Virtually Assisted Personalized Tracheostomy Tube Design in Pediatric Complex Airway Anomalies

Abstract

We sought to assess the feasibility of virtually assisted personalized tracheostomy tube (vapTT) implementation for patients with congenital airway anomalies (CAAs) and persistent tracheostomy tube (TT)–related respiratory failure at a tertiary pediatric hospital. Three patients (0–18 years) with CAAs and recurrent TT-related respiratory complications were managed with vapTT over 5 years. Patients underwent airway computed tomography acquisition with 3-dimensional reconstruction and TT virtual modeling for shape customization. Models were transferred to Bivona for fabrication based on industry-standard materials and processes. Clinical information and tracheoscopies assessing position, obstruction, and granulation were reviewed. Patients demonstrated resolution of visualized TT-related obstruction, granulation, or ulceration and de-escalation of respiratory support. Clinical events requiring urgent tracheoscopy decreased in all 3 patients. Sufficient relief of critical airway obstruction allowed progression of medical care and/or discharge. VapTTs are feasible for patients with CAA. This new frontier in personalized devices may serve uniquely challenging patient populations for whom standard treatments have failed.

Pediatric congenital airway anomalies (CAAs) are rare, challenging entities with significant morbidity and mortality. CAAs comprise anatomic variants that significantly deviate in morphology (ie, tracheal cartilaginous sleeve, severe airway tortuosity). CAAs are more prevalent in populations with craniofacial disorders or skeletal dysplasia, who often experience multilevel airway obstruction.¹

Tracheostomy tubes (TTs) among patients with CAA can result in TT-related complications. TTs are designed for patients with normal orientation of the tracheobronchial tree. Intrinsic tissue differences and airway tortuosity in CAAs can lead to tube malposition, obstruction, and chronic airway injury. Restriction of oxygenation and/or ventilation can be fatal if not recognized and treated expeditiously.²

Virtual 3-dimensional modeling can address challenging clinical problems.^3,4 We designed virtually assisted personalized TTs (vapTTs) for medically complex cases with CAAs. The impact on respiratory status and patient outcomes is discussed.

Methods

Patients treated with vapTTs were reviewed after institutional review board approval (Seattle Children’s Hospital, 2016–2020). Patients with CAAs experiencing recurrent acute hypoxic or hypercapnic respiratory failure due to ongoing TT obstruction following exhaustion of commercially available standard and customized options (defined as inability to resolve TT-related events despite iterative use of multiple TTs) were considered for vapTTs and included. Success or failure of each TT was assessed through serial tracheoscopy and intensive care unit surveillance. Patients tolerating industry-standard, conventionally customized TTs or cases prophylactically considered for vapTTs were excluded. VapTT success or failure was assessed by tracheoscopy and respiratory monitoring.

VapTT Design

Consensus on clinical candidacy was reached by multidisciplinary review (otolaryngology, radiology, pulmonology, critical care, respiratory therapy, and bioengineering). Caregivers provided verbal informed consent for vapTT design, manufacturing, and placement. Standard Bivona (Smith’s Medical) TT template files served as reference for personalization. High-resolution neck/chest computed tomography (CT) scans were obtained (0.5-mm slice thickness; Siemens Force scanner) with the TTs in situ to evaluate airway configuration and position when cannulated.

CT images were exported to Mimics 19 software (Materialise). Anatomic structures were modeled (lumen, bone, skin, esophagus, in-dwelling TTs). Clinical, endoscopic, and radiologic data were incorporated for accurate structure annotation (eg, to avoid granulation tissue on endoscopy being mislabeled as tracheal wall). Objects were exported to 3- matic software (Materialise) for modeling new TTs by using known dimensions and verifying optimal skin flange interface to ensure adequate trajectory and minimize outer dimension wall contact. Often, varying TT models were drawn (curvature, length, and diameter) to assess fit and ease of placement/removal. Models defined as optimal by the clinical team were returned to the DICOM imaging space and reviewed for overlap to the raw DICOM images. Detailed discussion with Bivona was critically helpful to ensure that functional patient details were incorporated (eg, mobility, postural bias, stoma site integrity). The patient’s body section encompassing the TTs (above stoma to past carina) and the model tubes were printed in a mixture of Agilus, Tissue Matrix, and Gel Matrix on a J750 Digital Anatomy Printer for final review. These models were used to demonstrate the plan and facilitate discussion with the patient’s medical team and family. Final in situ virtual product images were exported to standard tessellation language files with templates completed per the ordering work at Bivona. VapTTs were manufactured by Bivona using industry-quality standards and specifications. VapTT placement occurred in the intensive care unit or operating room with immediate and surveillance tracheoscopies to confirm position and monitor ulceration/granulation. Patients continued ongoing monitoring for resolution of events for which vapTT was indicated.

