See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/230637763

A preliminary study of computer assisted evaluation of congenital tracheal

stenosis: A new tool for surgical decision-making

Article  in  International Journal of Pediatric Otorhinolaryngology · August 2012

DOI: 10.1016/j.ijporl.2012.07.009 · Source: PubMed

CITATIONS READS

20 485

7 authors, including:

Lionel Meister J. Giordano

French National Centre for Scientific Research Aix-Marseille Université
15 PUBLICATIONS   82 CITATIONS    32 PUBLICATIONS   278 CITATIONS   

SEE PROFILE SEE PROFILE

Natalie Loundon Jean-Michel Triglia

Hôpital Armand-Trousseau (Hôpitaux Universitaires Est Parisien) Assistance Publique Hôpitaux de Marseille
145 PUBLICATIONS   1,843 CITATIONS    199 PUBLICATIONS   3,727 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

ShockPoration: Human Cells Transfection induced by Shock Wave Interaction View project

Light scattering and characterization of nanoparticles View project

All content following this page was uploaded by J. Giordano on 22 February 2019.

The user has requested enhancement of the downloaded file.

 International Journal of Pediatric Otorhinolaryngology 76 (2012) 1552–1557

Contents lists available at SciVerse ScienceDirect

International Journal of Pediatric Otorhinolaryngology

journal homepage: www.elsevier.com/locate/ijporl

A preliminary study of computer assisted evaluation of congenital tracheal

stenosis: A new tool for surgical decision-making
Olivia Mimouni-Benabu a, Lionel Meister b, Jérôme Giordano b, Pierre Fayoux c, Natalie Loundon d,
Jean Michel Triglia a, Richard Nicollas a,b,*
a
Department of Pediatric Otolaryngology, Head and Neck Surgery, La Timone Children’s Hospital, Aix Marseille Université, Marseille, France
b
UMR 6595 IUSTI, Aix-Marseille Université, Marseille, France
c
ENT Surgery Department, Claude Hurriez Hospital, Lille, France
d
Department of Pediatric Otolaryngology, Head and Neck Surgery, Armand Trousseau Children’s Hospital, Paris, France

A R T I C L E I N F O A B S T R A C T

Article history: Objective: Congenital tracheal stenosis is a rare but severe condition with tracheal narrowing. There is no
Received 20 December 2011 absolute correlation between luminal diameter and prognosis, and therapeutic decisions are difficult for
Received in revised form 4 July 2012 intermediate cases. The aim of this study was to develop a dynamic model of the ventilatory
Accepted 5 July 2012
consequences of congenital tracheal stenosis using computational fluid dynamics.
Available online 5 August 2012
Methods: In 8 children with congenital tracheal stenosis and 1 healthy child, 3-dimensional geometries
of the trachea were constructed with computed tomography images and specialized software (ITK-
Keywords:
SNAP). Airflow simulations were performed for each geometry using 2 physiologic inhalation flow rates
Trachea
Segmentation
under steady and laminar flow conditions. Flow velocity, static and total airway pressure, and pressure
Computational fluid dynamics drop across the entire trachea were determined.
Airflow Results: In the patients with congenital tracheal stenosis, the pressure drop from the tracheal inlet to
outlet, at flow rate 3 L/min, ranged from 14 to 430 Pa; the pressure drop at flow rate 7.3 L/min ranged
from 60 to 1825 Pa. The pressure drop enabled a classification based on the severity of stenosis. The
classification based on pressure drop was retrospectively consistent with the classification based on
clinical data from the patients.
Conclusions: Simulations with computational fluid dynamics may provide an objective method to
evaluate the severity of the symptoms in patients with congenital tracheal stenosis and may help guide
treatment.
ß 2012 Elsevier Ireland Ltd. All rights reserved.

