Application of numerical simulation of upper airway flow field in pediatric obstructive sleep apnea with adenoid hypertrophy

Introduction

Obstructive sleep apnea (OSA) is characterized by recurrent episodes of partial (hypopnea) or complete (apnea) upper airway (UA) collapse during sleep (1) with incidence ranging from 1–3% (2). Adenoid hypertrophy (AH) plays a significant role in the development of pediatric OSA (2,3). Traditional clinical diagnostic and evaluation methods primarily rely on symptom observation, nasal endoscopy, and imaging modalities such as X-ray and computed tomography (CT). While these methods can demonstrate adenoid anatomical characteristics, they have notable limitations in revealing UA aerodynamic features and the pathological interaction mechanisms with AH (4-6). Thus, assessing nasopharyngeal airflow obstructions caused by AH is difficult with the morphological evaluation methods (5) and new analysis methods are eagerly needed.

From the perspective of aerodynamics, UA stenosis results in air recirculation, producing intraluminal negative pressure within anatomically predisposed areas, a mechanism that may ultimately precipitate UA collapse (7). In-depth studies on UA flow dynamics are necessary to elucidate how the UA’s anatomical configuration relates to its ventilation performance (8). The rapid advancements in fluid dynamics, computer technology and medical imaging have enabled researchers to conduct detailed hydrodynamic analyses of the UA using computational fluid dynamics (CFD) software (9). This method overcomes the drawbacks of in vivo animal trials, offering a repeatable, non-invasive route to explore how UA architecture links to respiratory performance (10). It enables qualitative and quantitative analysis of UA obstructive segments in children with AH (11-13). Existing studies have demonstrated that respiratory CFD indices correlate to some extent with patients’ subjective symptoms (14,15). This technology not only offers a new perspective for elucidating how AH-induced airway obstruction impacts respiratory airflow but also shows broad application prospects in clinical contexts such as aided disease diagnosis, treatment plan optimization, and therapeutic effect evaluation (12,14,16,17). However, CFD is currently primarily applied in fields such as cardiovascular research (18-20), and aerosol inhalation (21,22) with relatively limited research on CFD-based evaluation of AH (3,12,23).

Based on CT image data and our previous studies (24,25), this study constructed three-dimensional (3D) geometric models of children’s UA and performed numerical simulations to analyze respiratory airflow. The research aimed to provide fundamental theoretical support for clinical studies on pediatric AH, as well as a theoretical foundation for analyzing pathogenesis, enabling precise diagnosis, and formulating individualized interventions for related conditions.

Methods

General information

A prospective study was designed to investigate the aerodynamics differences in UA between normal children and OSA patients. Select a typical 8 years old boy of OSA with AH, who was admitted to the Department of Otorhinolaryngology-Head and Neck Surgery at Shanghai Children’s Hospital in April 2025. He was diagnosed with OAS by overnight polysomnography (PSG) and underwent adenoidectomy (AT). The clinical manifestations included snoring, mouth breathing, nasal congestion, inattention during class, hyperactivity, and occasional waking up; physical signs: Slightly dull expression, thick and outward protruding upper lip, protruding front teeth, uneven dental arch, narrow maxillary arch, and high arch of hard palate. The Apnea-Hypopnea Index (AHI) index was 10 evens/hour preoperative, while the postoperative AHI was 1 even/hour. The preoperative lowest oxygen saturation level was 82%, while thar was 95% postoperatively. A healthy control child, matched for body type, weight, and age, was selected; this child had no history of respiratory diseases or relevant clinical symptoms, with head and neck CT revealing normal nasal cavity and UA structures. The pre- and post-operative nasal endoscopic findings, CT images, and clinical data of the child were collected. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shanghai Children’s Hospital, Shanghai Jiao Tong University (approval No. SHCH-IACUC-2022-XMSB-85). Informed consent was obtained from the patient’s parent or legal guardian.

Experimental equipment and software

Nasal endoscope (Olympus Medical Systems, Tokyo, Japan), multi-slice spiral CT scanner (GE LightSpeed VCT, Chicago, USA), medical image and 3D reconstruction software MIMICS 21.0 (Materialise, Leuven, Belgium), finite element (FE) meshing software HYPERMESH (Altair, USA), and CFD analysis software ANSYS 2020R1 (ANSYS, Canonsburg, USA).

Construction of 3D FE models

CT scan

The child was placed in a supine position on the scanning bed, with the head fixed using a head frame to ensure it was positioned in the median sagittal plane and the orbitomeatal plane parallel to the ground. During the CT scan, the child maintained steady breathing and avoided movements such as swallowing or speaking. Scan parameters were as follows: slice thickness 0.625 mm, slice interval 0.5 mm, tube voltage 120 kV, and tube current 240 mA. DICOM format images of CT scans from the eyebrow arch to neck base were collected.

Construction of 3D geometric models and FE models

The DICOM format images data were imported into MIMICS 21.0 software, with the loaded CT series fully covering the anatomical structures of the UA. Due to differences in material density among the respiratory cavity, surrounding soft tissues, and bones, CT images exhibited significant grayscale contrast. Based on these threshold characteristics, the UA was accurately segmented on each CT slice (Figure 1). To avoid interference from boundary effects, the tracheal base was appropriately extended to establish an outlet region. Following post-processing steps such as denoising, morphological correction, and surface smoothing, a 3D model reflecting the geometric morphology of the UA in normal child was ultimately constructed (Figure 2). A 3D FE model meshed with tetrahedral elements was constructed using HYPERMESH software, with detailed meshing shown in Figure 3. To enhance computational convergence, the model was built with shared nodes, comprising 355,536 nodes and 1,894,358 elements.

Figure 1 Regions of the upper airway in the normal child.

Figure 2 3D model of the upper airway in the normal child. 3D, three dimensional.

Figure 3 Detail of mesh.

