Critical patient pendelluft automatic detection method based on electrical impedance tomography and artificial intelligence

Abstract

The present application belongs to the field of intelligent medical treatment, and particularly relates to a critical patient Pendelluft automatic detection method based on electrical impedance tomography and artificial intelligence. The method comprises: acquiring EIT data of a lung region of a patient within a monitoring period; segmenting the lung region into multiple regions, and extracting EIT data of the multiple regions to obtain EIT signals of the multiple regions; obtaining a tidal resistance characteristic based on the EIT signals, the tidal resistance characteristic comprising: a regional tidal resistance amplitude difference and an adjacent regional tidal resistance amplitude difference; calculating an EIT ventilation characteristic based on the tidal resistance characteristic, the EIT ventilation characteristic being a swing breath; and judging whether the patient has Pendelluft based on the EIT ventilation characteristic. Based on real-time acquired EIT time series data, through processing of the time series data, an index reflecting whether Pendelluft exists and its severity is extracted, and based on real-time monitoring of Pendelluft, the clinical management of a clinician on a mechanically ventilated patient is assisted.

Description

Technical Field

[0001] This invention relates to the field of intelligent healthcare, and more specifically, to an automated method, device, medium, and program product for the Pendeluft detection of critically ill patients based on electrical impedance imaging and artificial intelligence. Background Technology

[0002] Pendelluft is defined as asynchronous alveolar ventilation caused by variations in time constants or dynamic pleural pressure across different lung regions. In patients with obstructive lung diseases (such as chronic obstructive pulmonary disease (COPD), significant differences in the time constants of the respiratory system across different lung regions lead to the pendulum phenomenon. In spontaneously breathing patients with lung injury (such as those with acute respiratory distress syndrome (ARDS), the uneven transmission of pleural pressure generated by diaphragmatic contractions is the primary cause of the pendulum phenomenon. In extreme cases, such as flail chest, the pendulum volume can reach up to 12.5% ​​of the total airway volume.

[0003] Because pendelluft can be potentially harmful by introducing localized overexpansion, tidal lung recruitment, and inflammation, it is necessary to monitor pendelluft in order to adjust treatment or ventilation strategies accordingly. In the early stages of surgery with limited technical options, pendelluft was observed in open hemithoracic surgery. In experimental settings, only ventilation asynchrony between the left and right lungs could be detected. The extent of pendelluft can also be assessed by calculating the associated lung mechanics. Pathological changes in the lungs can be observed using computed tomography and magnetic resonance imaging, which provide indirect information about the regional time constant. In previous studies, positron emission tomography (PET) was used to capture the gaps in the tracer nitrogen-13 to calculate the pendulum. In a recent study, a microphone array was used to detect the pendulum. In experimental settings, dark-field microscopy and multispectral oximetry were also used to assess the effect of sighing on the pendulum.

[0004] However, the methods described above cannot be used at the bedside for identifying and continuously measuring pendulums. Electrical impedance tomography (EIT) is a novel imaging technique, and in recent years, an increasing number of studies have utilized EIT to detect pendulums. Currently, EIT can be used for real-time monitoring of local lung ventilation, but its clinical real-time monitoring relies heavily on expert interpretation, and clinical implementation is hampered by the complexity of data interpretation.

[0005] Mechanical ventilation remains a cornerstone intervention for patients with acute respiratory failure. However, prolonged mechanical ventilation can lead to various complications and adverse outcomes. The timing and strategy of weaning from mechanical ventilation are key factors affecting patient prognosis, with approximately 20-30% of patients experiencing weaning difficulties, resulting in prolonged mechanical ventilation and increased mortality. This challenge is particularly pronounced in patients with acute respiratory distress syndrome (ARDS), where, despite advancements in mechanical ventilation strategy 3, mortality rates remain high (35-46%).

[0006] Among various respiratory parameters, the pendelluft phenomenon (the movement of air between different lung regions without changing total lung volume) has become a key indicator of ventilation heterogeneity. The detection and quantification of pendelluft can provide valuable insights into regional lung mechanics and guide ventilation strategies. Summary of the Invention

[0007] In view of the above problems, the present invention provides an automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence. By extracting regional ventilation parameters from real-time monitored EIT data, the occurrence and severity of Pendeluft can be monitored in a timely manner.

[0008] This application (first aspect) discloses an automated method for detecting Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence, the method comprising:

[0009] S101: Acquire EIT data of the patient's lung region over a period of time;

[0010] S102: The lung region is divided into multiple regions, and the EIT data of multiple regions are extracted to obtain the EIT signal of each region;

[0011] S103: Obtain moisture resistance characteristics based on the EIT signal, the moisture resistance characteristics including: moisture resistance amplitude difference within the region and moisture resistance amplitude difference between adjacent regions;

[0012] S104: Based on the tidal resistance characteristics, the EIT ventilation characteristics are calculated. The EIT ventilation characteristics are swing breathing, and the swing breathing is the ratio of the amplitude difference of tidal resistance between the maximum regions to the amplitude difference of tidal resistance between the average regions.

[0013] S105: Determine whether the patient has Pendleuft based on the EIT ventilation characteristics.

[0014] Furthermore, the swinging breathing is represented as:

[0015]

[0016] Where PI represents swing breathing, ΔA i,j The mean(ΔA) represents the difference in moisture resistance amplitude between adjacent regions i and j. max() represents finding the maximum value. i,j This represents the average value of the difference in moisture resistance amplitude between all regions;

[0017] Optionally, if a patient develops Pendeluft, the EIT ventilation characteristics are input into a grade discriminant classifier to determine the severity of the patient's Pendeluft.

[0018] Furthermore, the method further includes the following after S102:

[0019] S102-1: Perform frequency domain transformation on the EIT signals of each region to obtain the spectrum signals of each region;

[0020] S103 is replaced with S103': Spectral features are obtained based on the spectral signal, and the spectral features include: the phase difference between adjacent regions;

[0021] S104 is replaced by S104': EIT ventilation characteristics are calculated based on the spectral characteristics, where the EIT ventilation characteristics are ventilation cycle differences, and the ventilation cycle differences are the average phase differences between adjacent regions.

[0022] Furthermore, the ventilation cycle difference is expressed as follows:

[0023]

[0024] TDI represents ventilation cycle difference. This represents the phase difference between adjacent regions i and j, and N represents the total number of phase differences.

[0025] Optionally, the lung region can be divided into four areas: front, back, left, and right.

[0026] Optionally, the ventilation cycle difference is expressed as:

[0027]

[0028] TDI represents the average ventilation cycle difference among four adjacent regions. This represents the phase difference between four adjacent regions.

