specific forms such as risk scores and probability values.

[0055] After obtaining the prediction results of pulmonary complications after cardiac and great vascular surgery output by the model, it can be understood that the prediction of pulmonary complications after cardiac and great vascular surgery can be stored, displayed, forwarded, output prompts, etc. It is easy to understand that the next step of data application can be adjusted according to actual needs, specifically according to the pre-configured data application strategy and the currently configured data application strategy, and this application does not make specific limitations.

[0056] As an example, the prediction results of pulmonary complications after cardiovascular surgery can be shown to doctors to prompt them to arrange the next treatment plan for the target patient based on the situation.

[0057] As yet another example, the system may be triggered to analyze further treatment options for pulmonary complications in order to provide a reference for treatment options for doctors.

[0058] Obviously, these two examples are further extensions based on the predicted results in actual situations, and medical assistance plans can be taken to help take appropriate measures for high-risk groups, assist doctors in diagnosis and treatment, reduce the incidence of lung complications, and ensure timely treatment of lung complications.

[0059] from Figure 1 It can be seen from the illustrated embodiments that the goal of predicting pulmonary complications after cardiac and great vascular surgery is focused on. This application constructs a prediction model for pulmonary complications after cardiac and great vascular surgery based on a machine learning model, which is highly targeted at pulmonary complications after cardiac and great vascular surgery and provides a series of optimization configuration schemes in details, thereby achieving good prediction results, meeting the demand for high-quality prediction of high-precision pulmonary complications after cardiac and great vascular surgery in clinical work, and providing high-quality data support for clinical work.

[0060] Continue to the above Figure 1 Each step of the illustrated embodiment and possible implementation methods thereof in practical applications are described in detail.

[0061] As an exemplary embodiment, for the cardiac and vascular surgery pulmonary complications prediction model configured in the present application, the cardiac and vascular surgery pulmonary complications that can be predicted may specifically include pulmonary infection, atelectasis, pneumothorax, respiratory failure, acute respiratory distress syndrome (ARDS) and pulmonary edema.

[0062] Among them, for lung infection: it refers to the patient's new postoperative X-ray / CT and other imaging data confirmed by persistent patchy lung exudates; accompanied by corresponding clinical symptoms and signs such as purulent sputum, lung rales, fever, hypoxia, etc.; deep sputum specimens or sputum expelled by coughing are aspirated through endotracheal intubation for more than 2 times, and the bacterial culture results of qualified sputum specimens show positive results for the same pathogen. If two of the above three are met, the diagnosis is confirmed.

[0063] For atelectasis: it refers to atelectasis of the lobe or the entire lung with medical evidence shown by X-ray or CT, accompanied by weakened or disappeared breath sounds in the affected lung, or it can be confirmed during bronchoscopic suction treatment.

[0064] For pneumothorax: it refers to the pneumothorax confirmed by imaging that occurs again after the closed chest drainage tube is removed after surgery. If thoracentesis and aspiration are ineffective, the patient needs to undergo closed chest tube drainage again to be diagnosed.

[0065] For respiratory failure: When the patient has clinical manifestations such as dyspnea, cyanosis, altered consciousness, circulatory changes, and the blood gas analysis index, arterial oxygen partial pressure (PaO2) is lower than 60mmHg (1mmHg = 0.133kPa) or (and) arterial carbon dioxide partial pressure (PaCO2) is higher than 50mmHg, the patient is diagnosed with respiratory failure. Respiratory failure is divided into two types: Type I respiratory failure (hypoxemic respiratory failure): arterial oxygen partial pressure is lower than 60mmHg and arterial carbon dioxide partial pressure is normal or decreased; Type II respiratory failure (hypercapnic respiratory failure): arterial oxygen partial pressure is lower than 60mmHg and arterial carbon dioxide partial pressure is higher than 50mmHg.

[0066] For acute respiratory distress syndrome (ARDS): the onset is relatively rapid; the oxygenation index is <200 mmHg; the arterial oxygen partial pressure is <200 mmHg, and oxygen inhalation cannot improve it; the chest X-ray shows increased lung markings and blurred edges; the pulmonary artery pressure (PAWP) is ≤18 mmHg, and no clinical large vessel evidence of left heart failure is found.