AT was performed in accordance with the expert consensus on clinical diagnosis and treatment management of AH in children (26). Using the same modeling approach as described earlier, 3D geometric models of the child’s UA were constructed for both preoperative and postoperative states, as shown in Figure 4. In terms of quality analysis for geometric model construction, the established 3D UA models exhibited realistic morphology with clear detail representation, without geometric distortion or jagged artifacts. These models fully preserved the original features of nasal anatomical structures, with surface morphology not overly simplified, and were consistent with clinical measurement data. We demonstrated the accuracy of the model through clinical recognition of the structure and comparison with endoscopy. For FE model construction, the meshing strategy was scientifically rigorous and rationally designed: local refinement was implemented for fine structures with significant curvature changes (e.g., the nasal valve area and turbinate folds), while coarse meshing was employed for large, smooth regions (e.g., the nasal septum plane). This differential meshing scheme ensured the accuracy of flow field computations, effectively controlled the total number of elements, and significantly reduced computational costs in numerical simulations.

Figure 4 3D models of the upper airway in the child with AH before and after surgery. (A) AH model; (B) postoperative model. The differences between geometric models are marked with red circles. 3D, three dimensional; AH, adenoid hypertrophy.

CFD analysis

Based on CT scan data, 3D FE models of the UA were constructed for normal children, pre- and post-AT children, respectively. By applying boundary conditions and utilizing numerical simulation techniques, respiratory airflow dynamic characteristics were systematically analyzed, and airflow passage through the UA was simulated. Following multiple iterative simulations to achieve convergence accuracy, velocity, pressure, volume flow rate and WSS data for each node were stored in result files. Physical parameters at any location in the flow field could be derived from node information, and the distributions of velocity, pressure, and WSS of each section were visualized via contour plots. A comparative analysis of numerical simulation performance across various turbulence models in the human UA showed that under typical respiratory intensities, the k-ω turbulence model more accurately characterizes UA airflow properties (27), and the wall function adopts the k-ε model.

For this simulation, the medium was defined as air, with a density of ρ=1.225 kg/m³ and a dynamic viscosity of µ=1.7894×10⁻⁵ kg/(m·s). Under the rigid wall assumption, the airway in this study was treated as a rigid cavity, and airflow was simplified as an incompressible flow with a constant viscosity coefficient (3). Fine structures (e.g., nasal hairs) were neglected, and the effects of factors such as temperature, humidity, and gravity were not considered. A pressure boundary condition was applied to the inlet (located at the nostrils). Since the anterior nares connect to the external environment, a standard atmospheric pressure of 101,325 Pa was imposed. The outlet was specified as a pressure outlet. To simulate different breathing conditions, boundary pressures of −6, −10, and −20 Pa were imposed at the lower airway outlet. The wall boundary employed the no-slip condition (velocity =0), and flow characteristics near the wall were handled using the standard wall function.

Selection of observation sections

In addition to the inlet section (at the nostrils) and the outlet section (airway outlet), eight other respiratory tract sections were selected as flow field observation sections (Figure 5). Section 1: the end of the nasal vestibule and the beginning of the main nasal passage; Section 2: the starting point of the middle nasal concha and inferior nasal concha; Section 3: the middle of the nasal cavity, being the section with the most complex anatomical structure; Section 4: the region where the bilateral nasal cavities merge (i.e., the narrow plane between the end of the nasal cavity and the beginning of the nasopharynx); Section 5: the lower boundary of the nasopharynx (the plane of the free edge of the soft palate); Section 6: the lower boundary of the velopharyngeal (the epiglottic plane); Section 7: the lower boundary of the laryngopharynx (the plane of the lower edge of the cricoid cartilage); Section 8: the plane of the third tracheal ring.

Figure 5 Airway sections, eight respiratory tract sections were selected as flow field observation sections.

Results

The software autonomously computed the cross-sectional area (CSA) of UA for each slice and determined the smallest CSA. The CFD-POST module was utilized for postprocessing to analyze key aerodynamic parameters. These included airflow streamlines, turbulence, velocity, volumetric flow rates, pressure, and WSS.

3D computational model of the UA

Three different values were used for the grid independence verification of the model. The normal UA model was meshed with 909,274, 1,894,358, and 3,500,435 elements, while the AH model was meshed with 803,482, 1,559,357, and 3,101,421 elements, respectively. Element independence verification results showed that 1,894,358 elements for the normal model and 1,559,357 elements for the AH model were sufficient to meet flow field simulation accuracy requirements while ensuring computational efficiency. Following adenoid resection, the postoperative model was constructed using the same protocol and finally meshed with 1,980,825 elements.

CFD analysis

Analysis of airflow streamline

When the outlet pressure was −6 Pa, the streamline distribution of the UA is shown in Figure 6. In the normal model, the airflow streamlines were generally smooth and regular, with relatively turbulent flow observed in the anterior soft palate and at the epiglottis (Figure 6A). In the AH model, the distribution of airflow turbulence regions was largely similar to that of the normal model; however, due to airflow passing through the narrowed nasopharynx during inspiration, marked turbulence was observed in the upper nasal cavity (Figure 6B). After surgery, turbulence in the upper nasal cavity disappeared, and its airflow characteristics were consistent with the normal model (Figure 6C).

Figure 6 Streamlines of the upper airways. (A) Normal model; (B) AH model; (C) postoperative model. Turbulent flow was observed in the anterior soft palate (○) and upper nasal cavity (△). AH, adenoid hypertrophy.