1. Introduction subjective, based on respiratory status, endoscopy, tomodensito-

Congenital tracheal stenosis is a rare but potentially life-
threatening condition, with narrowing of the tracheal lumen
compared with the normal diameter. Although the pathogenesis of
congenital tracheal stenosis is not fully understood, the disease
commonly is caused by the presence of complete tracheal cartilage
rings, which create a rigid posterior tracheal wall. Stenosis may
involve part of the trachea (commonly the lower portion) or the
entire trachea and the main bronchi. It may cause respiratory
compromise ranging from mild dyspnea to respiratory distress.
Currently, the therapeutic decision about surgical treatment is
* Corresponding author at: Department of Pediatric Otolaryngology, Head and
Neck Surgery, La Timone Children’s Hospital, Aix Marseille Université, 264 rue Saint
Pierre, 13385 Marseille cedex 5, France. Tel.: +33 491386709; fax: +33 491384318.
E-mail address: richard.nicollas@ap-hm.fr (R. Nicollas).
metry, and the experience of the surgeon.
Patients with congenital tracheal stenosis may have diverse
symptoms and treatment. Patients may be classified in 3 groups,
based on clinical presentation:
 Severe: Early symptoms occur in the first hour of life. Emergency
surgery is required because of the severity of respiratory
compromise. Frequency of mortality with or without surgery
is almost 100% [1] or slightly better [2].
 Mild to minimally symptomatic: The diagnosis is made during
childhood or adolescence. Treatment may be non-operative and
individualized, depending on the patient’s respiratory difficul-
ties.
 Intermediate: Respiratory symptoms occur during the first year of
life. This group presents a challenge for otorhinolaryngologists.
When the clinical status permits, delaying surgery until after the
first year of life has been recommended because non-operative
treatment decreases the frequency of mortality from 73% to 16%

0165-5876/$ – see front matter ß 2012 Elsevier Ireland Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ijporl.2012.07.009
 O. Mimouni-Benabu et al. / International Journal of Pediatric Otorhinolaryngology 76 (2012) 1552–1557
[1]. Furthermore, the symptoms of congenital tracheal stenosis
may improve with increased age, and a normal tracheal lumen
diameter may be reached by age 9 years [3]. However, a minor
viral infection may drastically alter the patient’s outcome, cause
acute respiratory distress, and increase the frequency of
mortality by 20% [1].
An objective evaluation of symptom severity in patients with
congenital tracheal stenosis could assist the surgeon’s choice of the
most suitable treatment.
Analysis with computational fluid dynamics has been used in
physiologic airflow studies of the effect of the cartilage ring on
tracheal airflow [4,5], pollution dispersion [6], drug particle
deposition [7], flow rate distribution in the human lung [8], and
the effect of a turbulent laryngeal jet on airflow in the human intra-
thoracic airway. Recent advances in imaging, including spiral
computed tomography (CT) scanning with multiple detectors, and
technological advances in image processing, have enabled the use
of computational fluid dynamics for creating realistic airway
models.
The purpose of this study was to perform computer
simulations of models of the narrowed trachea of infants to
characterize the ventilatory consequences of congenital tracheal
stenosis, based on the pressure drop from the tracheal inlet to the
main bronchi, and to compare these data with retrospective
information from clinical charts. We hypothesized that
the pressure drop from the tracheal inlet to outlet may be
useful in clinical decision-making about the treatment of
congenital tracheal stenosis.
2. Patients and methods
2.1. Patients
Patients included 9 infants (age range, 1–48 months) who had
mild to severe congenital tracheal stenosis (8 patients) or normal
trachea (1 patient) and who were treated from 2006 to 2009 in the
otorhinolaryngology departments of Claude Hurriez Hospital (Lille,
France) and La Timone Children’s Hospital (Marseille, France).
Medical records were retrospectively evaluated for demographic
data, comorbid diagnoses, and outcomes. Congenital tracheal
stenosis was classified [1] as type 1 (generalized hypoplasia, 1
patient), type 2 (funnel-type stenosis, 2 patients), and type 3
(segmental stenosis, 5 patients).
Fig. 1. Computed tomography scans of 8 patients who had congenital tracheal stenosis, included in the present study.
1553
Pre-treatment computed tomography (CT) scans of the patients
were reviewed retrospectively (Figs. 1 and 2). These scans were
made with 0.25 mm  0.25 mm resolution and slice thickness
from 600 mm to 1.5 mm (Siemens 64-detector helical CT scanner,
Siemens).
Surgery for correction of congenital tracheal stenosis was done
in 4 (50%) patients. There were 4 patients (50%) who had associated
anomalies including cardiovascular anomalies, pulmonary agene-
sis, gastrointestinal, renal, and skeletal problems. Death occurred
in 4 patients (50%). A clinical classification was defined, based on
the clinical severity of symptoms and patient outcomes. Patients
were classified into 3 groups:
 Group 1 (mild congenital tracheal stenosis): asymptomatic or
mild respiratory symptoms, non-operative treatment possible
(patients A, B, and C).
 Group 2 (moderate congenital tracheal stenosis): diagnosis at
age 6 months, no surgical treatment was done for tracheal
stenosis but the patient died of postoperative complications after
cardiac surgery (patient D).
 Group 3 (severe congenital tracheal stenosis): diagnosis in the
neonatal period, and death occurred in the first month of life
without or after tracheal surgery (patients E, F, G, and H) (Table
1).
Parameters of tracheal geometry were determined from the CT
scans, including stenosis length, total length of trachea (from x to
y), narrowest cross-sectional area of stenosis, and normal tracheal
diameter. From these data, 2 ratios were calculated to describe
stenosis geometry: the length ratio (rl) = stenosis length/total
length of trachea; and the diameter ratio (rd) = narrowest cross-
sectional area of stenosis/normal tracheal diameter.
In addition to the clinical classification, a CT classification of the
patients was done by a pediatric radiologist and an otorhinolaryn-
gologist, based on the narrowness of the trachea and the extent of
the stenosis.
2.2. Computational fluid dynamics
Computational fluid dynamics allowed the numerical simula-
tion of airflows by solving the Navier–Stokes equations, which
govern general fluid mechanics. A segregated flow model with a
second-order convective scheme was used for the resolution of the
Navier–Stokes equations. The computations included 3 major
1554 O. Mimouni-Benabu et al. / International Journal of Pediatric Otorhinolaryngology 76 (2012) 1552–1557