[0029] Optionally, the phase difference between adjacent regions is expressed as:

[0030]

[0031] Among them, X i(f) This represents the frequency domain signal of the i-th region; This indicates taking the phase value corresponding to the dominant frequency of the spectrum;

[0032] Optionally, if a patient develops Pendeluft, the EIT ventilation characteristics are input into a grade discriminant classifier to determine the severity of the patient's Pendeluft.

[0033] Furthermore, the method includes:

[0034] After obtaining the EIT signals of each region, frequency domain transformation is performed on the EIT signals of each region to obtain the spectrum signals of each region;

[0035] Based on the EIT signal, the tidal resistance characteristics are obtained, which include: the difference in ventilation between adjacent regions and the time delay between adjacent regions;

[0036] Based on the spectral signal, spectral features are obtained, including: the phase difference between adjacent regions;

[0037] Based on the spectral characteristics and the tidal resistance characteristics, the EIT ventilation characteristics are calculated. The EIT ventilation characteristics are ventilation synchronization heterogeneity indexes, which are obtained based on the standard deviation of the phase difference between adjacent regions, the standard deviation of the ventilation volume difference between adjacent regions, and the mean of the time delay between adjacent regions.

[0038] The ventilation synchronization heterogeneity index is used to determine whether the patient has Pendelluft;

[0039] Optionally, the ventilation synchronization heterogeneity index is the ratio of the product of the standard deviation of the phase difference between adjacent regions and the standard deviation of the ventilation volume difference between adjacent regions to the mean of the time delay between adjacent regions.

[0040] Optionally, the ventilation synchronization heterogeneity index is calculated as follows:

[0041]

[0042] VSHI stands for Ventilation Synchronization Heterogeneity Index. σ represents the standard deviation of the phase difference between adjacent regions. V The standard deviation of ventilation difference between adjacent zones, μ t This represents the average time delay between adjacent regions;

[0043] Optionally, the standard deviation of the phase difference can be calculated as follows:

[0044] σ φ =std(Δφ) 1,2 ,Δφ 1.3 ,Δφ 2,4 ,Δφ 3,4)

[0045] Where, σ φ Δφ represents the standard deviation of the phase difference between adjacent regions. 1,2 ,Δφ 1.3 ,Δφ 2,4 ,Δφ 3,4 These represent the phase differences between adjacent regions 1 and 2, 1 and 3, 2 and 4, and 3 and 4, respectively.

[0046] Optionally, the standard deviation of ventilation variation can be calculated as follows:

[0047] σ V =std(ΔV) 1,2 ,ΔV 1,3 ,ΔV 2,4 ,ΔV 3,4 )

[0048] Where, σ V ΔV represents the standard deviation of the ventilation difference between adjacent zones. 1,2 ,ΔV 1,3 ,ΔV 2,4 ,ΔV 3,4 These represent the differences in ventilation volume between adjacent regions 1 and 2, 1 and 3, 2 and 4, and 3 and 4, respectively.

[0049] Optionally, the mean of the time delay can be calculated as follows:

[0050]

[0051] μ t Δt represents the mean time delay between adjacent regions. 1,2 Δt 1,3 Δt 2,4 Δt 3,4 These represent the time delays between adjacent regions 1 and 2, 1 and 3, 2 and 4, and 3 and 4, respectively.

[0052] Optionally, the regional time delay is represented as:

[0053]

[0054] Where, Δt i,j Indicates regional time delay. This represents the time-domain ventilation signal after preprocessing in the i-th region, and argmax() represents the weighted average of the peak times of each respiratory cycle within the entire monitoring period.

[0055] Optionally, if a patient develops Pendeluft, the EIT ventilation characteristics are input into a grade discriminant classifier to determine the severity of the patient's Pendeluft.

[0056] Furthermore, the method includes:

[0057] The lung region is divided into anterior and posterior areas;

[0058] Extract the EIT signals from the front and back regions, and obtain the spectral signals of the front and back regions after frequency domain transformation;

[0059] Based on the EIT signal, the tidal resistance characteristics are obtained, and the tidal resistance characteristics include: ventilation rate;

[0060] Based on the spectral signal, spectral features are obtained, including: the phase difference between adjacent regions;

[0061] The EIT ventilation characteristics are calculated based on the aforementioned spectral characteristics and tidal resistance characteristics. These EIT ventilation characteristics are optimized phase variation indices.

[0062] Optionally, the optimized phase variation index is expressed as:

[0063]

[0064] Where PVI represents the optimized phase variation index, VT represents the ventilation rate, and f s Respiratory rate, This represents the phase difference of the k-th time window, a constant of 10. 3 Dimensional adjustment factor;

[0065] Optional,

[0066]

[0067] arg(X {ant} (f d ,k)) represents the front region at the dominant frequency f in the k-th time window. d Complex spectral values ​​at;

[0068] arg(X {post} (f d ,k)) represents the front region at the dominant frequency f in the k-th time window. d The complex spectrum value at that location.

[0069] Furthermore, the method includes:

[0070] The dynamic airflow redistribution coefficient is calculated based on the aforementioned spectral characteristics and tidal resistance characteristics.

[0071] The dynamic airflow redistribution coefficient is used to determine whether the patient has Pendelluft.

[0072] Optionally, the dynamic airflow redistribution coefficient is expressed as:

[0073]

[0074] The lung region is divided into anterior and posterior zones, VT ant (k) represents the tidal volume in the front zone of the k-th time window, VT post (k) represents the moisture content in the region after the k-th time window. This represents the phase difference between the front and back regions of the k-th time window; M represents the number of selected observation time windows.

[0075] Optionally, a time window is defined as one respiratory cycle, with M being 3.

[0076] The second aspect of this application discloses a real-time PENDELLUFT monitoring system for mechanically ventilated patients based on EIT, comprising:

[0077] Acquisition module 201: Used to acquire EIT data of the patient's lung region over a period of time;

[0078] Region segmentation module 202: used to segment the lung region into multiple regions, extract EIT data from multiple regions, and obtain the EIT signal of each region;

[0079] Data preprocessing module 203: used to obtain moisture resistance characteristics based on the EIT signal, the moisture resistance characteristics including:

[0080] Moisture resistance amplitude difference within the region, and moisture resistance amplitude difference between adjacent regions;

[0081] Feature extraction module 204: used to calculate EIT ventilation features based on the tidal resistance features, wherein the EIT ventilation features are swing breathing, and the swing breathing is the ratio of the amplitude difference of tidal resistance between the maximum region to the amplitude difference of tidal resistance in the average region.

[0082] Prediction module 205: used to determine whether the patient has Pendleuft based on the EIT ventilation characteristics.

