Construction method and system of child OSA auxiliary diagnosis model based on acoustic immittance

Abstract

The invention discloses a construction method and system of a child OSA auxiliary diagnosis model based on acoustic immittance. According to the method, from the perspective of middle ear function and craniofacial anatomy joint analysis, a model is constructed; through systematic integration of acoustic immittance parameters (middle ear pressure, sound order value and the like), clinical signs (tonsil indexing, mouth breathing and the like) and facial imaging features (such as A / N ratio, soft palate length and the like determined based on nasopharynx side position X films), a children OSA auxiliary diagnosis model based on multi-modal data fusion is constructed. Compared with the prior art, the method has the advantages that easily obtained data parameters such as middle ear functions and craniofacial anatomy are creatively integrated, a high-precision and low-cost OSA auxiliary diagnosis tool is constructed, non-invasive and efficient screening replaces complex and expensive PSG detection, an objective and reliable solution is provided for early screening and grading diagnosis and treatment of children OSA, and the method is suitable for popularization and application. And the method is particularly suitable for large-scale application in regions deficient in medical resources.

Description

Technical Field

[0001] This invention belongs to the field of combining artificial intelligence with obstructive sleep apnea (OSA) in children, and specifically relates to a method and system for constructing an auxiliary diagnostic model for children's OSA based on acoustic impedance. Background Technology

[0002] Obstructive sleep apnea (OSA) is a common sleep disorder in children (2-18 years old), with a prevalence of approximately 1%-6%. Compared to adults, children with OSA have unique manifestations, such as hyperactivity, emotional difficulties, declining academic performance, and inattention. Furthermore, childhood OSA can cause a variety of serious complications, leading to cardiovascular disease, growth and developmental disorders, cognitive impairment, craniofacial developmental abnormalities, and secretory otitis media.

[0003] Currently, the diagnosis of childhood OSA relies on polysomnography (PSG). However, PSG testing has many limitations, such as requiring a specialized sleep laboratory, connecting multiple measurement channels, and requiring overnight parental care. In summary, PSG testing is complex, uncomfortable, time-consuming, costly, and often results in poor child cooperation, making it unsuitable for rapid or large-scale screening.

[0004] In addition to PSG, a complex method unsuitable for rapid or large-scale screening, existing technologies also employ questionnaires to screen children suspected of having OSA, such as collecting medical history data and inquiring about snoring during sleep. However, while existing questionnaires are practical, low-cost, and efficient methods, they suffer from strong subjectivity, lack of objective data support, and susceptibility to cultural and linguistic differences, leading to inaccurate and unobjective reflection of childhood OSA. Therefore, using questionnaires to assess childhood OSA still has significant limitations. For these reasons, childhood OSA is characterized by a high incidence but a low diagnosis rate.

[0005] Therefore, there is an urgent need in this field to develop an innovative solution that is fast and objective and accurate, relative to PSG and questionnaire methods, to assist in the diagnosis of OSA in children.

[0006] The information disclosed in the background section is only intended to enhance the understanding of the background of the present invention, and therefore may contain information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] To address the shortcomings of existing technologies, this invention proposes a method for constructing a diagnostic model for childhood OSA based on acoustic impedance, whereby childhood OSA refers to obstructive sleep apnea. The method includes the following steps: S1: Obtain basic information on all subjects suspected of having OSA, as well as OSA-related symptoms and clinical signs; S2: Obtain all acoustic impedance parameters for all subjects; S3: Obtain the nasopharyngeal lateral radiograph parameters for all subjects; S4: Obtain the obstructive apnea-hypopnea index (OAHI) of all subjects after overnight polysomnography (PSG), and further classify the subjects according to the obstructive apnea-hypopnea index (OAHI) into children with OSA and children without OSA. S5: Perform univariate and multivariate logistic regression analyses on the basic information, OSA-related symptoms, clinical signs, and nasopharyngeal lateral X-ray parameters of children with OSA to identify independent predictors of childhood OSA. S6: Based on the independent predictors, further construct the basic predictive model M0 for children's OSA based on multivariate logistic regression; S7: Based on the basic prediction model M0, the diagnostic value of each acoustic impedance index for OSA is explored through multivariate logistic regression, and different acoustic impedance indices are fused into the basic prediction model to obtain different acoustic impedance fusion models. S8: Compare the performance of various acoustic guidance anti-fusion models to find the optimal model.