[0067] For pulmonary edema: the main clinical manifestations are dyspnea, cyanosis, cough, coughing up white or bloody frothy sputum, scattered moist rales in both lungs, and imaging showing butterfly-shaped or flake-like blurred shadows centered on the hilum of the lung.

[0068] It can be understood that the embodiments herein provide specific content for the types of pulmonary complications after cardiac and great vascular surgery that can be predicted by the present application.

[0069] On this basis, the present application can also adaptively configure the model architecture of the pulmonary complications prediction model after cardiac and vascular surgery.

[0070] Specifically, as another exemplary embodiment, the prediction model for pulmonary complications after cardiac and great vascular surgery may be a fusion model composed of different models, and different models correspond to different pulmonary complications after cardiac and great vascular surgery.

[0071] That is to say, the prediction model for pulmonary complications after cardiac and great vessel surgery includes the pulmonary infection prediction model corresponding to the complication of pulmonary infection, the pulmonary atelectasis prediction model corresponding to the complication of atelectasis, the pneumothorax prediction model corresponding to the complication of pneumothorax, the respiratory failure prediction model corresponding to the complication of respiratory failure, the acute respiratory distress syndrome prediction model corresponding to the complication of acute respiratory distress syndrome, and the pulmonary edema prediction model corresponding to the complication of pulmonary edema.

[0072] As for the prediction models corresponding to different complications, it can be understood that the specific model architectures adopted may also be different, so as to have higher specificity and adaptability to the respective complications, and thus obtain better prediction accuracy and prediction efficiency. In this case, the specific model architectures adopted by the prediction models for different complications can also be selected based on the prediction effect after training when different model architectures are adopted.

[0073] In addition, for the overall prediction model for pulmonary complications after cardiac and great vascular surgery, only one data input is required. In the model, the input patient data can be input separately into the different models included, that is, the inputs of the different models included are the same. Alternatively, secondary data processing can be performed according to the data input requirements of different models, so that the specific patient data input into the different models included are different. Both situations are possible and can meet the data input requirements of the different models included for the prediction and processing of pulmonary complications after cardiac and great vascular surgery.

[0074] As for the latter, it is easy to understand that a more streamlined data input effect can be achieved in the secondary data processing stage, so as to enable the different models included to achieve more efficient and accurate postoperative complication prediction effects and avoid the interference that may be caused by additional and unnecessary input data.

[0075] Below, the different models in the prediction model for pulmonary complications after cardiac and great vascular surgery of the present application and the parameters / data that are focused on in the prediction process of different types of postoperative pulmonary complications are further described in detail.

[0076] Specifically, as another exemplary embodiment, corresponding to lung infection, the model input parameters of the lung infection prediction model in the lung complication prediction model after cardiac and great vessel surgery may specifically include:

[0077] The number of days of hospitalization in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, intraoperative antibiotic use, operation duration, intraoperative additional antibiotics, nasogastric tube, intraoperative blood transfusion, extracorporeal circulation duration, sedation, secondary endotracheal intubation, secondary thoracotomy, tracheotomy, nosocomial infection, intraaortic balloon pump, dialysis, blood creatinine on the day after surgery, blood creatinine 1 day after surgery, blood glucose on the day after surgery, arterial oxygen partial pressure 1 day after surgery, blood lactate on the day after surgery, and blood oxygen saturation on the day after surgery;

[0078] Corresponding to atelectasis, the model input parameters of the atelectasis prediction model in the pulmonary complications prediction model after cardiac and great vascular surgery may specifically include:

[0079] The number of days of hospitalization in the intensive care unit, the use of additional antibiotics during surgery, the use of nasogastric tubes, the amount of blood transfusion during surgery, the duration of extracorporeal circulation, nosocomial infection, white blood cells on the day after surgery, neutrophils on the day after surgery, albumin on the day after surgery, hemoglobin 1 day after surgery, albumin 1 day after surgery, serum albumin 1 day after surgery, and total hemoglobin 1 day after surgery;

[0080] Corresponding to pneumothorax, the model input parameters of the pneumothorax prediction model in the pulmonary complication prediction model after cardiac and great vascular surgery may specifically include:

[0081] Gender, alcohol consumption, length of stay in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, nasogastric tube, sedation, secondary thoracotomy, dialysis, serum creatinine on the day after surgery, hemoglobin on the 1st day after surgery, total hemoglobin on the 1st day after surgery, and blood lactate on the day after surgery;

[0082] Corresponding to respiratory failure, the model input parameters of the respiratory failure prediction model in the pulmonary complications prediction model after cardiac and large vessel surgery may specifically include:

[0083] The number of days of hospitalization in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, duration of operation, nasogastric tube, intraoperative blood transfusion, duration of extracorporeal circulation, sedation, secondary endotracheal intubation, tracheotomy, intraaortic balloon pump, dialysis, white blood cell count on the day after surgery, hemoglobin on the day after surgery, albumin on the day after surgery, white blood cell count 1 day after surgery, neutrophil count 1 day after surgery, hemoglobin 1 day after surgery, and total hemoglobin 1 day after surgery.