Analysis of airflow velocity

Under a pressure of −6 Pa, the flow field velocity distributions of the three models are shown in Figures 7,8. In the normal model, after airflow entered the nasal cavity, flow velocity gradually decreased as the cavity volume expanded. The first high-velocity region appeared in Section 5, with a peak velocity of approximately 0.97 m/s; the maximum velocity (V_max) occurred in Section 8 (corresponding to the third tracheal ring), reaching 1.22 m/s, with high-velocity flow primarily concentrated on the tracheal posterior wall. This was because the anatomical structures at these two locations were narrower than in other regions. In contrast, the initial nasal airflow velocity in the AH model was similar to that in the normal model; however, velocity increased significantly at the proximal nasopharynx (Section 4), reaching 1.47 m/s in the narrowed nasopharyngeal region—an approximately 213% increase relative to the peak velocity in the normal model. This model still exhibited high velocity (1.07 m/s) in Section 5, with the distribution of high-velocity flow differing markedly from the normal model, primarily concentrating in the tracheal middle part.

Figure 7 Velocity flow diagrams of different models under −6 Pa inspiratory pressure. (A) Normal model; (B) AH model; (C) postoperative model. Normal model: the first high-velocity region appeared in Section 5 (0.97 m/s), the Vmax occurred in Section 8 (1.22 m/s). AH model: The velocity increased significantly at the proximal nasopharynx (Section 4), reaching 1.47 m/s. The Section 5 has a speed of 1.07 m/s, following closely behind. Postoperative model: the velocity decreased to 0.49 m/s in Section 4, consistent with the normal model’s 0.47 m/s. The velocity in Section 5 was 0.93 m/s, approaching normal levels. AH, adenoid hypertrophy; V_max, maximum velocity.

Figure 8 Velocity curves. (A) −6 Pa inspiratory pressure; (B) −10 Pa inspiratory pressure; (C) −20 Pa inspiratory pressure. AH, adenoid hypertrophy.

In the postoperative model, the CSA of the previously narrowed region in Section 4 expanded, reducing airflow velocity here to 0.49 m/s—largely consistent with the normal model’s 0.47 m/s. The flow velocity in Section 5 was 0.93 m/s, approaching normal physiological levels, while high-velocity flow shifted to the tracheal posterior wall. Thus, flow field characteristics aligned more closely with those of the normal model.

The velocity curves for each section in the three models are shown in Figures 7,8. Under different outlet pressures, airflow velocity in the normal model followed regular patterns: the average velocity of each section increased as inspiratory pressure decreased, with V_max observed in Sections 5 and 8. In the AH model, Sections 5 and 8 also showed high-velocity regions, similar to the normal model. However, the V_max was observed at the narrowed Section 4, reflecting the accelerating effect of airway narrowing on airflow. Analysis of velocity contours revealed that in the normal model, airflow velocity distribution along the trachea was uniform. In the AH model, by contrast, velocity increased as tracheal diameter decreased. Moreover, as airflow passed through the narrowed region, sudden expansion of the downstream airway led to eddy formation, with flow disturbance significantly greater than in the normal model. In the postoperative model, expansion of the CSA in the previously narrowed region led to a significant decrease in peak velocity, with velocity distribution largely consistent with the normal model.

Analysis of airflow pressure

When the inspiratory pressure was −6 Pa, in the normal airway model (Figure 9A), pressure from Section 1 to Section 8 showed a decreasing trend, with notable fluctuations in the region near the glottis but an overall smooth trend. The minimum pressure was observed at Section 8 (−3.51 Pa). The pressure distribution of the AH model (Figure 9B) showed significant differences compared to the normal model: the narrowed region exhibited more pronounced pressure gradient changes, with the pressure difference increasing notably. In the AH model, the pressure difference at the corresponding narrowed region was −2.97 Pa—a significant 67.80% increase compared to the normal model. Notably, the minimum pressure in both the normal and AH models occurred at Section 8, reflecting a similar distribution pattern of extreme pressure locations.

Figure 9 Pressure of different models under −6 Pa inspiratory pressure. (A) Normal model; (B) AH model; (C) postoperative model. AH, adenoid hypertrophy.

Analysis of pressure contours for the postoperative model (Figure 9C) showed that with expansion of the CSA in the narrowed segment, pressure in the originally narrowed region rebounded markedly to −1.64 Pa, approaching the normal model’s pressure level (−1.77 Pa). The overall pressure distribution trend was highly consistent with that of the normal model.

A comparative analysis of average pressure across sections in the three models under different inspiratory pressures (Figures 9,10) showed that in the normal model, total pressure from the nasal inlet to the tracheal outlet decreased gradually in a gradient, with a local high-pressure region forming in the region near the glottis. In the AH model, pressure in the narrowed region differed significantly from the normal model, with a markedly greater pressure drop. In non-narrowed sections, the pressure drop trend aligned with the normal model, though pressure values were generally higher. In the postoperative model, with expansion of the CSA in the stenotic segment, pressure at Section 4 increased significantly, with a maximum increase of 44.78% observed at an inspiratory pressure of −6 Pa. As stenosis in this region resolved, the original pressure gradient gradually dissipated, and the overall pressure drop trend flattened—gradually approaching that of the normal model.

Figure 10 Average pressures at each section of the three models under different inspiratory pressures. (A) −6 Pa inspiratory pressure; (B) −10 Pa inspiratory pressure; (C) −20 Pa inspiratory pressure. AH, adenoid hypertrophy.

Notably, a relatively high-pressure area—similar to that in the normal model—reappeared in the region near the glottis of the postoperative model, indicating that the postoperative airway pressure distribution had markedly approached the normal physiological state.

Analysis of volume flow rate

Figure 11 compared volume flow rates between the normal model and the AH model: at −6 Pa, the normal model had a volume flow rate of 2.84 L/min, while the AH model had 0.87 L/min—with the normal model reaching 3.26 times that of the AH model. At −20 Pa, the normal model’s flow rate (6.43 L/min) exceeded the AH model (2.59 L/min) by 3.84 L/min.

Figure 11 Volume flow rates of the three models under different inspiratory pressures. AH, adenoid hypertrophy.