[0083] A third aspect of this application discloses a computer device, the device comprising: a memory and a processor; the memory being used to store program instructions; the processor being used to invoke the program instructions, which, when executed, are used to perform the steps of the method described above.

[0084] The fourth aspect of this application discloses a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0085] The fifth aspect of this application discloses a computer program product, including a computer program that, when executed by a processor, implements the steps of the above-described method.

[0086] This application has the following beneficial effects:

[0087] (1) The process of this application begins with the collection of continuous EIT data from a bedside monitor. Our system then processes this standardized data and performs automated Pendleft detection and analysis in less than 3 minutes. This streamlined workflow greatly reduces the time and expertise required for EIT data interpretation while maintaining high accuracy and reliability.

[0088] (2) Based on real-time acquired EIT time series data, this application extracts ventilation cycle differences and swing breathing through processing time series data. These indicators reflect the presence and severity of Pendleuft. Based on real-time monitoring of Pendleuft, it assists clinicians in the clinical management of mechanically ventilated patients. Attached Figure Description

[0089] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0090] Figure 1 This is a schematic diagram of the method flow provided in the first aspect of the present invention;

[0091] Figure 2 This is a schematic diagram of a program product provided in the second aspect of the present invention;

[0092] Figure 3 This is a schematic diagram of a computer device provided in an embodiment of the present invention;

[0093] Figure 4 This is a schematic diagram of the architecture of an exemplary computing device provided in an embodiment of the present invention;

[0094] Figure 5 This is a schematic diagram of the storage medium provided in an embodiment of the present invention;

[0095] Figure 6 This is a schematic diagram of a process for analyzing EIT time series data, extracting features, and then making predictions, provided by an embodiment of the present invention.

[0096] Figure 7This invention provides a schematic diagram of a process for analyzing and extracting features from EIT time series data, fusing these features, and then making predictions.

[0097] Figure 8 This is a performance diagram of a prediction model based on different features provided in an embodiment of the present invention;

[0098] Figure 9 This is a schematic diagram illustrating the effectiveness of Pendeluft classification based on multiple indicators provided in an embodiment of the present invention;

[0099] Figure 10 This is a schematic diagram illustrating the performance comparison of the indicators proposed in this invention for Pendeluft prediction, provided by an embodiment of this invention.

[0100] Figure 11 This is a schematic diagram illustrating the division of lung regions according to an embodiment of the present invention;

[0101] Figure 12 This is a schematic diagram of the calculation process for moisture resistance characteristics provided in an embodiment of the present invention. Detailed Implementation

[0102] To enable those skilled in the art to better understand the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0103] In some of the processes described in the specification, claims, and accompanying drawings of this invention, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as S101, S102, etc., are merely used to distinguish different operations and do not represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a sequential order, nor do they limit "first" and "second" to different types.

[0104] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0105] Figure 1This is a schematic diagram of an automated Pendeloff detection method for critically ill patients based on electrical impedance imaging and artificial intelligence, provided by an embodiment of the present invention. Specifically, the method includes the following steps:

[0106] S101: Acquire EIT data of the patient's lung region over a period of time;

[0107] S102: The lung region is divided into multiple regions, and the EIT data of multiple regions are extracted to obtain the EIT signal of each region;

[0108] S103: Obtain moisture resistance characteristics based on the EIT signal, the moisture resistance characteristics including: moisture resistance amplitude difference within the region and moisture resistance amplitude difference between adjacent regions;

[0109] S104: Based on the tidal resistance characteristics, the EIT ventilation characteristics are calculated. The EIT ventilation characteristics are swing breathing, and the swing breathing is the ratio of the amplitude difference of tidal resistance between the maximum regions to the amplitude difference of tidal resistance between the average regions.

[0110] S105: Determine whether the patient has Pendleuft based on the EIT ventilation characteristics.

[0111] This application is based on a study: Study design and background: A retrospective, multicenter study conducted in three tertiary hospitals in China between January 2020 and December 2024.

[0112] Participants: Of the 458 patients screened, 278 were mechanically ventilated and met the inclusion criteria.

[0113] Primary outcomes included diagnostic accuracy (sensitivity, specificity, and AUC-ROC) and time efficiency. Secondary outcomes included cost-effectiveness and the correlation between pendelluft severity and clinical outcomes.

[0114] The current multicenter study aims to: (1) develop and validate an automated pendelluft detection and grading system using EIT data; (2) compare its diagnostic accuracy and cost-effectiveness with conventional machine learning methods and expert assessments; and (3) investigate the relationship between pendelluft patterns detected by ChatGPT and clinical outcomes in mechanically ventilated patients.

[0115] Methods: Study design and patient selection: A multicenter retrospective diagnostic study was conducted in three tertiary hospitals in China (Shanghai Ruijin Hospital, Peking Union Medical College Hospital, and Fujian Provincial Hospital in Fuzhou) from January 2020 to December 2024.

[0116] Guidelines for the Diagnostic Accuracy Study (STARD) and the Enhanced Epidemiological Observation Study Report (STROBE) were followed. Of the initially screened 458 patients, we included 278 adult patients (≥18 years) who received pressure support mechanical ventilation and underwent EIT monitoring. Patients were excluded if their EIT records were incomplete, signal quality was poor (signal quality index <90%), they had chest wall deformities, severe hemodynamic instability (norepinephrine >0.5 μg / kg / min), or follow-up data were incomplete. The study protocol was approved by the institutional review committees of all participating centers.

[0117] EIT data acquisition and processing uses a PulmoVista 500 device. EIT measurements were performed in Lübeck, Germany, with 16 electrode bands normalized and positioned in the fourth intercostal space according to a pre-defined protocol. Data were recorded at 50 Hz during 10-minute intervals of stable ventilation, with continuous quality monitoring. The preprocessing procedure included removal of cardiac-related impedance variations, motion artifact correction, bandpass filtering (0.1–1.0 [ZZ1] Hz), and normalization to a 32 × 32 pixel matrix. Signal quality was rigorously controlled (minimum threshold >10 dB) and electrode contact was verified (>90% good contact required), as validated in previous studies, through signal-to-noise ratio assessment.

[0118] We combine deep learning capabilities with specialized ventilation pattern analysis. The model architecture begins with an EIT image encoder that processes regional ventilation distribution, capturing both controlled and spontaneous breathing patterns. This initial processing stage quantifies the swing amplitude by analyzing the percentage of tidal volume movement between independent and dependent lung regions. The core analytics component employs a 6-layer Transformer encoder and a specialized attention mechanism to comprehensively detect asynchronous regional ventilation. A classification head complements this, generating pendelluft severity levels with relevant confidence scores. Time-series analysis is particularly important for distinguishing different ventilation cycles and identifying subtle patterns of ventilation heterogeneity. The model was systematically trained and validated using a comprehensive dataset and benchmarked against traditional methods. This ensemble approach ensures reliable pendelluft detection while maintaining real-time analytical capabilities for clinical decision support.