Fig. 2. Congenital tracheal stenosis reconstructions in 3 patients with segmental stenosis (left and middle) and funnel-type stenosis (right).

steps: (1) constructing a geometrical model from CT data; (2) the DP classification was defined as type 1 (DP < 260 Pa), type 2
discretization of the computational domain into computational (DP  260 Pa), and type 3 (DP > 1400 Pa).
cells, resulting in the generation of a computational mesh; and (3)
solving the conservation equations on this mesh until a 2.3. Construction of trachea models
numerically converged solution was obtained. In the computa-
tions, 2 age-adapted flow rates were used because of the range of Specialized software (ITK-SNAP, University of Pennsylvania,
patient age. The pressure drop from the tracheal inlet (first ring) to Philadelphia, PA) [9] was used to obtain 3-dimensional images of
the main bronchial outlets (DP) was defined by the equation the trachea from the CT scans (Fig. 3) [9]. The segmented
DP = Pinlet (Poutlet1 + Poutlet2)/2, where Pinlet was the pressure at geometries were exported as Standard Tessellation Language
the first ring, Poutlet1 was the pressure at the left main bronchus and (STL) files into the computational fluid dynamics code with
Poutlet2 was the pressure at the right main bronchus. software (STAR-CCM+, CDadapco, City, State) that was used to
A congenital tracheal stenosis model classification was created, simulate airflow; this software was used to generate volumetric
based on DP. The flow rate used in simulations did not modify this computational meshes for all patients. Depending on the size of
classification, confirming that congenital tracheal stenosis com- tracheas, 500,000–1,300,000 cell meshes were obtained. For all
parisons may be done with any flow rate. For flow rate of 7.3 L/min, cases, a mesh base size of 0.2 mm was selected; this size was
determined by performing a convergence study with a prescribed
Table 1 tolerance <0.5%. To compute boundary layers accurately, mesh
Effect of patient and flow rate on pressure drop between tracheal inlet and main density was increased near the tracheal walls.
bronchial outlets in patients with congenital tracheal stenosis. To enable physiologic comparisons, simulations were comput-
Patient Age in months Clinical severitya DP (Pa)b ed during the inhalation phase of breathing, with 2 different flow
rates corresponding to a newborn (3 L/min) and a 4-year-old child
Flow rate (L/min)
3 7.3
(7.3 L/min) [7,10]. The tracheal walls were assumed to be rigid, and
a no-slip condition was imposed for all tracheas. The simulations
Healthy 40 None 1.2 4
were doc temperature (303.158 K), ambient atmospheric pressure
A 5 Mild 14 60
B 24 Mild 50 250 (101.325 kPa), normal air density (rair = 1.161 kg/m3), and normal
C 24 Mild 58 260 kinematic viscosity (nair = 1.8  10 5 m2/s). The pressure curves
were obtained by computing mean total pressure in 100 vertices
D 36 Moderate 106 484
along 100 streamlines seeded from the inlet.
E 48 Severe 379 1486 The Reynolds number is a dimensionless number used to
F 24 Severe 333 1591
G 10 Severe 430 1825
compare the effects of inertial forces to viscous forces for fluid
H 14 Severe 412 1825 flows. The Reynolds number gives information about the turbu-
a lence of fluid flow, which is important to determine whether
Clinical severity of congenital tracheal stenosis: group 1, mild; group 2,
moderate; group 3, severe. turbulence models are necessary to accurately simulate flow. In
b
Pressure drop between tracheal inlet and main bronchial outlets: DP = turbulent flows, Reynolds number is larger than 3000. In adults,
Pinlet (Poutlet1 + Poutlet2)/2. tracheal air flow may be considered turbulent because of the effect
 O. Mimouni-Benabu et al. / International Journal of Pediatric Otorhinolaryngology 76 (2012) 1552–1557 1555