[0084] It can be understood that the above parameter indicators can make a significant contribution to the prediction of specific pulmonary complications after cardiac and great vascular surgery. Therefore, based on these parameter indicators, the model can achieve a higher accuracy prediction effect of pulmonary complications after cardiac and great vascular surgery.

[0085] Furthermore, the present application may further simplify the above parameter indicators to take into account both prediction accuracy and prediction efficiency.

[0086] In this regard, after secondary simplification, the model input parameters of the lung infection prediction model can specifically include:

[0087] intensive care unit length of stay, intraoperative antibiotics, sedation, and tracheotomy;

[0088] After secondary simplification, the model input parameters of the atelectasis prediction model can specifically include:

[0089] Nasogastric tube, duration of extracorporeal circulation (≤2h), white blood cell count (>10*10 9 / L), intraoperative additional antibiotics, albumin (<35 g / L) on the first day after surgery, and total hemoglobin (9.0-12.0 g / dL (male) / 9.0-11.0 g / dL (female)) on the first day after surgery;

[0090] After secondary simplification, the model input parameters of the pneumothorax prediction model can specifically include:

[0091] Nasogastric tube, sedation, secondary thoracotomy, dialysis, and serum creatinine on the day after surgery (<53 μmoI / L (male) / <44 μmoI / L (female));

[0092] After secondary simplification, the model input parameters of the respiratory failure prediction model can specifically include:

[0093] Nasogastric tube, secondary endotracheal intubation, albumin on the day after surgery (<35 g / L) and white blood cell count 1 day after surgery (≥10*109 / L).

[0094] It can be understood that the above parameter setting schemes involved in the prediction and treatment of four types of pulmonary complications after cardiac and great vascular surgery (including the parameter setting schemes before secondary simplification) are screened out by the inventors of the present application through continuous conception from complex parameter indicators. They involve both common indicators in the diagnosis / treatment of pulmonary complications after cardiac and great vascular surgery, and indicators that have been discovered and habitually ignored by the inventors of the present application. In addition, parameter indicators with repeated contributions are eliminated, and parameter indicators with high specificity are retained. From the application level, this has better practical significance and application value.

[0095] At the same time, for the prediction model of pulmonary complications after cardiac and great vascular surgery built by the present application based on the machine learning model, or the initial machine learning model involved in training the prediction of pulmonary complications after cardiac and great vascular surgery, as another practical embodiment, a logistic regression model can be specifically used.

[0096] Among them, it should be noted that the Logistic regression model mentioned here refers to a processing model that follows the Logistic regression processing logic. Its specific model architecture is the same as that introduced earlier. It can adopt the existing model architecture, or it can be a model architecture obtained by further improving and optimizing the existing model architecture, or it can be a self-developed, novel model architecture. These are all possible in actual situations.

[0097] Thus, the prediction model / initial machine learning model for pulmonary complications after cardiac and vascular surgery provided in this application provides a more specific implementation solution.

[0098] In addition, corresponding to the application of the model, the present application may also involve the training phase of the model. In this regard, as another exemplary embodiment, the method of the present application may also include:

[0099] Obtaining initial sample patient data;

[0100] After labeling and preprocessing the initial sample patient data, sample patient data is obtained, wherein the preprocessing includes removing null values, normalizing, missing value processing, and outlier processing;

[0101] The sample patient data were used as training samples to train the initial machine learning model and obtain a postoperative pulmonary complications prediction model.

[0102] Among them, the initial sample patient data is usually corresponding to the model input data involved after the model is put into actual use in terms of data format and / or source.

[0103] When annotating the initial sample patient data or the data after preprocessing the initial sample patient data, it can be done by manual annotation or by using relevant automatic annotation tools. Among them, the use of automatic annotation tools requires special configuration of automatic annotation logic that matches the data involved in this application.