[0119] Pendelluft assessment and classification: Pendelluft amplitude is quantified as the impedance difference between the sum of impedance changes in all regions and the global impedance change, using a validated methodology. Patients are classified into four grades according to established criteria: Grade 0 (no pendelluft, <2.5% TIV), Grade 1 (mild, 2.5–5% TIV), Grade 2 (moderate, 5–10% TIV), and Grade 3 (severe, >10% TIV). This grading system has been validated through consensus among clinical experts and its relevance to clinical outcomes.

[0120] Statistical Analysis: Sample size was calculated based on AUC comparisons using the DeLong method. Statistical analysis was performed using R version 4.0.0 (R Foundation for Statistical Computation, Vienna, Austria). Model performance was evaluated by a comprehensive diagnostic accuracy index, calibration assessment using the Hosmer-Lemeshow test, and decision curve analysis. Statistical significance was assessed using the McNemar test with Bonferroni correction (adjusted significance threshold P < 0.01). Missing data were handled using multiple imputation techniques.

[0121] Results: Of the 458 patients screened, 278 were from three tertiary hospitals, and 245 (88.1%) had complete datasets suitable for the final analysis. Baseline characteristics stratified by pendelluft severity included a median age of 62.4 years (IQR: 51–73), with 156 (63.7%) being male. Compared with patients with low-grade pendelluft (n = 156), patients with high-grade pendelluft (n = 122) exhibited significantly worse baseline characteristics, including higher APACHE II scores (21.3 ± 5.6 vs 17.5 ± 4.2, P < 0.001), lower PaO2 / FiO2 ratios (180 ± 42 vs 220 ± 48, P < 0.001), and higher respiratory rates (28 ± 6 vs 22 ± 5 / min, P < 0.001).

[0122] Clinical outcome analysis showed a significant association between pendelluft severity and patient prognosis. The 28-day mortality rate gradually increased from 9.0% in low-grade pendelluft to 25.4% in high-grade pendelluft (adjusted HR 2.8, 95% CI: 1.6–4.9, P = 0.001). Kaplan-Meier survival analysis indicated significantly poorer survival in high-grade patients (log-rank P < 0.001). Early detection and intervention in 82 patients with grade 2–3 pendelluft improved oxygenation (mean P / F ratio improvement of 52 ± 18 mmHg, P < 0.001) and shortened ventilation time (median reduction of 2.8 days, P = 0.002).

[0123] Compared with traditional methods, the system of this application exhibits superior diagnostic performance, with an overall accuracy of 89.6% (95% CI: 86.2-92.4%), sensitivity of 88.7% (95% CI: 84.9-91.8%), and specificity of 90.2% for the Pendelluft detection.

[0124] (95% CI: 87.1–92.8%). The system's AUC-ROC was 0.91 (95% CI: 0.87–0.95), significantly outperforming traditional machine learning methods. 2 7. Models included CNN (AUC 0.85, P = 0.003), Random Forest (AUC 0.82, P < 0.001), and SVM (AUC 0.79, P < 0.001). Model calibration analysis showed a high degree of consistency between predicted and observed probabilities (Hosmer-Lemeshow test, P = 0.87). Center-specific performance analysis showed consistent model accuracy across different hospitals, with AUCs ranging from 0.89 to 0.92 (inter-center comparison P = 0.38). The implementation of this system significantly improved workflow efficiency, reducing the average analysis time from 12.5 minutes to 2.8 minutes and saving each patient RMB 850 (95% CI: RMB 620–1080).

[0125] The strong correlation between pendelluft severity and clinical outcomes exceeded previous observations. While earlier studies demonstrated an association between ventilation heterogeneity and mortality, our results revealed a specific threshold effect. The presence of pendelluft during the T-piece trial was significantly associated with increased mortality in patients who were difficult to wean off (28-day mortality: 37.8% vs 11.1%, p = 0.014). Importantly, however, we demonstrate that early detection and intervention can alter this trajectory. This finding supports the growing body of evidence for personalized mechanical ventilation strategies, as highlighted by recent multicenter trials. A key strength of our system is its robust performance across diverse clinical settings and patient populations. Previous AI implementations have often exhibited significant performance degradation when applied across centers due to variations in clinical practice and patient characteristics. Our model maintained consistent performance across three centers (AUC range 0.89–0.92) and different patient subgroups, likely due to our comprehensive training methodology, as recommended by recent AI validation guidelines, incorporating center-specific variations. Integration of this system with clinical workflows warrants particular attention. While previous studies have primarily focused on diagnostic accuracy, we demonstrate significant improvements in efficiency (analysis time reduced from 12.5 minutes to 2.8 minutes) and cost-effectiveness (a saving of RMB 850 per patient). This aligns with recent health economics analyses of AI implementation in intensive care, but importantly, our study provides detailed workflow integration metrics, including uptime (99.4%) and false alarm rate (3.2%), addressing key implementation issues raised by Wang et al.

[0126] This application presents an analytical workflow for monitoring swing breathing based on real-time collected EIT data, which mainly includes the following key stages:

[0127] Phase 1: EIT Data Acquisition

[0128] A continuous data sequence is acquired through an EIT imaging device, denoted as I(1), I(2)…I(t), where I(y) represents the EIT data acquired at time t. During the acquisition of EIT image data, the temporal continuity and quality of the image sequence are guaranteed.

[0129] In some embodiments, the starting time of the time window is marked as 1, and the t-th time is marked as t;

[0130] Phase Two: Time Series Analysis of the Collected EIT Data Figure 12 )

[0131] 2.1 First, the lung region is segmented into regions 1, 2, ..., n, ..., N;

[0132] In some embodiments, N=2, and only the lung region is divided into gravity-dependent and non-gravity-dependent regions.

[0133] In some embodiments, N=2, and the lung region is divided only into anterior and posterior regions.

[0134] In some embodiments, N=4, the lung region is divided into front, back, left, and right regions. Figure 11 (As shown).

[0135] 2.2 Extracting time-series signals from segmented regions to obtain EIT data sequences for each region.

[0136] Time signal sequence X i (1),X i (2)…X i (t) represents the EIT signal of the i-th region;

[0137] The EIT ventilation signal for the i-th region is calculated as follows:

[0138]

[0139] Where 0 to T represents one ventilation cycle.