Fig. 3. Screenshot of software (ITK-SNAP) used to evaluate congenital tracheal stenosis. The CT scans and resulting segmentation were observed simultaneously.

of the laryngeal jet [4], and turbulent flow models have been used
in most airflow studies [4–6,11].
A k– turbulent model was compared with a laminar model.
3. Results
Streamlines colored according to velocity showed that velocity
was uniform along the trachea in the normal patient model;
however, in pathologic models of the patients with congenital
tracheal stenosis, velocity increased with decreased tracheal
diameter, with maximum velocity at the sites of tightest stenosis
(Fig. 4). For pathologic cases, vortices were generated distal to the
stenosis (Fig. 4). In most pathologic cases, reverse flow occurred
downstream to the stenosis because of sudden airway expansion;
this was not observed in the healthy model or case A (in which the
stenosis was progressive).
In all pathologic cases, total pressure decrease was observed
from the trachea inlet to bronchial outlets (Fig. 5). Pressure profiles
were different for the different cases, depending on the stenosis
geometry. However, a pressure drop was present in all cases, and
the slope of the pressure drop correlated with stenosis severity. In
the healthy patient, the pressure drop from inlet to outlet (flow

Fig. 4. Streamlines with velocity color map for a normal (left) and a pathologic (right) model of congenital tracheal stenosis (flow rate, 7.3 L/min).
1556 O. Mimouni-Benabu et al. / International Journal of Pediatric Otorhinolaryngology 76 (2012) 1552–1557
Fig. 5. Distribution of total pressure between tracheal inlet and bronchial outlets at both flow rates studied. The pressure drop for a normal model was 4 Pa (left), and pressure
drop for a pathologic model was 1591 Pa (right).
rate 7.3 L/min) was linear (4 Pa); in the pathologic cases, the
pressure drop (flow rate 7.3 L/min) ranged from 60 to 1825 Pa
(Table 1).
Patients with similar types of congenital tracheal stenosis were
compared in pairs to identify geometric attributes affecting the
severity of congenital tracheal stenosis. Comparison of patients D
(larger length and diameter ratios, smaller pressure drop) and G
(smaller length and diameter ratios, larger pressure drop)
suggested that the cross-sectional area had more effect on
pressure drop than the stenosis length (Table 1). Comparison of
the stenosis in patient C (gradual stenosis) with patient E (sudden
change in diameter) suggested that diameter variation had a major
effect on pressure drop (Table 1).
The order of severity of congenital tracheal stenosis from mild
to severe, based on clinical information (healthy, A, B, C, D, E, F, G,
H) was similar to the order of severity based on computation of DP
(healthy, A, B, C, D, E, F, H, G), except for cases G and H (Table 1). The
order of severity of congenital tracheal stenosis from mild to severe
based on CT classification of patients (healthy, G, B, A, F, C, E, D, H)
differed from the clinical or computational classifications, espe-
cially for patients G and D; patient G had been classified on CT in
the mild group but required surgical repair in his first month of life,
and patient D was classified on CT in the severe group but he did
not undergo any tracheal surgery.
A k– turbulent model did not give any significantly different
results than a laminar model. In patient E, the pressure drop was
1456 Pa with the k– turbulent model and 1486 Pa using a
laminar model. In the infants, calculation of the Reynolds number
showed that airflow was laminar because of the smaller geometry
and lower velocity than adults; for the 9 studied cases, computed
Reynolds numbers ranged from 650 to 1500.
4. Discussion
This study confirmed that a realistic description of airflow in the
narrowed trachea of infants with congenital tracheal stenosis can
be obtained using CT images. Computational fluid dynamics
enabled comparison of different patients with congenital tracheal
stenosis and increased efficiency in the evaluation of both imaging
and clinical history. There was a strong association between the
pressure drop and the clinical outcome, with higher pressure drop
associated with worse clinical outcome because of the higher