For the AH model, volume flow rate increased from 0.87 L/min at −6 Pa to 2.59 L/min at −20 Pa, while the normal model’s flow rate rose significantly from 2.84 L/min at −6 Pa to 6.43 L/min at −20 Pa—indicating distinct growth patterns. For the postoperative model, flow rate increased from 4.30 L/min at −6 Pa to 7.07 L/min at −20 Pa, showing a marked rise and trending toward normal physiological levels.

Analysis of wall shear stress (WSS)

The WSS of the three models was generally low, with maximum values all occurring at the narrowest anatomical segment of the respective airway (Figure 12). Notably, maximum WSS tended to increase as inspiratory pressure decreased: at −20 Pa, the peak in the normal model was 0.50 Pa; due to airway narrowing, the peak in the AH model increased to 0.66 Pa; while the peak in the postoperative model decreased to 0.53 Pa, largely consistent with that of the normal model.

Figure 12 WSS of different models at −6 Pa inspiratory pressure. (A) Normal model; (B) AH model; (C) postoperative model. AH, adenoid hypertrophy; WSS, wall shear stress.

CFD analysis showed that WSS was positively correlated with flow velocity: as flow velocity increased, WSS gradually rose, with significant stress concentration occurring especially in the narrowest airway segments. Notably, worsened stenosis further elevated WSS.

These mechanical relationships are visually depicted in Figure 13, which clearly demonstrates how airway geometry influences WSS distribution.

Figure 13 WSS of three models under different inspiratory pressures. AH, adenoid hypertrophy; WSS, wall shear stress.

Discussion

Pediatric OSA is characterized by frequent occurrence of the UA obstruction during sleep, which interferes with children’s normal ventilation and sleep structure, and causes a series of pathophysiological changes (1,28). Common symptoms include daytime drowsiness, fatigue, snoring, irritability, morning headaches, and memory loss (29). The adenoids are susceptible to pathologic hypertrophy triggered by viral infections or allergic reactions (30). AH is the most common cause of pediatric OSA (31,32), and AT has been the preferred treatment approach (2). The gold standard for diagnosis of pediatric OSA is overnight PSG, which is costly, time consuming and not widely available (33). What’s more, the relevant anatomic information for investigating UA functions and surgical decisions is not available (34). Currently, nasal endoscopy is considered the standard for clinically evaluating the size of the adenoids. Additionally, lateral cephalogram and CT are also commonly used to assess the size, shape, and position of adenoids in children (35,36). However, these methods only provide two-dimensional data of a relatively simple 3D form, so they cannot evaluate the condition of UA ventilation (37,38).

In recent years, advancements in computer techniques, the FE analysis software and particularly CFD have propelled airway research. CFD simulates fluid behaviour adopting the Navier-Stokes equations and has been applied to clinical medicine, assisting studies on airway dynamics (3). The CFD simulation generally uses 3D imaging reconstructed from CT or magnetic resonance imaging for airway modelling (38). CT scanning is a fast and noninvasive, so, we used CT images for 3D modeling in the present study. Given specific inlet parameters, the UA’s structure and boundary conditions intrinsically mold its aerodynamic attributes, which include airflow streamline, pressure, velocity, and WSS (3). In recent years, CFD has been extensively utilized to assess respiratory function in patients (10), enabling both qualitative visualisation of the airflow and accurate quantitative analyses of airflow fields (39-41). There is evidence that the parameters obtained by CFD analysis can differentiate patients with OSA from non-OSA (42), assess the severity of OSA, and evaluate the post-operative efficacy (38,42). In addition, close agreement was achieved between numerical predictions and experimental data (41). Nevertheless, most previous studies have focused on adult subjects, while aerodynamic characteristics in children has been less researched. For children with AH, clinicians are still unable to quantify the relationship between adenoid morphology and the airflow characteristics of UA, which inevitably leads to a degree of uncertainty in diagnostic and treatment decisions (41).

This study developed individualized 3D UA models from CT data, reconstructed airway geometric remodeling processes caused by adenoid morphological changes, and revealed the impact of AH on airway airflow dynamics using CFD simulations. Results showed that increased adenoid volume decreased the CSA of the nasopharyngeal airway to 31.7% of the normal model’s (normal CSA: 302.3 mm², AH CSA: 95.8 mm²). According to the law of mass conservation in fluid mechanics, airflow velocity is inversely proportional to the CSA of UA, resulting in a significant increase in peak velocity in the stenotic region. Simulation results further showed that respiratory space occupied by AH induced turbulent effects, manifested as disorganized airflow streamlines and increased streamline curvature. The airflow streamlines became disorganized, and downstream separation occurred, forming a high-speed jet in the velopharyngeal (8). This turbulence, in turn, disrupted the normal respiratory flow field and increased airway resistance—consistent with clinically documented symptoms such as stridor and nasal congestion, validating the practical value of CFD technology in quantifying airway obstruction. Furthermore, snoring during sleep is fundamentally linked to AH-induced airway stenosis. In severe cases, recurrent partial or complete airway obstruction can occur, leading to obstructive apnea and systemic hypoxia. These findings clarify the core pathophysiological mechanism driving the deterioration of nighttime sleep quality.

Within the 3D models developed in this study, AH led to a significant reduction in the nasopharyngeal CSA forming a characteristic “hourglass-shaped” stenotic structure with anterior and posterior cavities. As inspiratory pressure increased (i.e., became more negative), overall airflow velocity increased in a gradient manner: velocity peaked at the narrowest point of UA, while decreasing relatively in other regions. In the AH model, airflow velocity upstream and downstream of the stenosis did not differ significantly from the normal model, whereas velocity in the stenotic segment was markedly higher than in the normal model. In the current study, the CFD-evaluated V_max was higher in the OSA group than in the control group, and the result is in accordance with other reports (3,37). This fluid dynamic characteristic aligns with Bernoulli’s principle (43): increased flow velocity accompanies decreased local static pressure, and airway narrowing due to AH further exacerbates the velocity-pressure imbalance within the stenotic segment. High-velocity airflow formed a jet when passing through the stenotic segment, inducing flow field disturbances and turbulence in the anterior and posteroinferior regions of the soft palate—consistent with flow field characteristics observed in related studies (42,44).