[0008] Preferably, the following steps are also included: S9: Construct a nomogram prediction model based on the best model.

[0009] Preferred basic information includes: age, gender, and body mass index (BMI).

[0010] Preferred OSA-related symptoms include: snoring, mouth breathing, and breath-holding.

[0011] Preferably, clinical signs include tonsil grading.

[0012] Preferably, the acoustic impedance parameters include: equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, and pressure gradient.

[0013] Preferred parameters for lateral nasopharyngeal X-ray include: maximum adenoid width A, upper respiratory tract width N, A / N ratio, anterior skull base length (ASL), soft palate length (SPL), maxillary anteroposterior dimensions (ANS-PNS), mandibular length (Go-Me), and skull base angle (NS-S.Ba). Preferably, in step S8, Delong's test was used to compare the performance of various acoustic duct anti-fusion models in order to find the optimal model.

[0014] Furthermore, the present invention also discloses a computer storage medium, wherein the storage medium includes computer instructions that, when executed on a computer, cause the computer to execute any of the foregoing methods for constructing a child OSA-assisted diagnostic model based on acoustic impedance.

[0015] Furthermore, this invention also discloses a system for an auxiliary diagnostic model of pediatric OSA based on acoustic impedance, wherein the system comprises: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for constructing a child OSA auxiliary diagnostic model based on acoustic impedance as described above, and to perform auxiliary diagnosis on suspected OSA patients based on the optimal model.

[0016] Compared with the prior art, the present invention has the following advantages: This invention fully utilizes the basic information of children with OSA, OSA-related symptoms, and clinical signs, and combines various tympanometry indicators to enable the construction method described in this invention to be used relatively objectively and accurately to construct the optimal model, and further realize an auxiliary diagnostic system. This is conducive to the rapid triage of children with OSA and the priority recommendation of PSG testing, while breaking through the clinical bottlenecks of traditional screening questionnaires being highly subjective, PSG resources being scarce, and the diagnostic capabilities of primary healthcare institutions being insufficient. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating the method for constructing a child OSA-assisted diagnostic model based on acoustic impedance in one embodiment of the present invention. Figure 2 This is a Nomogram of the optimal model obtained based on the construction method in another embodiment of the present invention; Figure 3 This is another embodiment of the present invention, showing the ROC curve of the best model under the training set and external validation set conditions respectively; Figure 4 In another embodiment of the present invention, the calibration curve and DCA curve of the best model are shown in the training set and external validation set respectively. Detailed Implementation

[0018] The following will refer to the appendix. Figures 1 to 4 Specific embodiments of the invention are described in detail below. While specific embodiments of the invention are shown in the accompanying drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the invention and to fully convey the scope of the invention to those skilled in the art.

[0019] It should be noted that certain terms are used in the specification and claims to refer to specific components. Those skilled in the art will understand that different terms may be used to refer to the same component. This specification and claims do not distinguish components based on differences in terminology, but rather on differences in function. The terms "comprising" or "including" used throughout the specification and claims are open-ended and should be interpreted as "comprising but not limited to." The following descriptions are preferred embodiments for carrying out the invention; however, these descriptions are for the purpose of understanding the general principles of the specification and are not intended to limit the scope of the invention. The scope of protection of this invention is determined by the appended claims.

[0020] To facilitate understanding of the embodiments of the present invention, further explanations and descriptions will be provided below with reference to the accompanying drawings and specific embodiments. The accompanying drawings do not constitute a limitation on the embodiments of the present invention.