[0140] 2.3 After processing the EIT data sequence using Fourier analysis, the frequency signals of each region were obtained: X 1(f) X 2(f) ,…,X i(f) ,…,X N(f) , where X i(f) The frequency domain signal representing the i-th region represents the frequency signal of the i-th region;

[0141] Time signals and frequency signals are two different perspectives on the same signal. Time signals show how the signal evolves over time t, while frequency signals show the frequency composition of the signal.

[0142] 2.4 Quantitative Ventilation Cycle Characteristics:

[0143] By analyzing the spectrum, we can:

[0144] ① Identify the primary ventilation rate: Obtain the patient's basic ventilation rate by finding the maximum peak value in the power spectrum.

[0145] ② Calculate signal amplitude: Measure the peak-to-trough difference of the signal, reflecting the Impedance Tidal Variation Index (ITVI):

[0146]

[0147] Where max(X) n [f vent [) represents the maximum value of the moisture resistance signal in the nth region, min(X n [f vent [) represents the minimum value of the moisture resistance signal in the nth region, max(X {global}} represents the maximum value of the moisture resistance signal across all regions;

[0148] ③ Extract phase information: Analyze the phase angle of the signal to reflect the ventilation timing characteristics:

[0149] ④ Determine the cycle duration: Calculate the duration of a complete ventilation cycle based on the dominant frequency.

[0150] ⑤VT represents tidal volume; Unit: mL

[0151] Where V(t) represents the real-time flow rate (measured by the ventilator or estimated using the derivative of EIT impedance change); the integration time range is one basic unit: a single natural respiratory cycle T. cycle From the start of the inspiratory phase to the start of the next inspiratory phase; achievement standard: three consecutive stable cycles (cycle variation coefficient <10%); the time window is represented as [t0, t0+3T]. cycle ]

[0152] The tidal volume of the i-th region is expressed as:

[0153] 2.5 Calculate the differences between different regions:

[0154] In some embodiments, the lung is divided into anterior and posterior, left and right regions, and the calculations are performed separately:

[0155] ① Differences in moisture resistance amplitude between regions: Calculate the difference between the average moisture resistance amplitudes of different regions to reflect the non-uniformity of ventilation distribution.

[0156] The difference in signal peak values ​​between adjacent regions i and j is called the moisture resistance amplitude difference. This represents the time-domain signal amplitude difference in the i-th region, i.e., the peak-to-valley value.

[0157] ② Phase difference between regions: Analyze the phase difference of ventilation waveforms in different regions to reflect ventilation synchronicity.

[0158]

[0159] Among them, X i(f) This represents the frequency domain signal of the i-th region; This indicates taking the phase value corresponding to the dominant frequency of the spectrum;

[0160] ③ Time delay: Measure the time difference between different areas reaching peak ventilation, reflecting inconsistencies in ventilation timing.

[0161] Regional time delay is expressed as:

[0162]

[0163] in, represents the preprocessed time-domain ventilation signal of the i-th region, where preprocessing includes noise reduction, signal enhancement and other preprocessing methods; argmax represents the weighted average of the peak times of each respiratory cycle within the entire monitoring period;

[0164] ④ Differences in ventilation between regions:

[0165] ΔV i,j =VT i -VT j

[0166] VT i VT represents the ventilation volume of the i-th region within the respiratory cycle. j This represents the ventilation volume of the j-th region within the respiratory cycle.

[0167] Peak ventilation location method; the determination of peak ventilation must meet the following conditions:

[0168] I) Preprocessing: Bandpass filtering is performed on the raw EIT signal (0.1-5Hz to preserve the respiratory correlation frequency band).

[0169] Ⅱ) Peak detection:

[0170] • Use a sliding window (3s length, 50% overlap) to find local maxima.

[0171] • Ensure that the peak interval conforms to the physiological respiratory cycle range (0.15-0.3Hz).

[0172] III) Main frequency calibration: When the frequency domain main frequency f max When the peak value differs from the main peak value in the time domain, the peak value in the time domain shall prevail.

[0173] Phase 3: Feature Extraction and Fusion

[0174] 3.1 Extract key indicators:

[0175] ① The ventilation cycle difference index (TDI) is obtained by calculating the cumulative phase difference between adjacent lung regions. The specific calculation formula is as follows:

[0176]

[0177] Where i and j represent adjacent regions, this indicator can effectively reflect the inconsistency of ventilation timing between different regions.

[0178] In some embodiments, the lungs are divided into four quadrants / regions, and the average phase difference between adjacent regions is calculated to obtain the ventilation cycle difference of the four regions:

[0179]

[0180] Lung region division as follows Figure 11 As shown.

[0181] In some embodiments, the specific steps for calculating the phase difference include:

[0182] First, perform an FFT transformation (1024 points, Hanning window) on the EIT signal of each ROI;

[0183] Secondly, identify the phase value (phase angle of the dominant frequency point) corresponding to the spectral peak;

[0184] Then, calculate the absolute value of the difference in the phase angle of the dominant frequency between adjacent ROIs;

[0185] Finally, the TDI is calculated by averaging all adjacent pairs.

[0186] In some embodiments, the lung region is divided into four regions: front, back, left, and right. When calculating TDI, adjacent regions include:

[0187]

[0188] Then, the TDI is obtained by averaging the phase differences.

[0189] ② The swing breathing index (PI) is obtained by calculating the difference between the maximum signal peaks in the regions and the mean of the peak and trough values ​​of the amplitude across all regions, i.e.:

[0190] The swinging breathing is represented as:

[0191]

[0192] in, This represents the peak difference in moisture resistance between adjacent regions i and j. The mean(ΔA) represents the difference in moisture resistance amplitude within the i-th region, also known as the peak-to-valley value. i,j This represents the average of the peak and trough values ​​of amplitude across all regions.

[0193] ③ Furthermore, this application develops the Ventilation Synchronization Heterogeneity Index (VSHI), whose calculation model integrates the dual dimensions of temporal asynchrony (Δφ / Δt) and amplitude heterogeneity (ΔV), and the formula is defined as:

[0194]

[0195] VSHI stands for Ventilation Synchronization Heterogeneity Index.

[0196] Phase difference standard deviation:

[0197]

[0198] Standard deviation of ventilation difference:

[0199] σ V =std(ΔV) 1,2 ,ΔV 1,3 ,ΔV 2,4 ,ΔV 3,4 )

[0200] Average latency:

[0201]

[0202] When divided according to the four quadrants (front, back, left, right), it corresponds to Figure 11 The regions (1-F, 2-B, 3-R, 4-R) are defined; first, all adjacent combinations are calculated:

[0203] -Front and Left

[0204] -Front and Right

[0205] -back and left

[0206] -back and right

[0207] In some embodiments, the calculated average phase difference is shown in the table below:

[0208] Region pairing Anatomical proximity (Yes / No) Average phase difference (rad) FL yes 0.32 FR yes 0.28 BL yes 0.41 BR yes 0.37 FB no N / A

[0209] 2. The VSHI calculation rules include all parameter differences for these four pairs of combinations.