Notably, the local negative pressure gradient induced by sudden flow velocity increases in the stenotic segment promotes an inward collapse tendency of the airway wall. Moreover, the physiological reduction in pharyngeal muscle tone during sleep further exacerbates this negative pressure-driven dynamic airway collapse. Nonanatomic factors, such as soft tissue compliance and reduced intraluminal pressure in the airway during inspiration, may play crucial roles in airway obstruction and the pathogenesis of OSA (45). This abnormal fluid dynamic effect not only increases the respiratory workload of pediatric patients but also, with long-term persistence, may lead to clinical sequelae such as stunted growth due to excessive energy expenditure (46,47). The aforementioned fluid dynamic pathological mechanism provides a key pathophysiological basis for explaining common clinical symptoms like stridor and inspiratory dyspnea. We found that two obvious whirlpools are formed in the anterior upper part of the pediatric nasal cavity and in the oropharynx, which was caused by the sudden increase in the nasal CSA, resulting in local flow separation and counterflow. As the palatopharyngeal region lacks bony support and its soft tissues are more compliant, it tends to collapse under the large pressure difference. Our research supports literature reports (41,48). In addition, we achieved a relatively spacious nasopharyngeal cavity and low airflow velocity after AT, which have brought less turbulence, less soft palate vibrating, and less mouth breathing and snoring, consistent with literature reports (2).

Children with impaired airway ventilation often present with nasal congestion; even with compensatory mouth breathing, they may still experience hypoxia due to insufficient ventilation. In contrast, postoperatively, airway ventilation function improves significantly, and children’s oxygen supply is fundamentally enhanced—supported by the significant difference in volumetric flow rate within the flow field (Figure 11). In children, AH significantly reduces respiratory gas volumetric flow rates through mechanical obstruction and airway remodeling. This pathological change is the core mechanism underlying OSA (49,50). Clinical evidence indicates that AT can specifically relieve mechanical airway obstruction and effectively restore normal respiratory function. Postoperatively, with expansion of the luminal CSA at the stenotic site, stridor and dyspnea are significantly alleviated. The results of this study demonstrated that the dynamic characteristic of sudden increases in airflow velocity at the stenotic cross-section manifests clinically as stridor and exertional dyspnea. The consistency between these clinical observations and CFD analysis results further reinforces the clinical rationale for AT as the first-line treatment for AH. By means of 3D models obtained from CT scans coupled with CFD simulations, our study had shown the potential of CFD to predict clinical intervention outcomes through the quantification of flow, consistent with literature reports (41,51,52).

While existing studies have demonstrated a correlation between respiratory WSS and airway wall injury, research on the mechanisms underlying AH remains relatively limited (41,53-57). Children’s airways differ markedly from adults’ in anatomical structure, physiological function, and immune defense mechanisms; their mucosal epithelium is especially vulnerable to mechanical injury caused by abnormally elevated WSS (58). Our study demonstrated that under increased inspiratory negative pressure, WSS at the stenotic site in the AH model exceeded that in the normal model. This abnormal elevation in WSS can directly damage mucosal epithelial cells at the adenoid surface, impair the mucosal barrier, and further increase mucosal permeability—facilitating the invasion of allergens and pathogens and triggering chronic inflammatory responses. Prolonged persistent inflammatory stimuli can further promote AH, establishing a pathological cycle: “mechanical stenosis → shear stress injury → inflammatory response → tissue hyperplasia”. This cycle exacerbates airway structural remodeling and ventilatory dysfunction. Additionally, our study found that WSS near the Eustachian tube’s pharyngeal orifice was relatively elevated, which may explain the mechanism behind the high incidence of otitis media in AH patients, providing a basis for preventive interventions.

The size of minimal CSA (CSA_min) plays a pivotal role in the aerodynamic characteristics of the UA. The smaller the CSA_min, the greater the airflow velocity, pressure drop and WSS, the lower the minimum pressure, and the higher the overall airway resistance (8,59-61). The OSA severity is significantly negatively correlated with CSA_min (62), which is determined by the size of the adenoids (41). It can thus be concluded that smaller CSA_min is associated with a higher risk of OSA (8). Consistent with this, CFD analysis in our study showed that the differences in Vmax, minimum pressure, and maximum pressure (P_max) were consistently localized to the region of CSA_min. We suggest that clinicians use the UA CSA_min threshold as an initial screening tool for identifying patients at potential risk of OSA. This approach enables early detection of risk factors, and ensures that high-risk patients can be promptly referred to otolaryngology for further evaluation.

Based on CFD technology, this study developed individualized airway models of children with AH and performed numerical simulations. Through comparative analysis with normal models, it systematically explored the correlations between UA morphological characteristics, fluid dynamic parameters, and clinical symptoms. Our research supports the use of CFD as a viable, non-invasive method for evaluating airway patency and identifying children at risk of OSA. However, this research was limited by the complexity of adenoid anatomical structures, significant individual variability, and small sample size. Future research will be conducted to overcome these shortcomings. This novel diagnostic tool also needs future research to confirm the diagnostic performance and usefulness in the real-life situation.

Conclusions

This study demonstrates CT image-based CFD techniques can provide a clear visualization and an objective description of the UA flow field. The CFD could be a useful tool to differentiate subjects with and without OSA, and measure the effectiveness of treatment outcomes.

Acknowledgments

None.