[0021] In one embodiment, the present invention proposes a method for constructing a diagnostic model for childhood OSA based on acoustic impedance, wherein childhood OSA refers to obstructive sleep apnea, and the method includes the following steps: S1: Obtain basic information on all subjects suspected of having OSA, as well as OSA-related symptoms and clinical signs; S2: Obtain all acoustic impedance parameters for all subjects; S3: Obtain the nasopharyngeal lateral radiograph parameters for all subjects; S4: Obtain the obstructive apnea-hypopnea index (OAHI) of all subjects after overnight polysomnography (PSG), and further classify the subjects according to the obstructive apnea-hypopnea index (OAHI) into children with OSA and children without OSA. S5: Perform univariate and multivariate logistic regression analyses on the basic information, OSA-related symptoms, clinical signs, and nasopharyngeal lateral X-ray parameters of children with OSA to identify independent predictors of childhood OSA. S6: Based on the independent predictors, further construct the basic predictive model M0 for children's OSA based on multivariate logistic regression; S7: Based on the basic prediction model M0, the diagnostic value of each acoustic impedance index for OSA is explored through multivariate logistic regression, and different acoustic impedance indices are fused into the basic prediction model to obtain different acoustic impedance fusion models. S8: Compare the performance of various acoustic guidance anti-fusion models to find the optimal model.

[0022] For the above embodiments, it can be understood that the present invention makes full use of the basic information of children with OSA, OSA-related symptoms and clinical signs, and combines various indicators of acoustic impedance, so that the construction method described in the present invention can be used to construct the best model relatively objectively and accurately, and can further realize a pediatric OSA auxiliary diagnostic system based on the constructed best model. This is conducive to the rapid triage of children with OSA and the priority recommendation of PSG testing, while breaking through the clinical bottlenecks of traditional screening questionnaires being highly subjective, PSG resources being scarce, and the diagnostic capabilities of primary medical institutions being insufficient.

[0023] In another embodiment, the construction method further includes the following steps: S9: Construct a nomogram prediction model based on the best model.

[0024] In another embodiment, the basic information includes: age, gender, and body mass index (BMI).

[0025] In another embodiment, OSA-related symptoms include snoring, mouth breathing, and breath-holding.

[0026] In another embodiment, clinical signs include tonsil grading.

[0027] In another embodiment, the acoustic impedance parameters include: equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, and pressure gradient.

[0028] In another embodiment, the parameters of the nasopharyngeal lateral radiograph include: maximum adenoid width A, width of the skeletal upper respiratory tract N, A / N ratio, anterior skull base length (ASL), soft palate length (SPL), anteroposterior maxillary dimension (ANS-PNS), mandibular length (Go-Me), and skull base angle (NS-S.Ba). In another embodiment, in step S4, the subjects are classified according to the following criteria: For children without OSA, OAHI < 1 time / hour; for children with OSA, OAHI ≥ 1 time / hour.

[0029] In another embodiment, in step S8, Delong's test was used to compare the performance of various acoustic duct anti-fusion models in order to find the optimal model.