[0104] In addition, it can be seen that in the embodiments herein, the present application specifically lists a series of data preprocessing operations that can be used, namely, null value removal processing, normalization processing, missing value processing and outlier processing,

[0105] For the removal of null values, the specific operation is to remove useless / blank data; for the normalization, the specific operation is to organize the data into a standard style through normalization and delete duplicates; for the missing value processing, the specific operation is to delete the data with a default rate greater than 50%, and then fill the remaining data with the median and mean; for the outlier processing, the specific operation is to remove the values ​​with an abnormal range greater than 0.5% from the data after missing value processing.

[0106] In this way, data preprocessing can be performed through labeling to obtain high-quality training samples that can be used specifically for model training.

[0107] In addition, it should be understood that, in practical applications, the specific data preprocessing operations mentioned above may be replaced by other types of specific data preprocessing operations, or specific data preprocessing operations may be added / deleted.

[0108] As for specific types of data preprocessing operations, it is understandable that existing means can usually be used directly and conveniently. Of course, the possibility of further optimizing and improving existing means or even adopting new self-developed means under actual circumstances is not ruled out.

[0109] In addition, it can be understood that in specific operations, the configuration work of the model (including the training phase) may continue to involve a series of settings in terms of details. The contents involved here usually belong to the scope of the prior art, so this application will not go into details here.

[0110] The above is an introduction to the method for predicting pulmonary complications after cardiac and great vessel surgery provided in this application. In order to facilitate better implementation of the method for predicting pulmonary complications after cardiac and great vessel surgery provided in this application, this application also provides a device for predicting pulmonary complications after cardiac and great vessel surgery from the perspective of functional modules.

[0111] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of the device for predicting pulmonary complications after cardiac and great vessel surgery in this application. In this application, the device for predicting pulmonary complications after cardiac and great vessel surgery 200 may specifically include the following structure:

[0112] The acquisition unit 201 is used to acquire patient data recorded on the system of a target patient, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data;

[0113] The prediction unit 202 is used to input the patient data into a pre-configured prediction model for pulmonary complications after cardiac and great vessel surgery, so that the prediction model for pulmonary complications after cardiac and great vessel surgery carries out corresponding prediction processing for pulmonary complications after cardiac and great vessel surgery, wherein the prediction model for pulmonary complications after cardiac and great vessel surgery is obtained by training an initial machine learning model with sample patient data labeled with corresponding pulmonary complications after cardiac and great vessel surgery results;

[0114] The extraction unit 203 is used to extract the prediction result of pulmonary complications after cardiac and great vessel surgery output by the prediction model of pulmonary complications after cardiac and great vessel surgery.

[0115] In an exemplary embodiment, the pulmonary complications after cardiac and great vessel surgery that can be predicted by the pulmonary complications prediction model after cardiac and great vessel surgery specifically include pulmonary infection, atelectasis, pneumothorax, respiratory failure, acute respiratory distress syndrome and pulmonary edema.

[0116] In another exemplary embodiment, the prediction model for pulmonary complications after cardiac and great vascular surgery is specifically a fusion model composed of different models, and different models correspond to different pulmonary complications after cardiac and great vascular surgery.

[0117] In another exemplary embodiment, corresponding to lung infection, the model input parameters of the lung infection prediction model in the lung complication prediction model after cardiac and great vessel surgery include:

[0118] The number of days of hospitalization in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, intraoperative antibiotic use, operation duration, intraoperative additional antibiotics, nasogastric tube, intraoperative blood transfusion, extracorporeal circulation duration, sedation, secondary endotracheal intubation, secondary thoracotomy, tracheotomy, nosocomial infection, intraaortic balloon pump, dialysis, blood creatinine on the day after surgery, blood creatinine 1 day after surgery, blood glucose on the day after surgery, arterial oxygen partial pressure 1 day after surgery, blood lactate on the day after surgery, and blood oxygen saturation on the day after surgery;

[0119] Corresponding to atelectasis, the model input parameters of the atelectasis prediction model in the pulmonary complications prediction model after cardiac and great vascular surgery include:

[0120] The number of days of hospitalization in the intensive care unit, the use of additional antibiotics during surgery, the use of nasogastric tubes, the amount of blood transfusion during surgery, the duration of extracorporeal circulation, nosocomial infection, white blood cells on the day after surgery, neutrophils on the day after surgery, albumin on the day after surgery, hemoglobin 1 day after surgery, albumin 1 day after surgery, serum albumin 1 day after surgery, and total hemoglobin 1 day after surgery;

[0121] Corresponding to pneumothorax, the model input parameters of the pneumothorax prediction model in the pulmonary complications prediction model after cardiac and great vascular surgery include:

[0122] Gender, alcohol consumption, length of stay in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, nasogastric tube, sedation, secondary thoracotomy, dialysis, serum creatinine on the day after surgery, hemoglobin on the 1st day after surgery, total hemoglobin on the 1st day after surgery, and blood lactate on the day after surgery;

[0123] Corresponding to respiratory failure, the model input parameters of the respiratory failure prediction model in the pulmonary complications prediction model after cardiac and vascular surgery include:

[0124] The number of days of hospitalization in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, duration of operation, nasogastric tube, intraoperative blood transfusion, duration of extracorporeal circulation, sedation, secondary endotracheal intubation, tracheotomy, intraaortic balloon pump, dialysis, white blood cell count on the day after surgery, hemoglobin on the day after surgery, albumin on the day after surgery, white blood cell count 1 day after surgery, neutrophil count 1 day after surgery, hemoglobin 1 day after surgery, and total hemoglobin 1 day after surgery.

[0125] In another exemplary embodiment, after secondary simplification, the model input parameters of the lung infection prediction model specifically include:

[0126] intensive care unit length of stay, intraoperative antibiotics, sedation, and tracheotomy;

[0127] After secondary simplification, the model input parameters of the atelectasis prediction model include:

[0128] nasogastric tube, duration of extracorporeal circulation, white blood cells on the day after surgery, additional antibiotics during surgery, albumin on the day after surgery, and total hemoglobin on the day after surgery;

[0129] After secondary simplification, the model input parameters of the pneumothorax prediction model specifically include:

[0130] nasogastric tube, sedation, secondary thoracotomy, dialysis, and serum creatinine on the day after surgery;

[0131] After secondary simplification, the model input parameters of the respiratory failure prediction model include:

[0132] Nasogastric tube, secondary endotracheal intubation, albumin on the day after surgery, and white blood cell count on the 1st day after surgery.

[0133] In another exemplary embodiment, the initial machine learning model specifically adopts a Logistic regression model.

[0134] In yet another exemplary embodiment, the apparatus further includes a training unit 204, configured to:

[0135] Obtaining initial sample patient data;

[0136] After the initial sample patient data is labeled and preprocessed, the sample patient data is obtained, wherein the preprocessing includes null value removal processing, normalization processing, missing value processing and outlier processing;

[0137] The sample patient data were used as training samples to train the initial machine learning model and obtain a postoperative pulmonary complications prediction model.

[0138] This application also provides a processing device from the perspective of hardware structure, see Figure 3 , Figure 3 301, a memory 302, and an input / output device 303. The processor 301 is used to execute the computer program stored in the memory 302 to implement the following Figure 1 The steps of the method for predicting pulmonary complications after cardiac and vascular surgery in the corresponding embodiment; or, the processor 301 is used to execute the computer program stored in the memory 302 to implement the following Figure 2 Corresponding to the functions of each unit in the embodiment, the memory 302 is used to store the processor 301 executing the above Figure 1 The computer program required by the method for predicting pulmonary complications after cardiac and vascular surgery in the corresponding embodiment.

[0139] Exemplarily, the computer program may be divided into one or more modules / units, one or more modules / units are stored in the memory 302, and executed by the processor 301 to complete the present application. One or more modules / units may be a series of computer program instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the computer device.

[0140] The processing device may include, but is not limited to, a processor 301, a memory 302, and an input / output device 303. Those skilled in the art will appreciate that the illustration is merely an example of a processing device and does not constitute a limitation on the processing device, and may include more or fewer components than shown in the illustration, or a combination of certain components, or different components. For example, the processing device may also include a network access device, a bus, etc., and the processor 301, the memory 302, the input / output device 303, etc. are connected via a bus.

[0141] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the processing device, and uses various interfaces and lines to connect various parts of the entire device.