[0210] In some embodiments, the VSHI calculation code is as follows:

[0211]

[0212]

[0213] Clinical validation showed that in a cohort of 50 ARDS patients, this indicator had an independent predictive AUC of 0.83 for the Pendleuft phenomenon (p<0.001), and the AUC increased by 6 percentage points to 0.89 when used in combination with traditional respiratory parameters.

[0214] ④ In addition, this application introduces the Dynamic Airflow Redistribution Coefficient (DRC), which is expressed as follows: This coefficient is derived from the product of the ratio of tidal volume in the front and rear zones and the phase difference.

[0215]

[0216] The lung region is divided into anterior and posterior zones, VT ant (k) represents the tidal volume in the pre-k time window, VT post (k) represents the moisture content in the region after the k-th time window. This indicates the phase difference between the front and rear regions;

[0217] In some embodiments, a time window is one respiratory cycle, and M is the number of observation time windows, where M is 3.

[0218] ⑤ The algorithm for the phase variation index is improved, and an optimized phase variation index (PVI) in the form of time history integral is proposed:

[0219]

[0220] PVI unit: mL, where VT represents tidal volume;

[0221] Where V(t) represents the real-time flow rate (measured by the ventilator or estimated using the derivative of EIT impedance change); the integration time range is one basic unit: a single natural respiratory cycle T. cycle From the start of the inspiratory phase to the start of the next inspiratory phase;

[0222] Achievement criteria: Three consecutive stable periods (period variation coefficient <10%)

[0223] The time window is represented as [t0, t0+3T] cycle ]

[0224] The phase difference of the k-th time window is expressed as:

[0225] VT stands for tidal volume (unit: mL);

[0226] f s Respiratory rate (breaths per minute);

[0227] constant 10 3 Dimensional adjustment factor (rad·min / mL → standard unit)

[0228] in,

[0229] arg(X {ant} (f d ,k)) indicates that the front region at the dominant frequency f d The complex spectral value at the location (calculated via STFT);

[0230] f d The range is 0.2-0.35Hz (normal adult respiratory rate range).

[0231] The calculation of CSI (Cyclic Synchronous Heterogeneity Index) requires consideration of both phase differences and amplitude heterogeneity.

[0232] The formula is expressed as:

[0233]

[0234] Where the parameters are:

[0235] α = 0.6, β = 0.4 (container weighting coefficients)

[0236] γ = 1.2, k = 0.8 (nonlinear transformation index)

[0237] Δφ ij =∣φ i (f d )-φ j (f d )∣

[0238] ΔV ij =∣V i (f d )-V j (f d )∣

[0239] Verification benchmark: Respiratory cycle variability must be eliminated within the time window integration (phase difference spectrum smoothing).

[0240] Time alignment is ensured through cubic spline interpolation.

[0241] Implement early warning: When f d If the frequency falls within 0.05-0.2Hz (heartbeat frequency band), the reset mechanism will be automatically triggered;

[0242] CSI calculation is disabled when the impedance signal-to-noise ratio is <15dB. The improved index not only retains the function of measuring phase difference, but also captures the time-series cumulative effect through integration. In a comparative study of 45 cases, the new algorithm improved the positive predictive value by 27% and the cross-validation AUC reached 0.81 (p<0.001).

[0243] The fourth stage is prediction based on extracted features.

[0244] In some embodiments, the features obtained in the third stage are input into a machine learning model to predict the occurrence of Pendeluft. By establishing a validation framework that includes multi-center clinical data, the key performance indicators are shown in Table 1.

[0245] Table 1 Performance analysis of the indicators proposed in this application

[0246]

[0247]

[0248] As can be seen from the table, these indicators have all passed rigorous validation and statistical tests, and may have good efficacy in clinical applications. In this patented solution, the newly designed indicators demonstrate excellent predictive ability for the presence and severity of Pendeloff on the training set.

[0249] The four core indicators VSHI (AUC=0.83), DRC (AUC=0.78), TDI (AUC=0.77) and CSI (AUC=0.81) all passed rigorous validation. The sample size covered 30 cases, and the Delong test showed that the AUC of each indicator was significantly greater than 0.7 of the baseline.

[0250] After integrating the above indicators, the comprehensive model achieved an AUC of 0.91 (95% CI: 0.87-0.95) in 278 multicenter validations, with a macro-average AUC of 0.89 (research papers TABLE2 & S4).

[0251] Ranked validation showed that the AUC distinguishing the Grade 0 and Grade 3 groups was 0.94, and the AUC distinguishing the Grade 2 and Grade 3 groups was 0.82. All comparisons were statistically significant (ANOVA, p < 0.001). Figure 9 ).

[0252] Clinical validation further confirmed that when using a dual standard of VSHI ≥ 10 and model confidence ≥ 0.5, the sensitivity for severe Pendeluft reached 89.6% (CI: 85.3-93.2%). Sample analysis was performed using the Hosmer-Lemeshow test to confirm good calibration (Brier score = 0.15), and all validation procedures conformed to the hybrid architecture design of ResNet-50+BiLSTM.

[0253] In some embodiments, the predictive performance of the new index proposed in this application was independently verified, as shown in Table 2.

[0254] Table 2 shows the performance of the metrics used to verify the existence of Pendeluft.

[0255]

[0256] In some embodiments, the significance of the proposed new index in distinguishing between groups was determined by inter-group analysis of the proposed new index, as shown in Table 3.

[0257] Table 3. Validation of differences between groups (ANOVA test results)

[0258]

[0259] In some embodiments, the model performance of the new metrics proposed in this application for predicting the presence and severity of Pendeluft was verified, as shown in Table 4.

[0260] Table 4. Predictive validity verification of the new indicators on the training set.

[0261]

[0262] In some embodiments, the above features are fused and then predicted using a deep learning model, including: (1) constructing a feature matrix: organizing the ventilation cycle features and swing breathing features into a unified matrix form; (2) feature fusion processing. The system fuses the above basic indicators with other relevant features to construct a comprehensive feature vector F. This feature vector not only includes direct features such as TDI and PI, but also includes the interaction features between them and the pattern features that change over time. The dimension of the fused feature vector is N×1, where N represents the total number of features. To ensure the comparability of features, the system will standardize the fused features. (3) High-order feature extraction. This step uses deep learning methods to further process the fused features. First, the fused feature vector F is input into a deep neural network, and the hidden layer features H = DNN(F) are obtained through multi-layer nonlinear transformation. Then, the system introduces an attention mechanism, and the features are weighted and enhanced through H' = Attention(H) to highlight the contribution of important features. Finally, the high-order features are mapped to the prediction probability space through the Softmax function to obtain the final prediction result Y = Softmax(H'). This process establishes an end-to-end mapping relationship from the original physiological signal features to the prediction results.