[0030] The following provides a more specific example for further explanation: In another embodiment, the present invention proposes a method for constructing a child OSA-assisted diagnostic model based on acoustic impedance, comprising the following steps: (1) Based on strict medical inclusion and exclusion criteria, the following basic information of hundreds of suspected OSA children, i.e. subjects, was collected: age, sex, body mass index (BMI), and the following OSA-related symptoms: snoring, mouth breathing, breath-holding, and the following clinical signs, including tonsil grading. (2) Collect the acoustic impedance results of the child, including equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, and pressure gradient; (3) Measure the nasopharyngeal lateral radiographs of the child, including the maximum width of the adenoids (A), the width of the upper respiratory tract (N), the A / N ratio, the anterior skull base length (ASL), the soft palate length (SPL), the anteroposterior dimensions of the maxilla (ANS-PNS), the length of the mandible (Go-Me), and the skull base angle (NS-S.Ba). (4) Obtain the OAHI index of the subjects after overnight PSG, and classify the subjects into OSA patients and non-OSA children; (5) Perform univariate and multivariate logistic regression analysis on the basic information of children with OSA, OSA-related symptoms, clinical signs and nasopharyngeal lateral X-ray parameters to find independent predictive factors; (6) Based on independent predictors, construct the basic predictive model Model 0 for children's OSA using multivariate logistic regression; (7) Based on Model 0, the diagnostic value of each acoustic impedance index for OSA was explored by multivariate logistic regression, and different acoustic impedance indices were fused into the basic prediction model to obtain different acoustic impedance fusion models Model 1-Model 5 as new models corresponding to each acoustic impedance index; (8) The Delong's test was used to compare the effectiveness of each model to determine the best model. Then, a nomogram prediction model was constructed based on the best model, and the model was externally validated and evaluated.

[0031] In another embodiment, in step (4), the subjects are classified according to the following criteria: For children without OSA, OAHI < 1 time / hour; for children with OSA, OAHI ≥ 1 time / hour.

[0032] In another embodiment, in step (5), the expression for the univariate logistic regression analysis is: Log{ p / (1- p )}= β 0+ β i X i inp The probability of developing OSA; X i For single independent variables (such as age, sex, body mass index, snoring, mouth breathing, breath-holding, tonsil grading, equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, pressure gradient, maximum adenoid width (A), width of the skeletal upper airway (N), A / N ratio, anterior skull base length (ASL), soft palate length (SPL), anteroposterior maxillary dimensions (ANS-PNS), mandibular length (Go-Me), skull base angle (NS-S.Ba), etc.); β i The regression coefficients passed the significance test. p <0.05) Select candidate variables from all independent variables, where i ranges from 1 to n.

[0033] For example, the expression for multivariate logistic regression analysis is as follows: Log{ p / (1- p )}= β 0+ β 1 X 1+ β 2 X 2+…+ β n X n Therefore, multivariate joint modeling is performed on the variables after single-factor screening, and the final m independent predictors are determined by stepwise regression (AIC criterion), where m is less than or equal to n.

[0034] In another embodiment, in step (6), the expression for constructing the basic prediction model is: P (OSA) = 1 + 1 / e -(β0+β1X1+...+βmXm) in P The probability of developing OSA X j For each of the final m independent predictors, there is a single independent variable (e.g., snoring, mouth breathing, tonsil grading, age, A / N ratio, soft palate length). β j for X j The corresponding regression coefficients, j, range from 1 to m.

[0035] In another embodiment, in step (7), the expression for any new model after fusion is: P 新 (OSA) = 1 + 1 / e -(β0+β1X1+...+βmXm +γY) Where Y is a specific acoustic impedance index, such as equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, or pressure gradient; γ is the regression coefficient corresponding to Y.

[0036] In another embodiment, in step (8), a nomogram prediction model is constructed using R software based on the best model to visualize the model. The performance of the prediction model is then evaluated using evaluation methods based on an external validation set, including all appropriate means of evaluating model performance such as receiver operating characteristic (ROC) curve analysis, further calibration curves, Hosmer-Lemeshow tests, clinical efficacy curves, etc.

[0037] See Figure 1 In another embodiment, the present invention proposes a method for constructing a child OSA-assisted diagnostic model based on acoustic impedance, comprising the following steps: (1) Collect demographic data of children suspected of having OSA. The demographic data of children suspected of having OSA includes: basic information (age, sex, body mass index), OSA-related symptoms (snoring, mouth breathing, apnea) and clinical signs (tonsil grading). All the above information comes from legitimate medical records. These children all underwent sleep apnea monitoring, nasopharyngeal lateral X-ray, and tympanometry. Before the examination, the guardian of each subject signed an informed consent form.