degree of respiratory resistance and effort in congenital tracheal
stenosis. A reproducible classification based on objective measures
of airflow and pressure drop was consistent with clinical outcome.
The reverse flow observed downstream to tracheal stenosis is
similar to that described in the nasal valve region, in which a
recirculation zone behind the nasal valve may be observed
because of a sudden increase in diameter [12]. This also is
consistent with flow observations and classifications found in
industrial pipes [13].
The similarity of severity classification based on clinical and
pressure drop information shows the utility of objective flow data
in the management of the congenital tracheal stenosis, especially
because the CT classification did not correctly predict outcome in 2
(25%) patients with congenital tracheal stenosis.
Static pressure measurement in the human model upper airway
using a k– turbulent model was closer to experimental results
than other models [14]. This turbulence model has been used in
several computational fluid dynamics studies [5–7]. Therefore, it
was used in the present study to verify the relevance of using a
laminar model. The similarity of results between turbulent and
laminar models is consistent with laminar flow behavior in the
infant trachea.
Limitations of the present study include the use of CT data from
several pediatric centers, necessary because congenital tracheal
stenosis is a rare problem. Different centers may use different
imaging protocols, especially slice thickness, and slice thickness
>0.6 mm is not accurate for tracheal reconstruction. In addition,
image segmentation with the software (ITK-SNAP) required
human intervention from a specialist such as an otorhinolaryn-
gologist or radiologist to build the model as close to anatomic
reality as possible, and automation was not feasible. Furthermore,
the variation of patient age from 1 to 48 months caused a wide
range of geometric parameters such as tracheal length (range, 30–
55 mm) and tracheal section (range, 4.5–9 mm), potentially
confounding comparisons between different patients.
In contrast with subglottic stenosis, congenital tracheal stenosis
includes a wide spectrum of lesions, and currently available
classifications do not enable staging of all cases, guiding surgical
treatment, or sorting patients by severity [1,2]. These classifica-
tions are based on functional symptoms (Anton-Pacheco), tracheal
shape (Cantrell and Guild), or stenosis extension and associated co-
morbidities (Hoffer) [15]. However, the present work may allow an
objective criterion to classify patients with congenital tracheal
stenosis. The present classification, based on pressure drop along
the trachea, was retrospectively consistent with classifications
based on clinical outcomes.
Future work may include a determination of a safe pressure
drop threshold for non-operative and surgical treatment. With a
 O. Mimouni-Benabu et al. / International Journal of Pediatric Otorhinolaryngology 76 (2012) 1552–1557
low enough pressure drop, tracheal growth may reliably predict
patient recovery, avoiding surgical correction of congenital
tracheal stenosis that may be associated with operative mortality.
Performing computational fluid dynamics simulations on more
patients with congenital tracheal stenosis will be necessary, and
this may require a multicenter study because of the rarity of this
condition.
Choice of sections may affect the reliability of computational
fluid dynamics for the classification of congenital tracheal stenosis.
The sections used thus far were obtained from 3-dimensional
segmentation of the trachea based on CT scans, and all images were
oriented in the cross-sectional planes, corresponding to the axial
direction in the scanner during image acquisition. However, the
trachea may not be oriented in the axial direction, and cross-
sectional planes may not correspond to perpendicular planes of the
trachea. Therefore, it would be a major improvement to define the
tracheal sections in the direction of the longitudinal axis of
the trachea. We are currently working on extracting the
geometrical cross-section along the trachea, using the longitudinal
axis of the trachea. Extracting a plane section normal to this line