Footnote

Data Sharing Statement: Available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2566/dss

Funding: This study was supported by the Fundamental Research Funds for the Central Universities (Shanghai Jiao Tong University ‘Jiao Tong Star’ Program Medical-Engineering Cross Research Fund) (grant No. YG2025ZD06), Seed Program Project for Medical New Technology Research and Transformation of the Shanghai Municipal Health Commission (No. 2024ZZ2036), and the Fundamental Research Funds for the Central Universities (No. YG2023ZD23).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://qims.amegroups.com/article/view/10.21037/qims-2025-1-2566/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of Shanghai Children’s Hospital, Shanghai Jiao Tong University (approval No. SHCH-IACUC-2022-XMSB-85). Informed consent was obtained from the patient’s parent or legal guardian.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Gottlieb DJ, Punjabi NM. Diagnosis and Management of Obstructive Sleep Apnea: A Review. JAMA 2020;323:1389-400. [Crossref] [PubMed]
2. Hu C, Han D, Zhou B, Zhang L, Li Y, Zang H, Li L. Investigation of resectability degree for adenoidal surgery in OSA children with the method of computational fluid dynamics. Acta Otolaryngol 2017;137:82-5. [Crossref] [PubMed]
3. Yang G, Liu A, Pan J, Zhao Y, Yu L, Sun L, Tong X, Li Y, Liu Y. Efficacy comparison between dentofacial orthopedic treatment and adenotonsillectomy on children with obstructive sleep apnea and skeletal class II malocclusion: a retrospective study. BMC Oral Health 2025;25:1325. [Crossref] [PubMed]
4. Gao D, Sun X, Yang Y, Yang J, Cheng L. Diagnostic value of CBCT in Chinese children with adenoid hypertrophy. Laryngoscope Investig Otolaryngol 2022;7:1308-14. [Crossref] [PubMed]
5. Iwasaki T, Sugiyama T, Yanagisawa-Minami A, Oku Y, Yokura A, Yamasaki Y. Effect of adenoids and tonsil tissue on pediatric obstructive sleep apnea severity determined by computational fluid dynamics. J Clin Sleep Med 2020;16:2021-8. [Crossref] [PubMed]
6. Calvo-Henriquez C, Branco AM, Lechien JR, Maria-Saibene A, DeMarchi MV, Valencia-Blanco B, Boronat-Catalá B, Rangel-Chávez J, Martin-Martin C. What is the relationship between the size of the adenoids and nasal obstruction? A systematic review. Int J Pediatr Otorhinolaryngol 2021;151:110895.
7. Hu B, Yin G, Fu S, Zhang B, Shang Y, Zhang Y, Ye J. The influence of mouth opening on pharyngeal pressure loss and its underlying mechanism: A computational fluid dynamic analysis. Front Bioeng Biotechnol 2022;10:1081465. [Crossref] [PubMed]
8. Wang C, Hu J, Qu Q, Ni X, Li Q. The effect of minimum cross-sectional area site and size on upper airway flow field characteristics: a pilot study. Oral Surg Oral Med Oral Pathol Oral Radiol 2026;141:538-47. [Crossref] [PubMed]
9. Shang Y, Dong J, Tian L, Inthavong K, Tu J. Detailed computational analysis of flow dynamics in an extended respiratory airway model. Clin Biomech (Bristol) 2019;61:105-11. [Crossref] [PubMed]
10. Faizal WM, Ghazali NNN, Khor CY, Badruddin IA, Zainon MZ, Yazid AA, Ibrahim NB, Razi RM. Computational fluid dynamics modelling of human upper airway: A review. Comput Methods Programs Biomed 2020;196:105627. [Crossref] [PubMed]
11. Zdanski C, Davis S, Hong Y, Miao D, Quammen C, Mitran S, Davis B, Niethammer M, Kimbell J, Pitkin E, Fine J, Fordham L, Vaughn B, Superfine R. Quantitative assessment of the upper airway in infants and children with subglottic stenosis. Laryngoscope 2016;126:1225-31. [Crossref] [PubMed]
12. Lin EL, Bock JM, Zdanski CJ, Kimbell JS, Garcia GJM. Relationship between degree of obstruction and airflow limitation in subglottic stenosis. Laryngoscope 2018;128:1551-7. [Crossref] [PubMed]
13. Crosby T, Adkins L, McWhorter A, Kunduk M, Dunham M. Computational fluid dynamics model of laryngotracheal stenosis and correlation to pulmonary function measures. Respir Physiol Neurobiol 2023;312:104037. [Crossref] [PubMed]
14. Mason EC, McGhee S, Zhao K, Chiang T, Matrka L. The Application of Computational Fluid Dynamics in the Evaluation of Laryngotracheal Pathology. Ann Otol Rhinol Laryngol 2019;128:453-9. [Crossref] [PubMed]
15. Cheng T, Carpenter D, Cohen S, Witsell D, Frank-Ito DO. Investigating the effects of laryngotracheal stenosis on upper airway aerodynamics. Laryngoscope 2018;128:E141-9. [Crossref] [PubMed]
16. Mimouni-Benabu O, Meister L, Giordano J, Fayoux P, Loundon N, Triglia JM, Nicollas R. A preliminary study of computer assisted evaluation of congenital tracheal stenosis: a new tool for surgical decision-making. Int J Pediatr Otorhinolaryngol 2012;76:1552-7. [Crossref] [PubMed]
17. Rahimi-Gorji M, Pourmehran O, Gorji-Bandpy M, Gorji TB. CFD simulation of airflow behavior and particle transport and deposition in different breathing conditions through the realistic model of human airways. Journal of Molecular Liquids 2015;209:121-33.
18. An K, Zhang F, Ouyang W, Pan X. Blood flow dynamics in the ascending aorta of patients with bicuspid aortic valve before and after transcatheter aortic valve replacement: a computational fluid dynamics study. BMC Cardiovasc Disord 2024;24:717. [Crossref] [PubMed]