[0038] Inclusion criteria: 1) Children aged 2-12 years with suspected OSA; 2) Has undergone a PSG check; 3) Has undergone a lateral nasopharyngeal radiograph; 4) Has undergone acoustic impedance testing.

[0039] Exclusion criteria: 1) Baseline data is missing; 2) Children with craniofacial diseases; 3) Children with middle ear diseases other than OME, such as cholesteatoma of the external auditory canal, tympanic membrane trauma, congenital ear malformation, etc. 4) Children with muscular and nervous system disorders, such as congenital abnormalities in craniofacial structure or neuromuscular regulation.

[0040] (2) Collect the acoustic impedance results of the child, including equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, and pressure gradient; For example, the acoustic impedance detection in step (2) specifically involves: tympanic pressure measurements are performed using the Titan device (Interacoustics, Denmark, with the IMP440 / WBT440 modules installed). Probe tone stimulation is conducted at 226 Hz. Tympanograms / acoustic blockage rates are measured using a positive and negative pressure scanning method of +200–300 daPa. The equivalent ear canal volume, middle ear resonant frequency, acoustic compliance, middle ear pressure, and pressure gradient are automatically analyzed by software. The left and right ears of the subjects are randomly selected for testing.

[0041] (3) Measure the nasopharyngeal lateral X-ray of the child, including the maximum width of the adenoids (A), the width of the upper respiratory tract (N), the A / N ratio, the length of the anterior skull base, the length of the soft palate, the anteroposterior dimensions of the maxilla, the length of the mandible, and the skull base angle; For example, the nasopharyngeal lateral radiograph examination in step (3) is specifically performed as follows: A lateral cephalometric radiograph is obtained in a standardized manner using an Orthophos X-ray device OC-100 (Instrumentarium Imaging Company, Finland). The distance between the measuring instrument and the head is fixed at 150 cm, and the subject stands naturally. The subject's line of sight is parallel to the ground, and the radiograph is taken while the subject is not speaking or swallowing.

[0042] For example, the measurement and evaluation of the nasopharyngeal lateral radiographs were jointly performed by two otolaryngologists with more than 10 years of clinical experience. The specific procedure was as follows: Researcher A manually traced the image outline using 0.03-inch (approximately 0.76 mm) thick cellulose acetate tracing paper, and Researcher B independently verified the trace. To assess intra-personal error, 10 images were randomly selected two weeks later and measured a second time by the same researcher. If the difference between the two measurements was significant (linear parameter deviation > 0.5 mm or angular parameter deviation > 0.5 degrees), a third measurement was performed, and the average of the two closest results was taken as the valid data. It should be noted that due to the geometric characteristics of X-ray projection, there is an approximately 7-8% magnification effect on the linear parameters in the image. To correct for this deviation, the actual length was converted using the built-in scale of the image during measurement. The calculation formula is: True length = Image measurement value × (Actual scale length / Image scale display length).

[0043] (4) Obtain the OAHI index of the subjects after overnight PSG, and classify the subjects into OSA and non-OSA children; In this embodiment, step (4) specifically involves: all subjects undergoing PSG monitoring at the Sleep Laboratory of the Sleep Center, Department of Otolaryngology-Head and Neck Surgery, Second Affiliated Hospital of Xi'an Jiaotong University / Department of Otolaryngology-Head and Neck Surgery, Xi'an Children's Hospital. All records are evaluated by certified clinical polysomnography experts. The Obstructive Sleep Apnea-Hypopnea Index (OAHI) is used to assess the severity of sleep apnea. Based on PSG, children are divided into: 1) Non-OSA children: OAHI < 1; 2) OSA children: OAHI ≥ 1.