[0142] The memory 302 can be used to store computer programs and / or modules. The processor 301 implements various functions of the computer device by running or executing the computer programs and / or modules stored in the memory 302 and calling the data stored in the memory 302. The memory 302 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, at least one application required for a function, etc.; the data storage area can store data created according to the use of the processing device, etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0143] When the processor 301 is used to execute the computer program stored in the memory 302, the following functions can be implemented:

[0144] Acquire patient data recorded on the system of a target patient, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data;

[0145] Inputting the patient data into a pre-configured cardiac and great vascular surgery pulmonary complication prediction model, so that the cardiac and great vascular surgery pulmonary complication prediction model carries out corresponding cardiac and great vascular surgery pulmonary complication prediction processing, wherein the cardiac and great vascular surgery pulmonary complication prediction model is obtained by training an initial machine learning model with sample patient data labeled with corresponding cardiac and great vascular surgery pulmonary complication occurrence results;

[0146] Extract the prediction results of pulmonary complications after cardiac and great vascular surgery output by the prediction model for pulmonary complications after cardiac and great vascular surgery.

[0147] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described cardiac and vascular surgery pulmonary complications prediction device, processing device and its corresponding units can refer to the following. Figure 1 The description of the method for predicting pulmonary complications after cardiac and great vascular surgery in the corresponding embodiment will not be repeated here in detail.

[0148] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0149] To this end, the present application provides a computer-readable storage medium, in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the present application as follows: Figure 1 The steps of the method for predicting pulmonary complications after cardiac and vascular surgery in the corresponding embodiment, the specific operations can be referred to as follows Figure 1 The description of the method for predicting pulmonary complications after cardiac and great vascular surgery in the corresponding embodiment will not be repeated here.

[0150] The computer-readable storage medium may include: a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0151] Due to the instructions stored in the computer-readable storage medium, the present application can be executed. Figure 1 The steps of the method for predicting pulmonary complications after cardiac and vascular surgery in the corresponding embodiment, therefore, the present application can be implemented as follows: Figure 1 The beneficial effects that can be achieved by the method for predicting pulmonary complications after cardiac and great vascular surgery in the corresponding embodiment are detailed in the previous description and will not be repeated here.

[0152] The above is a detailed introduction to the method, device, processing equipment and computer-readable storage medium for predicting pulmonary complications after cardiac and great vascular surgery provided by the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for predicting pulmonary complications after cardiac and vascular surgery, characterized in that: The method comprises: Acquire patient data recorded on the system of a target patient, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data; Inputting the patient data into a preconfigured cardiac and great vessel surgery pulmonary complication prediction model, so that the cardiac and great vessel surgery pulmonary complication prediction model carries out corresponding cardiac and great vessel surgery pulmonary complication prediction processing, wherein the cardiac and great vessel surgery pulmonary complication prediction model is obtained by training an initial machine learning model with sample patient data labeled with corresponding cardiac and great vessel surgery pulmonary complication occurrence results; The prediction results of pulmonary complications after cardiac and great vessel surgery output by the prediction model for pulmonary complications after cardiac and great vessel surgery are extracted.

2. The method according to claim 1, characterized in that The pulmonary complications after cardiac and great vascular surgery that can be predicted by the pulmonary complications after cardiac and great vascular surgery prediction model specifically include pulmonary infection, atelectasis, pneumothorax, respiratory failure, acute respiratory distress syndrome and pulmonary edema.

3. The method according to claim 2, characterized in that The model for predicting pulmonary complications after cardiac and great vascular surgery is specifically a fusion model composed of different models, and different models correspond to different pulmonary complications after cardiac and great vascular surgery.