19. Cao H, Li Y, Chen F, Peng Y, Liu Z, Zheng T, Chen M. Standardizing plaque impact on overall coronary hemodynamics using a binomial equation based on the equivalent energy loss. J Biomech 2025;179:112461. [Crossref] [PubMed]
20. Bhardwaj JS, Chanda S. Structural and hemodynamic analysis of Weaire-Phelan scaffolds made of Ti-alloy as bone replacement component: A preclinical investigation. PLoS One 2024;19:e0312880. [Crossref] [PubMed]
21. Sommerfeld M, Sgrott OL Jr, Taborda MA, Koullapis P, Bauer K, Kassinos S. Analysis of flow field and turbulence predictions in a lung model applying RANS and implications for particle deposition. Eur J Pharm Sci 2021;166:105959. [Crossref] [PubMed]
22. Ahookhosh K, Saidi M, Aminfar H, Mohammadpourfard M, Hamishehkar H, Yaqoubi S. Dry powder inhaler aerosol deposition in a model of tracheobronchial airways: Validating CFD predictions with in vitro data. Int J Pharm 2020;587:119599. [Crossref] [PubMed]
23. McDaniel LS, Poynot WJ, Gonthier KA, Dunham ME, Crosby ATW. Image-Based 3-Dimensional Characterization of Laryngotracheal Stenosis in Children. OTO Open 2018;2:2473974X17753583.
24. Chen J, Xu H, Li X. Von Mises stress peak (VMSP) and laryngomalacia severity score (LSS) are extremely useful in the selection of treatment for laryngomalacia. Eur Arch Otorhinolaryngol 2023;280:3287-93. [Crossref] [PubMed]
25. Xu H, Chen J, Pu S, Li X. Three-dimensional finite element modeling for evaluation of laryngomalacia severity in infants and children. J Int Med Res 2020;48:300060520926407. [Crossref] [PubMed]
26. Children’s otolaryngology committee of the pediatric branch of the Chinese medical doctor association, Pediatric Allergic Committee of Chinese Maternal and Child Health Association, Otolaryngology-Head and Neck Surgery Branch of Asia-Pacific Association of Medicine and Bio-Immunology. Expert consensus on clinical diagnosis and treatment management of adenoid hypertrophy in children. Chinese Journal of Practical Pediatrics 2025;40:89-95.
27. Zhang Z, Kleinstreuer C. Low-Reynolds-Number Turbulent Flows in Locally Constricted Conduits: A Comparison Study. AIAA Journal 2003; [Crossref]
28. Gianoni-Capenakas S, Rossi C, Templier L, Muwanguzi M, Gomes AC, Vich ML, Lange CF. Computational fluid dynamics analysis of the respiratory function of orthodontic patients. A scoping review. BMC Med Imaging 2026;26:163.
29. Desai R, Komperda J, Elnagar MH, Viana G, Galang-Boquiren MTS. Evaluation of upper airway characteristics in patients with and without sleep apnea using cone-beam computed tomography and computational fluid dynamics. Orthod Craniofac Res 2023;26.
30. Sokolovs-Karijs O, Brīvība M, Gudrā D, Saksis R, Zodāne AA, Rozenberga M, Bunka L, Frigotto A, Osīte J, Reinis A, Segliņa G, Krūmiņa A. The Clinical Role of the Adenoid Microbiome in the Development of Adenoid Hypertrophy and Otitis Media with Effusion. J Clin Med 2026;15:2187. [Crossref] [PubMed]
31. Zhang X, Li Y, Wu X, Lin M, Wang P, Zhao B, Liu Y. Deep learning-based automatic adenoid segmentation and a novel volume-based index for adenoid hypertrophy assessment. BMC Oral Health 2026;26:366. [Crossref] [PubMed]
32. Yang Q, Cui Y, Huang X, Liu J, Ma X, Gao GF, Pan H, Qin S. Disease-stage-specific immunometabolic remodeling in pediatric obstructive sleep apnea: a single-cell transcriptomic atlas of adenoid tissue. EMBO Mol Med 2026;18:2483-513. [Crossref] [PubMed]
33. Marcus CL, Brooks LJ, Draper KA, Gozal D, Halbower AC, Jones J, Schechter MS, Ward SD, Sheldon SH, Shiffman RN, Lehmann C, Spruyt K. Diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics 2012;130:e714-55. [Crossref] [PubMed]
34. Riley RW, Powell NB, Li KK, Troell RJ, Guilleminault C. Surgery and obstructive sleep apnea: long-term clinical outcomes. Otolaryngol Head Neck Surg 2000;122:415-21. [Crossref] [PubMed]
35. Duan H, Xia L, He W, Lin Y, Lu Z, Lan Q. Accuracy of lateral cephalogram for diagnosis of adenoid hypertrophy and posterior upper airway obstruction: A meta-analysis. Int J Pediatr Otorhinolaryngol 2019;119:1-9. [Crossref] [PubMed]
36. Quinlan CM, Otero H, Tapia IE. Upper airway visualization in pediatric obstructive sleep apnea. Paediatr Respir Rev 2019;32:48-54. [Crossref] [PubMed]
37. Yanagisawa-Minami A, Sugiyama T, Iwasaki T, Yamasaki Y. Primary site identification in children with obstructive sleep apnea by computational fluid dynamics analysis of the upper airway. J Clin Sleep Med 2020;16:431-9. [Crossref] [PubMed]
38. Hsu WC, Kang KT, Chen YJ, Weng WC, Lee PL, Hsiao HT. Computational fluid dynamics study in children with obstructive sleep apnea. Clin Otolaryngol 2024;49:109-16. [Crossref] [PubMed]
39. Cheng ES, Lai DK, Mao YJ, Lee TT, Lam WK, Cheung JC, Wong DW. Computational Biomechanics of Sleep: A Systematic Mapping Review. Bioengineering (Basel) 2023;10:917. [Crossref] [PubMed]