[0044] It should be noted that in the above embodiments of the present invention, steps (1) to (4) specifically involve: collecting relevant data from 931 children aged 2-12 years suspected of having OSA at the Sleep Center of the Department of Otolaryngology-Head and Neck Surgery, Second Affiliated Hospital of Xi'an Jiaotong University from July 2020 to February 2024, as a training set for subsequent regression analysis, basic prediction model and fusion to obtain a new model; and using relevant data from 353 children suspected of having OSA collected at Xi'an Children's Hospital from October 2020 to May 2023 as an external validation set for further validation of the model.

[0045] (5) Perform univariate and multivariate logistic regression analysis on the basic information of the child, OSA-related symptoms, clinical signs and nasopharyngeal lateral X-ray parameters to find independent predictive factors; The expression for the univariate logistic regression analysis in step (5) is as follows: Log{ p / (1- p )}= β 0+ β i X i in p The probability of developing OSA; X i For single independent variables (such as age, sex, body mass index, snoring, mouth breathing, breath-holding, tonsil grading, equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, pressure gradient, maximum adenoid width (A), width of the skeletal upper airway (N), A / N ratio, anterior skull base length (ASL), soft palate length (SPL), anteroposterior maxillary dimensions (ANS-PNS), mandibular length (Go-Me), skull base angle (NS-S.Ba), etc.); β i The regression coefficients passed the significance test. p <0.05) Select candidate variables from all independent variables, where i ranges from 1 to n.

[0046] For example, the expression for multivariate logistic regression analysis is as follows: Log{ p / (1- p )}= β 0+ β 1 X 1+ β 2 X 2+…+ β n X n Therefore, multivariate joint modeling is performed on the variables after single-factor screening, and the final m independent predictors are determined by stepwise regression (AIC criterion), where m is less than or equal to n.

[0047] (6) After selecting the independent predictors of childhood OSA, construct a basic predictive model for childhood OSA; For example, the expression for constructing the basic prediction model is: P (OSA) = 1 + 1 / e -(β0+β1X1+...+βmXm) in P The probability of developing OSA X j For each of the final m independent predictors, there is a single independent variable (e.g., snoring, mouth breathing, tonsil grading, age, A / N ratio, soft palate length). β j for X j The corresponding regression coefficients, j, range from 1 to m.

[0048] (7) Explore the diagnostic value of various acoustic impedance indices for children's OSA by using Logistic regression, and integrate different acoustic impedance indices into the basic prediction model to obtain different acoustic impedance fusion models, and use them as various new models, such as the following 5 new models in Table 2: Model 1-Model 5; For example, based on the aforementioned training set, the following table (Table 1) was obtained through univariate logistic regression analysis: Table 1. Univariate Logistic Regression Analysis of OSA in Children We found that snoring, mouth breathing, tonsil grading, age, A / N ratio, and soft palate length are risk factors for childhood OSA. P<0.05). Multivariate logistic regression analysis of these variables still revealed that snoring, mouth breathing, tonsil grading, age, A / N ratio, and soft palate length were risk factors for childhood OSA. Further details are provided in Table 2. Table 2 Multivariate Logistic Regression Analysis of OSA in Children ** p <0.01, *** p <0.001 On the one hand, as shown in Table 2, a basic model for predicting childhood OSA (Obstructive Sleep Apnea) was constructed based on multivariate logistic regression, using snoring, mouth breathing, tonsil grading, age, A / N ratio, and soft palate length as predictive variables (Model 0, AUC = 0.845). This fully confirms that snoring, mouth breathing, tonsil grading, age, A / N ratio, and soft palate length are the six independent predictors ultimately identified.

[0049] On the other hand, as shown in Table 2, the five acoustic impedance indices of equivalent ear canal volume, middle ear resonant frequency, acoustic compliance value, middle ear pressure, and pressure gradient are respectively integrated into Model 0 to obtain the new fused models Model 1 to Model 5, with AUC values ​​of 0.847, 0.853, 0.849, 0.867, and 0.851, respectively. Therefore, it can be determined that the optimal model in this embodiment is Model 4.