[0263] This multi-level feature extraction and fusion method has significant advantages: it can not only capture various features of swinging respiration, but also automatically learn the complex relationships between features through deep learning, improving the accuracy and reliability of predictions. Simultaneously, the introduction of an attention mechanism allows the model to adaptively focus on the most relevant features, further enhancing system performance. This design provides an accurate and reliable technical solution for assessing swinging respiration in clinical practice.

[0264] In some embodiments, the extracted indicators, namely the ventilation cycle difference index (TDI) and the swing breathing index (PI), are combined to form a basic feature vector, and the interaction features between them (such as TDI×PI) are calculated to form an extended feature set. Next, the features are extended in a spatiotemporal dimension, that is, the changing trends of the indicators are analyzed in the time dimension (such as calculating statistical features within a short time window), and the correlation between regions is analyzed in the spatial dimension (such as calculating the feature correlation between adjacent regions). Finally, all features are weighted and fused using the learned weight matrix W to obtain the final fused feature vector F = W⊙[basic features, interaction features, spatiotemporal features], where the weight matrix W is obtained through model training and optimization.

[0265] The result of feature fusion is then fed into a specially designed deep learning model, which employs a deep network structure containing input layers, convolutional layers, pooling layers, and fully connected layers, effectively handling the complex relationship between temporal features and image features.

[0266] Next, after the model training is complete, the system enters the prediction and evaluation phase. In this phase, the model makes predictions on new data, while simultaneously validating the results and evaluating performance to ensure the reliability of the predictions.

[0267] Finally, the system presents the analysis results through various visualization methods, including heatmaps and time curves, and combines them with clinical data for comprehensive prognostic analysis, providing strong support for medical decision-making. This multi-level analysis method not only improves the accuracy of swing respiration detection but also provides a more comprehensive basis for clinical decision-making.

[0268] Since this application is used for bedside monitoring and processes real-time acquired EIT data, the feature extraction in this application is often based on respiratory cycle extraction, or for accuracy, a time window of three respiratory cycles is used to monitor and balance the influence of noise. Those skilled in the art should understand that this application is capable of processing real-time EIT data from bedside monitoring.

[0269] Figure 3 This is a schematic diagram of a computer device provided in an embodiment of the present invention, such as... Figure 3As shown, the device 2000 may include: one or more processors 2010 and one or more memories 2020; wherein the memories store computer-readable code that, when run by the one or more processors, can perform the methods described above.

[0270] The processor in this embodiment can be an integrated circuit chip with signal processing capabilities. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), an off-the-shelf programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, operations, and logic block diagrams disclosed in this embodiment. The general-purpose processor can be a microprocessor or any conventional processor, and can be based on an x86 or ARM architecture.

[0271] In general, the various exemplary embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. When aspects of embodiments of this disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it will be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as non-limiting examples in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0272] For example, the method or apparatus according to embodiments of this disclosure can also be used by means of Figure 4 The architecture of the computing device 3000 shown is used for implementation. For example... Figure 4 As shown, the computing device 3000 may include a bus 3010, one or more CPUs 3020, a read-only memory (ROM) 3030, a random access memory (RAM) 3040, a communication port 3050 connected to a network, an input / output component 3060, a hard disk 3070, etc. The storage devices in the computing device 3000, such as the ROM 3030 or the hard disk 3070, may store various data or files used for processing and / or communication of the methods provided in this disclosure, as well as program instructions executed by the CPU. The computing device 3000 may also include a user interface 3080. Of course, Figure 4 The architecture shown is merely exemplary and can be omitted as needed when implementing different devices. Figure 4 One or more components in the computing device shown.

[0273] This invention also includes a computer-readable storage medium, such as... Figure 5The diagram illustrates a storage medium 4000 provided in an embodiment of the present invention. The computer storage medium 4020 stores computer-readable instructions 4010. When the computer-readable instructions 4010 are executed by a processor, the method described above according to embodiments of the present disclosure can be performed. The computer-readable storage medium in the embodiments of the present disclosure may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. Non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), Synchronous Link Dynamic Random Access Memory (SLDRAM), and Direct Memory Bus Random Access Memory (DR RAM). It should be noted that the memory used in the methods described herein is intended to include, but is not limited to, these and any other suitable types of memory.

[0274] This disclosure also provides a computer program product or computer program that, when executed by a processor, implements the steps of the above-described method, such as... Figure 2 As shown, the computer program product or computer program includes:

[0275] Acquisition module 201: Used to acquire EIT data of the patient's lung region over a period of time;

[0276] Region segmentation module 202: used to segment the lung region into multiple regions, extract EIT data from multiple regions, and obtain the EIT signal of each region;

[0277] Data preprocessing module 203: used to obtain moisture resistance characteristics based on the EIT signal, the moisture resistance characteristics including:

[0278] Moisture resistance amplitude difference within the region, and moisture resistance amplitude difference between adjacent regions;

[0279] Feature extraction module 204: used to calculate EIT ventilation features based on the tidal resistance features, wherein the EIT ventilation features are swing breathing, and the swing breathing is the ratio of the amplitude difference of tidal resistance between the maximum region to the amplitude difference of tidal resistance in the average region.

[0280] Prediction module 205: used to determine whether the patient has Pendleuft based on the EIT ventilation characteristics.

[0281] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0282] In general, the various exemplary embodiments of this disclosure can be implemented in hardware or dedicated circuitry, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. When aspects of embodiments of this disclosure are illustrated or described as block diagrams, flowcharts, or using some other graphical representation, it will be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as non-limiting examples in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0283] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0284] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between apparatuses or units through some interfaces, and may be electrical, mechanical, or other forms.

[0285] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0286] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0287] The exemplary embodiments of this disclosure described in detail above are merely illustrative and not restrictive. Those skilled in the art will understand that various modifications and combinations of these embodiments or their features can be made without departing from the principles and spirit of this disclosure, and such modifications should fall within the scope of this disclosure.