40. Wedel J, Steinmann P, Štrakl M, Hriberšek M, Cui Y, Ravnik J. Anatomy matters: The role of the subject-specific respiratory tract on aerosol deposition - A CFD study. Comput Methods Appl Mech Eng 2022;401:115372. [Crossref] [PubMed]
41. Hu Z, Dong J, Lou M, Zhang J, Ma R, Wang Y, Gong M, Wang B, Tong Z, Ren H, Zheng G, Zhang Y. Effect of different degrees of adenoid hypertrophy on pediatric upper airway aerodynamics: a computational fluid dynamics study. Biomech Model Mechanobiol 2023;22:1163-75. [Crossref] [PubMed]
42. Taherian S, Rahai H, Lopez S, Shin J, Jafari B. Evaluation of human obstructive sleep apnea using computational fluid dynamics. Commun Biol 2019;2:423. [Crossref] [PubMed]
43. Weese J, Lungu A, Peters J, Weber FM, Waechter-Stehle I, Hose DR. CFD- and Bernoulli-based pressure drop estimates: A comparison using patient anatomies from heart and aortic valve segmentation of CT images. Med Phys 2017;44:2281-92. [Crossref] [PubMed]
44. Mutuku JK, Chen WH. Flow Characterization in Healthy Airways and Airways with Chronic Obstructive Pulmonary Disease (COPD) during Different Inhalation Conditions. Aerosol Air Qual Res 2018;18:2680-94.
45. Fogel RB, Malhotra A, White DP. Sleep. 2: pathophysiology of obstructive sleep apnoea/hypopnoea syndrome. Thorax 2004;59:159-63.
46. Naseri A, Shaghaghian S, Abouali O, Ahmadi G. Numerical investigation of transient transport and deposition of microparticles under unsteady inspiratory flow in human upper airways. Respir Physiol Neurobiol 2017;244:56-72. [Crossref] [PubMed]
47. Yang MM, Higano NS, Gunatilaka CC, Hysinger EB, Amin RS, Woods JC, Bates AJ. Subglottic Stenosis Position Affects Work of Breathing. Laryngoscope 2021;131:E1220-6. [Crossref] [PubMed]
48. Guo Y, Su L, Cai H, Gao X, Xue H, Wu X. Analysis of the effect of adenoid size on upper airway flow in children based on computational fluid dynamics. Chinese Journal of Otorhinolaryngology-skull Base Surgery 2023;29:75-80.
49. Cengiz AB, Tansuker HD, Ozyilmaz C, Eroglu S, Ogreden S, Oktay MF. Role of Peak Nasal Inspiratory Flow for Measuring the Effectiveness of Surgery in Children with Adenoidal Hypertrophy. Cureus 2020;12:e12378. [Crossref] [PubMed]
50. Molero-Ramirez H, Tamae Kakazu M, Baroody F, Bhattacharjee R. Polysomnography Parameters Assessing Gas Exchange Best Predict Postoperative Respiratory Complications Following Adenotonsillectomy in Children With Severe OSA. J Clin Sleep Med 2019;15:1251-9. [Crossref] [PubMed]
51. Van Gaver H, Op de Beeck S, Dieltjens M, De Backer J, Verbraecken J, De Backer WA, Van de Heyning PH, Braem MJ, Vanderveken OM. Functional imaging improves patient selection for mandibular advancement device treatment outcome in sleep-disordered breathing: a prospective study. J Clin Sleep Med 2022;18:739-50. [Crossref] [PubMed]
52. Chen YF, Sheen MH, Chang HP, Tseng YC. Pediatric obstructive sleep apnea: Computational fluid dynamics analysis of upper airway. J Dent Sci 2022;17:589-91. [Crossref] [PubMed]
53. Bates A, Mcconnell K, Fleck R, Dumoulin C, Woods J, Amin R. 0759 Quantification of Neuromuscular Effort in Upper Airways of Pediatric Patients with Obstructive Sleep Apnea. Sleep 2019;42:A305.
54. Huh D, Fujioka H, Tung YC, Futai N, Paine R 3rd, Grotberg JB, Takayama S. Acoustically detectable cellular-level lung injury induced by fluid mechanical stresses in microfluidic airway systems. Proc Natl Acad Sci U S A 2007;104:18886-91. [Crossref] [PubMed]
55. Tarran R, Button B, Boucher RC. Regulation of normal and cystic fibrosis airway surface liquid volume by phasic shear stress. Annu Rev Physiol 2006;68:543-61. [Crossref] [PubMed]
56. Evans DJ, Green AS, Thomas NK. Wall shear stress distributions in a model of normal and constricted small airways. Proc Inst Mech Eng H 2014;228:362-70. [Crossref] [PubMed]
57. Kimbell JS. Fleischman Gm, Pitkin E, Davis SD, Zdanski C. Wall Shear Stresses And Airflow Patterns In The Upper Airway Of A Child With Subglottic Stenosis: Comparisons Before And After Airway Reconstruction With Normal Children. Am J Respir Crit Care Med 2013;187:A4584.
58. Tsega EG. Computational Fluid Dynamics Modeling of Respiratory Airflow in Tracheobronchial Airways of Infant, Child, and Adult. Comput Math Methods Med 2018;2018:9603451. [Crossref] [PubMed]
59. Wen XL, Wu BZ, Li Y, Yi B, Peng X. Analysis of the aerodynamic characteristics of the upper airway in obstructive sleep apnea patients. J Dent Sci 2024;19:329-37. [Crossref] [PubMed]
60. Taha YM, Abu El Sadat SM, Gaber RM, Farid MM. Ability of upper airway metrics to predict obstructive sleep apnea severity: a systematic review. Dentomaxillofac Radiol 2025;54:245-55. [Crossref] [PubMed]
61. Altheer C, Papageorgiou SN, Antonarakis GS, Papadopoulou AK. Do patients with different craniofacial patterns have differences in upper airway volume? A systematic review with network meta-analysis. Eur J Orthod 2024;46:cjae010.
62. Wang X, Chen H, Jia L, Xu X, Guo J. The relationship between three-dimensional craniofacial and upper airway anatomical variables and severity of obstructive sleep apnoea in adults. Eur J Orthod 2022;44:78-85. [Crossref] [PubMed]