[0050] (8) The Delong's test was used to compare the performance of each model. A nodal plot was constructed based on the best model, and the model was externally validated and evaluated; among which, The expression for Delong's test in step (8) is: Z = (AUC 1- AUC 2) / √{Var( AUC 1)+Var( AUC 2)-2Cov( AUC 1, AUC 2)} If |Z|>1.96( p If the AUC value is less than 0.05, the difference between the two models is considered significant. As shown in Table 3, Delong's test compares the differences between Model 1-Model 5 and Model 0. Model 4 has the best performance, improving the AUC value by 0.022 compared to Model 0. p<0.05).

[0051] Table 3. Pairwise comparisons of the area under the recipient's work characteristic curve using Delong's test. For example, the model can be visualized using R software, and a nodal plot can be built based on the best model, Model 4, as shown below. Figure 2 As shown.

[0052] As mentioned earlier, in this invention, the training set is based on 931 suspected OSA patients from the Department of Otolaryngology-Head and Neck Surgery at the Second Affiliated Hospital of Xi'an Jiaotong University, and the external validation set is based on 353 suspected OSA patients from Xi'an Children's Hospital. The performance of the final optimal model is compared on both the training set and the external validation set for evaluation. Figure 3 Figures a and b show the ROC curves of the training set and the external validation set. The AUC value of the training set is 0.867 (95% CI: 0.838-0.897), and the AUC value of the external validation set is 0.839 (95% CI: 0.789-0.890). Figure 4 Figures a and b show the calibration curves for the training set and the external validation set.

[0053] Furthermore, the Hosmer-Lemeshow test also showed that there were no statistically significant differences between the training cohort (χ2 = 9.979, p = 0.267) and the validation cohort (χ2 = 8.480, p = 0.388), indicating that the nomogram was well calibrated.

[0054] Figure 4 The clinical efficacy curves for c and d in the figure show that the optimal model constructed in this invention has good performance in both training and external validation set scenarios. The expression for the clinical efficacy curve is as follows: Net income = TP / N - FP / N·pt / (1-pt) Where pt represents the threshold probability, and the model curve is clinically valuable when it is above the "full intervention" or "no intervention" line.

[0055] In summary, compared with PSG and questionnaire methods, this invention has substantial advantages in the following aspects: it simultaneously collects basic information of children, OSA-related symptoms, clinical signs, tympanometry results, and nasopharyngeal lateral X-ray data, objectively and scientifically realizing a pediatric OSA auxiliary diagnostic model that is both fast and accurate compared to existing technologies. Furthermore, the optimal model in the example includes two important risk factors for pediatric OSA: middle ear pressure and craniofacial diseases (e.g., A / N ratio, soft palate length). Based on the model constructed by the method of this invention, due to its large sample size, extensive feature screening, and external validation from other hospitals, the developed model has broad applicability.

[0056] The auxiliary diagnostic model obtained by the construction method of this invention can provide objective auxiliary diagnostic evidence for clinicians through a system including a processor and a memory, enabling rapid triage of children with OSA and priority recommendation of PSG testing. This is beneficial for preventing many adverse health problems through early identification and treatment of OSA, while breaking through the clinical bottlenecks of traditional screening questionnaires being highly subjective, PSG resources being scarce, and the diagnostic capabilities of primary healthcare institutions being insufficient.

[0057] In another embodiment, the present invention also discloses a computer storage medium comprising computer instructions that, when executed on a computer, cause the computer to perform any of the methods described above.

[0058] In another embodiment, the present invention also discloses a system for an acoustic impedance-based assisted diagnostic model for pediatric OSA, wherein the system comprises: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the method for constructing a child OSA auxiliary diagnostic model based on acoustic impedance as described above, and to perform auxiliary diagnosis on suspected OSA patients based on the optimal model.