Results

Three patients underwent vapTT design, fabrication, and utilization. Case summaries, airway models, vapTT indication, and outcomes are presented (Table 1). Previously trialed TTs are described (Table 2). Comparison of airway models is presented (Figure 1). Obstructive clinical events concerning for TT obstruction requiring urgent tracheoscopy decreased in all patients after vapTT placement. All patients had resolution of TT-related airway obstruction and de-escalation of respiratory support following vapTT implementation (patient 1 in Figure 2). Sufficient relief of critical airway obstruction allowed progression of medical care and/or discharge.

Table 1.

Patient-Specific Airway Abnormalities, Personal Tracheostomy Tube Indications and Modifications, and Posttreatment Outcomes.

Characteristics	Patient 1	Patient 2	Patient 3
Congenital airway anomaly	Tracheobronchial stenosis, TCS, tracheal tortuosity	TCS	Thoracic cavity constriction causing vertical airway buckling and high ventilatory pressures. Progressive tracheal tortuosity with focal tracheomegaly
Primary diagnosis	Apert syndrome	Pfeiffer syndrome	Metatropic dysplasia
Indication for vapTTa	UAO; RLD; recurrent hypoxic, hypercapnic respiratory failure. Cartilage exposure, ulceration, and granulation	UAO; RLD; recurrent hypoxic, hypercapnic respiratory failure	UAO; RLD; recurrent hypoxic, hypercapnic respiratory failure. Cartilage exposure, granulation formation, and thoracic vertebral erosion and remodeling
Triplanar customizationsb	45° in the axial plane and 20° anterior angulation in the sagittal plane at 1 cm from the distal tip of the tube	22.5° anterior angulation in the sagittal plane 5 mm from the distal tip of the tube	22° lateral bend in the coronal plane 1 cm from the tip of the tube. Mirrored S-shaped bend, custom length, custom length and position of the Aire-Cuf.
Outcome	Resolution of obstructive granulation tissue and ulceration causing recurrent acute respiratory failure. Stabilization allowed discharge on home-appropriate ventilator after a 3-y hospitalization	Resolution of obstructive granulation tissue causing recurrent acute respiratory failure	Resolution of obstructive granulation tissue, ulceration, and hypercarbia. Stabilization allowed discharge on home-appropriate ventilator.

^aAll patients had chronic respiratory failure with long-term ventilatory requirements necessitating tracheostomy tube dependence.

^bTriplanar angulations define the point-defining angles for the custom tracheostomy tube angulations. All bends were contoured beyond the descriptions provided in a layered model of the airway to create a 3-dimensional product.

Table 2.

Standard and Conventionally Customized TTs Used Prior to Consideration of Virtually Assisted Personalized TTs.

Patient	Standard	Standard Flextend	Hyperflex Flextend custom length	Custom bend and/or length	Total
1	1	0	4	17	22
2	0	1	0	1	2
3	0	1	0	3	4

Abbreviation: TT, tracheostomy tube.

Figure 1.

Three-dimensional airway modeling for study patients: coronal and sagittal modeling of patients 1-3 (left to right) with vapTT in situ. Normal airway model (far left panel) for comparison. vapTT, virtually assisted personalized tracheostomy tube.

Figure 2.

Patient 1: tracheoscopy before (A-C) and after (D-E) vapTT placement: (A) Ulceration/granulation from tracheostomy tube–wall contact. (B) Midlesion granulation/ulceration. (C) Luminal narrowing (yellow outline). (D) Remucosalization. (E) Improved airway patency. Star, tracheostomy tube tip. vapTT, virtually assisted personalized tracheostomy tube.

Discussion

CAAs are complex anatomic variants where commercially available TTs can be insufficient for adequate oxygenation and ventilation or result in chronic airway trauma. Conventional TT customizations can alter the shaft length, diameter, connector, neck flange, cuff, or options (ie, hyperflex) but cannot accommodate tortuous anatomy.⁵ We describe 3 tracheostomy-dependent patients with CAAs and recurrent life-threatening obstructive events despite exhaustion of commercially available custom TTs. Patients offered vapTTs had severe craniofacial differences or skeletal dysplasia with multilevel airway obstruction and tortuosity of the tracheobronchial tree. VapTTs resolved these episodes and achieved sufficient relief of critical airway obstruction, permitting progression of medical care and in some instances discharge after prolonged hospitalizations (3 years). Despite being a small retrospective series, personalized medical engineering’s utility on these patients cannot be undervalued.

There are limitations to consider. Software limitations in visualizing certain tissue interfaces (eg, cartilage or granulation tissue) require endoscopic and clinical correlation. Another potential gap lies in the information used to model the vapTT. While CT provides outstanding anatomic information, it is a single view in time, taken in a position that may not be the child’s common habitus. Integrating other factors, such as the thickness of stoma padding desired, the child’s activity level, or default biases in positional range, may weigh heavily on the final design. Last, truly objective parameters are limited in demonstrating the profound clinical implications that vapTTs had on our patients. Further studies to explore the safety, efficacy, and long-term outcomes of these novel vapTTs will be instrumental in extending this work.