may permit the computation of the true geometric cross-sectional
area. The morphologic description of the trachea, combined with
the numeric investigation similar to the present analysis, may
enable the development of a simplified model to investigate
congenital tracheal stenosis.
Simulation with computational fluid dynamics offered an
objective method for evaluation of tracheal airflow in congenital
tracheal stenosis, and may become an additional clinical tool for
classification and treatment of this condition. The present data
were computed as steady simulations, and it may be useful to
compute unsteady simulations, but it may be difficult to obtain
time-dependent data about pressure or mass flow in infants.
Nevertheless, these data may serve as boundary conditions for
unsteady simulations. Furthermore, it may be useful to study the
effect of inspiration and expiration in patients with congenital
tracheal stenosis. Experimental physical models may be developed
using the STL files and additive manufacturing, and experimental
measurements may be compared with the present calculated
results.
View publication stats
1557
In summary, despite the wide geometric range of the present
models, the present classification based on computed pressure
drop along the trachea was retrospectively consistent with the
clinical history of 8 patients with congenital tracheal stenosis.
Future development of a pressure drop threshold, over which
breathing may be too laborious for infant respiratory capacity, may
help surgeons make therapeutic decisions based on objective
measures of tracheal airflow and may improve clinical outcomes.
References
[1] P. Herrera, C. Caldarone, V. Forte, P. Campisi, H. Holtby, P. Chait, et al., The current
state of congenital tracheal stenosis, Pediatr. Surg. Int. 23 (2007) 1033–1044.
[2] J.L. Antón-Pacheco, I. Cano, J. Comas, L. Galletti, L. Polo, A. Garcı́a, et al., Manage-
ment of congenital tracheal stenosis in infancy, Eur. J. Cardiothorac. Surg. 29
(2006) 991–996.
[3] W. Cheng, D.E. Manson, V. Forte, S.H. Ein, I. MacLusky, B.C. Papsin, et al., The role of
conservative management in congenital tracheal stenosis: an evidence-based
long-term follow-up study, J. Pediatr. Surg. 41 (2006) 1203–1207.
[4] C.L. Lin, M.H. Tawhai, G. McLennan, E.A. Hoffman, Characteristics of the turbulent
laryngeal jet and its effect on airflow in the human intra-thoracic airways, Respir.
Physiol. Neurobiol. 157 (2007) 295–309.
[5] H.Y. Luo, Y. Liu, Modeling the bifurcating flow in a CT-scanned human lung airway,
J. Biomech. 41 (2008) 2681–2688.
[6] J. Russo, R. Robinson, M.J. Oldham, Effects of cartilage rings on airflow and particle
deposition in the trachea and main bronchi, Med. Eng. Phys. 30 (2008) 581–589.
[7] P.W. Longest, S. Vinchurkar, Inertial deposition of aerosols in bifurcating models
during steady expiratory flow, J. Aerosol Sci. 40 (2009) 370–378.
[8] T. Gemci, V. Ponyavin, Y. Chen, H. Chen, R. Collins, Computational model of airflow
in upper 17 generations of human respiratory tract, J. Biomech. 41 (2008)
2047–2054.
[9] P.A. Yushkevich, J. Piven, H.C. Hazlett, R.G. Smith, S. Ho, J.C. Gee, et al., User-guided
3D active contour segmentation of anatomical structures: significantly improved
efficiency and reliability, Neuroimage 31 (2006) 1116–1128.
[10] J.P. Goldsmith, E.H. Karotkin, Assisted Ventilation of the Neonate, 2nd ed.,
Saunders, Philadelphia, 1988.
[11] M. Mihaescu, E. Gutmark, S. Murugappan, R. Elluru, A. Cohen, J.P. Willging,
Modeling flow in a compromised pediatric airway breathing air and heliox,
Laryngoscope 119 (2009) 145–151.
[12] L.M. Hopkins, J.T. Kelly, A.S. Wexler, A.K. Prasad, Particle image velocimetry
measurements in complex geometries, Exp. Fluids 29 (2000) 91–95.
[13] I.E. Idel’cik, Mémento des Pertes de Charges, 3rd ed., Eyrolles, 1986.
[14] G. Mylavarapu, S. Murugappan, M. Mihaescu, M. Kalra, S. Khosla, E. Gutmark,
Validation of computational fluid dynamics methodology used for human upper
airway flow simulations, J. Biomech. 42 (2009) 1553–1559.
[15] M.E. Hoffer, L.W. Tom, R.F. Wetmore, S.D. Handler, W.P. Potsic, Congenital
tracheal stenosis. The otolaryngologist’s perspective, Arch. Otolaryngol. Head
Neck Surg. 120 (1994) 449–453.