[0059] In summary, this invention, from the perspective of combined analysis of middle ear function and craniofacial anatomy, systematically integrates acoustic impedance parameters (middle ear pressure, acoustic compliance, etc.), clinical signs (tonsil grading, mouth breathing, etc.), and facial imaging features (e.g., A / N ratio determined based on nasopharyngeal lateral X-ray, soft palate length, etc.) to construct a multimodal data fusion-based pediatric OSA-assisted diagnostic model. Compared to existing technologies, this invention innovatively integrates readily available data parameters from middle ear function and craniofacial anatomy to construct a high-precision, low-cost OSA-assisted diagnostic tool. This non-invasive and efficient screening method replaces the complex and expensive PSG test, providing an objective and reliable solution for early screening and tiered diagnosis and treatment of childhood OSA, and is particularly suitable for large-scale application in areas with limited medical resources.

[0060] The applicant has provided a detailed description of the embodiments of the present invention in conjunction with the accompanying drawings. However, those skilled in the art should understand that the above embodiments are merely preferred examples of the present invention and are not limited to the specific embodiments described above. The detailed description is intended to help readers better understand the spirit of the present invention and is not intended to limit the scope of protection of the present invention. On the contrary, any improvements or modifications made based on the inventive spirit of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for constructing a diagnostic model for childhood OSA based on acoustic impedance, wherein childhood OSA refers to obstructive sleep apnea. The method comprises the following steps: S1: obtaining basic information and OSA-related symptoms and clinical signs of all suspected OSA children; S2: obtaining acoustic immittance indicators of all subjects; S3: obtaining nasopharyngeal lateral X-ray parameters of all subjects; S4: obtaining obstructive apnea hypopnea index (OAHI) of all subjects after overnight polysomnography (PSG), and further classifying the subjects according to the obstructive apnea hypopnea index (OAHI) into OSA children and non-OSA children; S5: performing single factor and multi-factor logistic regression analysis on the basic information, OSA-related symptoms, clinical signs and nasopharyngeal lateral X-ray parameters of the OSA children to find independent predictors of pediatric OSA; S6: constructing a basic prediction model M0 of pediatric OSA based on multi-factor logistic regression according to the independent predictors; S7: exploring the diagnostic value of each acoustic immittance indicator for OSA by multi-factor logistic regression based on the basic prediction model M0, and fusing different acoustic immittance indicators into the basic prediction model to obtain different acoustic immittance fusion models; S8: comparing the performance of each acoustic immittance fusion model to find the best model.

2. The construction method according to claim 1, further comprising the following step: preferably, S9: constructing a nomogram prediction model based on the best model.

3. The construction method according to claim 1, wherein The basic information includes age, gender, and body mass index (BMI).

4. The construction method according to claim 1, wherein The OSA-related symptoms include snoring, mouth breathing, and breath holding.

5. The construction method according to claim 1, wherein The clinical signs include tonsil grading.

6. The construction method according to claim 1, wherein The acoustic immittance indicators include equivalent ear canal volume, middle ear resonance frequency, acoustic immittance, middle ear pressure, and pressure gradient.

7. The construction method according to claim 1, wherein The nasopharyngeal lateral X-ray parameters include adenoid maximum width A, bone upper respiratory tract width N, A / N ratio, anterior skull base length (ASL), soft palate length (SPL), upper jaw anteroposterior dimension (ANS-PNS), mandible length (Go-Me), and skull base angle (N.S-S.Ba).

8. The construction method of claim 1, wherein, In step S8, Delong's test is used to compare the performance of each acoustic immittance fusion model to find the best model.

9. A computer storage medium, wherein, The storage medium comprises computer instructions that, when executed on a computer, cause the computer to perform the construction method of the acoustic immittance-based pediatric OSA auxiliary diagnosis model according to any one of claims 1 to 8.

10. A system for an acoustic immittance-based child OSA auxiliary diagnostic model, wherein, The system comprises: a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein the processor implements the construction method of the acoustic immittance-based pediatric OSA auxiliary diagnosis model according to any one of claims 1 to 8 when executing the program, and performs auxiliary diagnosis on suspected OSA children based on the best model.

Images