4. The method according to claim 3, characterized in that Corresponding to the pulmonary infection, the model input parameters of the pulmonary infection prediction model in the pulmonary complications prediction model after cardiac and great vessel surgery include: The number of days of hospitalization in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, intraoperative antibiotic use, operation duration, intraoperative additional antibiotics, nasogastric tube, intraoperative blood transfusion, extracorporeal circulation duration, sedation, secondary endotracheal intubation, secondary thoracotomy, tracheotomy, nosocomial infection, intraaortic balloon pump, dialysis, blood creatinine on the day after surgery, blood creatinine 1 day after surgery, blood glucose on the day after surgery, arterial oxygen partial pressure 1 day after surgery, blood lactate on the day after surgery, and blood oxygen saturation on the day after surgery; Corresponding to the atelectasis, the model input parameters of the atelectasis prediction model in the pulmonary complications prediction model after cardiac and great vessel surgery include: The number of days of hospitalization in the intensive care unit, the additional antibiotics during the operation, the nasogastric tube, the amount of blood transfusion during the operation, the duration of extracorporeal circulation, the nosocomial infection, the white blood cells on the day after surgery, the neutrophils on the day after surgery, the albumin on the day after surgery, the hemoglobin 1 day after surgery, the albumin 1 day after surgery, the serum albumin 1 day after surgery, and the total hemoglobin 1 day after surgery; Corresponding to the pneumothorax, the model input parameters of the pneumothorax prediction model in the postoperative pulmonary complications prediction model for cardiac and great vascular surgery include: Gender, alcohol consumption, length of stay in the intensive care unit, emergency surgery, preoperative white blood cell count, preoperative neutrophil count, nasogastric tube, sedation, secondary thoracotomy, dialysis, blood creatinine on the day after surgery, hemoglobin on the 1st day after surgery, total hemoglobin on the 1st day after surgery, and blood lactate on the day after surgery; Corresponding to the respiratory failure, the model input parameters of the respiratory failure prediction model in the postoperative pulmonary complications prediction model for cardiac and great vascular surgery include: The number of days of hospitalization in the intensive care unit, the emergency surgery, the preoperative white blood cell count, the preoperative neutrophil count, the duration of the operation, the nasogastric tube, the amount of blood transfusion during the operation, the duration of extracorporeal circulation, the sedation, the secondary endotracheal intubation, the tracheotomy, the aortic balloon pump, dialysis, the white blood cell count on the day after surgery, the hemoglobin on the day after surgery, the albumin on the day after surgery, the white blood cell count 1 day after surgery, the neutrophil count 1 day after surgery, the hemoglobin 1 day after surgery and the total hemoglobin 1 day after surgery.

5. The method according to claim 4, characterized in that After secondary simplification, the model input parameters of the lung infection prediction model specifically include: The number of days of intensive care unit hospitalization, the intraoperative antibiotic supplementation, the sedation and tracheotomy; After secondary simplification, the model input parameters of the atelectasis prediction model specifically include: The nasogastric tube, the duration of extracorporeal circulation, the white blood cells on the day after surgery, the additional antibiotics during surgery, the albumin on the day after surgery, and the total hemoglobin on the day after surgery; After secondary simplification, the model input parameters of the pneumothorax prediction model specifically include: The nasogastric tube, the sedation, the secondary thoracotomy, the dialysis, and the blood creatinine on the day after surgery; After secondary simplification, the model input parameters of the respiratory failure prediction model specifically include: The nasogastric tube, the secondary endotracheal intubation, the albumin on the day after surgery and the white blood cell count on the 1st day after surgery.

6. The method according to claim 1, characterized in that The initial machine learning model specifically adopts a Logistic regression model.

7. The method according to claim 1, characterized in that The method further comprises: Obtaining initial sample patient data; After labeling and preprocessing the initial sample patient data, the sample patient data is obtained, wherein the preprocessing includes null value removal processing, normalization processing, missing value processing and outlier processing; The sample patient data is used as a training sample to train the initial machine learning model to obtain the postoperative pulmonary complications prediction model.

8. A device for predicting pulmonary complications after cardiac and vascular surgery, characterized in that: The device comprises: An acquisition unit, used for acquiring patient data of a target patient recorded on the system, wherein the target patient is specifically a patient who has undergone cardiac and vascular surgery, and the patient data includes general information data, preoperative factor data, intraoperative factor data, and postoperative factor data; A prediction unit, used for inputting the patient data into a pre-configured prediction model for pulmonary complications after cardiac and great vessel surgery, so that the prediction model for pulmonary complications after cardiac and great vessel surgery carries out corresponding prediction processing for pulmonary complications after cardiac and great vessel surgery, wherein the prediction model for pulmonary complications after cardiac and great vessel surgery is obtained by training an initial machine learning model with sample patient data annotated with corresponding occurrence results of pulmonary complications after cardiac and great vessel surgery; The extraction unit is used to extract the prediction results of pulmonary complications after cardiac and great vessel surgery output by the prediction model for pulmonary complications after cardiac and great vessel surgery.

9. A processing device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a computer program, and the processor executes the method according to any one of claims 1 to 7 when calling the computer program in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute the method according to any one of claims 1 to 7.

Images