Claims

1. An automated method for detecting Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence, characterized in that, The method includes: S101: Acquire EIT data of the patient's lung region over a period of time; S102: The lung region is divided into multiple regions, and EIT data of multiple regions are extracted to obtain the EIT signal of each region; the EIT signal of each region is transformed in the frequency domain to obtain the spectrum signal of each region; S103: Obtain moisture resistance characteristics based on the EIT signal, the moisture resistance characteristics including: moisture resistance amplitude difference within a region, moisture resistance amplitude difference between adjacent regions, the moisture resistance amplitude difference between adjacent regions being obtained based on the difference in moisture resistance amplitude differences within adjacent regions; obtain spectral characteristics based on the spectral signal, the spectral characteristics including: phase difference between adjacent regions; S104: Based on the tidal resistance characteristics, EIT ventilation characteristics including swing breathing are calculated, wherein the swing breathing is the ratio of the maximum value of the tidal resistance amplitude difference between adjacent regions to the average value of the tidal resistance amplitude difference between all adjacent regions; based on the spectral characteristics, EIT ventilation characteristics including ventilation cycle differences are calculated, wherein the ventilation cycle differences are the average value of the phase difference between adjacent regions. S105: Determine whether the patient has Pendleuft based on the EIT ventilation characteristics.

2. The automatic detection method for Pendeloff in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 1, characterized in that, The swinging breathing is represented as: in, Indicates swaying breathing. This represents the difference in moisture resistivity amplitude between adjacent regions i and j, and max() represents finding the maximum value. This represents the average value of the difference in moisture resistance amplitude between all regions.

3. The automatic detection method for Pendeloff in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 2, characterized in that, If a patient develops Pendleuft, the EIT ventilation characteristics are input into a classifier to determine the severity of the patient's Pendleuft.

4. The automatic detection method for Pendeloff in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 1, characterized in that, The difference in ventilation cycles is expressed as follows: in, Indicates differences in ventilation cycles. N represents the phase difference between adjacent regions i and j, and N represents the total number of phase differences.

5. The automatic detection method for Pendeloff in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 1, characterized in that, The lung region is divided into four areas: front, back, left, and right. The difference in ventilation cycle is expressed as follows: TDI represents the average ventilation cycle difference among four adjacent regions. These represent the phase difference between four adjacent regions.

6. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 4, characterized in that, The phase difference between adjacent regions i and j is expressed as: in, This represents the phase difference between adjacent regions i and j. This represents the frequency domain signal of the i-th region; This indicates taking the phase value corresponding to the dominant frequency of the spectrum; If a patient develops Pendleuft, the EIT ventilation characteristics are input into a classifier to determine the severity of the patient's Pendleuft.

7. The automatic detection method for Pendeloff in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 1, characterized in that, The moisture resistance characteristics also include: ventilation volume differences between adjacent regions and time delays between adjacent regions; the EIT ventilation characteristics also include the ventilation synchronization heterogeneity index. The ventilation synchronization heterogeneity index is calculated based on the spectral characteristics, the ventilation volume difference between adjacent regions, and the time delay between adjacent regions. The ventilation synchronization heterogeneity index is obtained based on the standard deviation of the phase difference between adjacent regions, the standard deviation of the ventilation volume difference between adjacent regions, and the mean of the time delay between adjacent regions.

8. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 7, characterized in that, The ventilation synchronization heterogeneity index is calculated as follows: in, Indicates the ventilation synchronization heterogeneity index. The standard deviation of the phase difference between adjacent regions. The standard deviation of ventilation differences between adjacent zones. This represents the average time delay between adjacent regions.

9. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 7, characterized in that, The lung region is divided into four regions: front, back, left, and right. The standard deviation of the phase difference between adjacent regions is calculated as follows: in, The standard deviation of the phase difference between adjacent regions. These represent the phase differences between adjacent regions 1 and 2, 1 and 3, 2 and 4, and 3 and 4, respectively.

10. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 7, characterized in that, The lung region is divided into four areas: front, back, left, and right. The standard deviation of the ventilation difference between adjacent areas is calculated as follows: in, The standard deviation of ventilation differences between adjacent zones. These represent the differences in ventilation volume between adjacent regions 1 and 2, 1 and 3, 2 and 4, and 3 and 4, respectively.

11. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 7, characterized in that, The lung region is divided into four areas: front, back, left, and right. The average time delay between adjacent areas is calculated as follows: This represents the mean time delay between adjacent regions. These represent the time delays between adjacent regions 1 and 2, 1 and 3, 2 and 4, and 3 and 4, respectively.

12. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 11, characterized in that, The time delay between adjacent regions is expressed as: in, Indicates the time delay between adjacent regions. This represents the time-domain ventilation signal after preprocessing in the i-th region. This represents the time-domain ventilation signal after preprocessing in the j-th region. This indicates that the weighted average of the peak times of each respiratory cycle within the entire monitoring period is taken.

13. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 1, characterized in that, When the multiple regions are two regions including a front region and a back region, the EIT signals of the front region and the back region are extracted respectively, and the frequency domain transformation is performed to obtain the spectrum signals of the front region and the back region. Based on the spectral signal, spectral features are obtained, including: the phase difference between adjacent regions; Based on the EIT signal, the tidal resistance characteristics are obtained, and the tidal resistance characteristics further include: ventilation rate; Based on the spectral characteristics and the ventilation volume, the EIT ventilation characteristics, including the optimized phase variation index, are calculated; the optimized phase variation index is expressed as: in, The optimized phase variation index is represented by VT, which represents ventilation rate. Respiratory rate, This represents the phase difference of the k-th time window, a constant. This is the dimension adjustment factor.

14. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 13, characterized in that, The method for calculating the phase difference of the k-th time window includes: This indicates that in the k-th time window, the front region is at the dominant frequency. Complex spectral values ​​at; This indicates that in the k-th time window, the back region is at the dominant frequency. The complex spectrum value at that location.

15. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 13, characterized in that, The EIT ventilation characteristics also include a dynamic airflow redistribution coefficient, which is calculated based on the spectral characteristics and the ventilation rate; the dynamic airflow redistribution coefficient is expressed as: The lung region is divided into anterior and posterior zones. This represents the ventilation volume in the front zone of the k-th time window. This represents the ventilation volume in the region following the k-th time window. ) represents the phase difference between the front and back regions of the k-th time window; M represents the number of selected observation time windows.

16. The automatic detection method for Pendeluft in critically ill patients based on electrical impedance imaging and artificial intelligence according to claim 15, characterized in that, One time window is one respiratory cycle, and M is 3.

17. A computer device, characterized in that, The device includes: a memory and a processor; the memory is used to store a computer program; the processor executes the computer program to implement the steps of the method according to any one of claims 1-16.

18. A computer-readable storage medium, characterized in that, It stores a computer program thereon, which, when executed by a processor, implements the steps of the method as described in any one of claims 1-16.

19. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the steps of the method described in any one of claims 1-16.

Images