Magnetic resonance imaging was a viable and non‐intrusive method for diagnosing tracheomalacia in children

Abbreviations

CT

computed tomography

MRI

magnetic resonance imaging

Tracheomalacia is a narrowing of the tracheal lumen, which occurs when the trachea is abnormally weak and the anterior and posterior walls appose. It is a dynamic phenomenon that can affect a segment or the entire tracheal wall. A brassy or barking cough is the most common clinical characteristic. ¹

There are no established management algorithms for diagnosing tracheomalacia and flexible bronchoscopy is frequently used. ¹ Radiological, non‐invasive, methods include computed tomography (CT) and magnetic resonance imaging (MRI). MRI studies are radiation free and deliver static and dynamic scans, with comparable precision to CT. They can also be conducted during free breathing, making them suitable for paediatric patients who may have difficulty following breathing instructions. ² , ³ There is no universally acknowledged classification system for severity, but diagnosing tracheomalacia requires a 50% reduction in the cross‐sectional area of the trachea during expiration. ¹ Drugs, such as beta‐2 agonists, ipratropium and mucoactive agents, have not proved effective.

The lack of data highlights the need to clarify the effectiveness and limitations of current diagnostic methods and identify a gold standard technique for children. Our aim was to evaluate the effectiveness of MRI studies and compare them with CT scans and flexible bronchoscopy.

This retrospective study was carried out at the University General Hospital of Athens from January 1, 2020 to July 31, 2023. It focused on children with suspected tracheomalacia, due to recurrent episodes of a brassy cough.

All patients underwent CT and MRI scans during the same visit, regardless of the results, and no anaesthesia was required. Multidetector CT scanning was performed on an Aquilion ONE/PRISM Edition 320‐slice scanner (Canon Medical Systems Corporation, Tochigi, Japan). The scanning protocol comprised dynamic paired end‐inspiratory and forced‐expiratory scans with the helical acquisition technique, 1 mm collimation and no interslice interval. Participants were asked to inhale deeply before the procedure and maintain it during the end‐inspiratory phase of scanning. They were then asked to take another deep breath, exhale completely and maintain this position during the end‐expiratory phase of scanning. Scanning was performed with low radiation dose parameters and deep learning reconstruction algorithm technology. The multidetector CT diagnosis was based on the percentage reduction of the cross‐sectional area ratio of the trachea during end‐expiration and end‐inspiration.

The MRI scans used a 1.5T Philips Ingenia MRI scanner (Philips Healthcare, Amsterdam, the Netherlands) and comprised two primary sequence types. First, static sequences used single‐shot turbo spin‐echo T2‐weighted images in transverse and sagittal planes during end‐inspiratory and end‐expiratory phases. Second, ciné sequences used balanced turbo field‐echo sequences in transverse and sagittal planes during end‐inspiration, forced expiration and free breathing in more than one respiratory cycle. The diagnosis was also based on the estimated cross‐sectional area ratio. The CT and MRI studies were independently evaluated by two paediatric radiologists (EA and SP), with 30 and 7 years of experience in paediatric imaging, who reached a consensus.

A subset of 13 patients underwent flexible bronchoscopy under deep sedation, after the radiological examinations, primarily because they had a chronic wet cough or the findings could influence their clinical management. One author (KD) performed all these procedures and evaluated the degree of tracheomalacia. He was aware of the radiological findings, including the presence or absence of tracheomalacia, but not the estimated degree of tracheomalacia.

The research protocol was approved by the hospital's ethics committee (number 1819037872/5.7.19Λ.) and the parents or legal guardians provided written consent for the research data to be used.

Comparisons were performed with Fisher's exact test, the t‐test and repeated measures analysis of variance. The Wald test was used for the post hoc estimations after the analysis of variance.

We studied 24 children (16 boys) at a median age of 11 (range: 5–17) years. All underwent CT and MRI scans and 13 also underwent flexible bronchoscopy.

The means and standard deviations for the expiratory reduction in the cross‐sectional luminal area of the trachea in the 13 patients who underwent all three procedures were 0.53 (0.22) for the MRI scans, 0.46 (0.24) for the CT scans and 0.38 (0.33) for the flexible bronchoscopy (p = 0.002). The post hoc tests showed significant differences between the MRI and CT scans (p = 0.008) and the MRI scans and flexible bronchoscopy (p < 0.001).

The means and standard deviations of the cross‐sectional area ratio during end‐expiration and end‐inspiration were 0.56 (0.03) for the MRI scans and 0.40 (0.03) for the CT scans (p < 0.001).

MRI scans detected the diagnostic threshold of a 50.0% expiratory reduction for tracheomalacia in 21/24 cases (87.5%). Flexible bronchoscopy did this in 8/13 cases (61.5%) and CT in 8/24 cases (33.3%).

MRI scans detect tracheomalacia more frequently and were more likely to accurately estimate the degree. These non‐invasive scans have been increasingly recognised as valuable for assessing diseases of the larger airways. There is no radiation and minimal patient compliance means they can even diagnose tracheomalacia in non‐sedated neonates. ³ The results of single respiratory‐phase CT scanning may be misleading, as tracheomalacia is a dynamic phenomenon. Mean radiation during CT scans is extremely low, but there is still some exposure. ⁴ Flexible bronchoscopy is invasive, requires anaesthesia and short hospitalisation and the results remain largely subjective, as there is no standardised criteria. Its assessment of airway dynamics during uncontrolled breathing may lead to misdiagnosis, particularly in non‐severe tracheomalacia. In addition, anaesthesia can significantly alter airway morphology, further complicating diagnostic accuracy. ⁵

An MRI scan should be the primary diagnostic tool for assessing tracheomalacia in children whose only symptom is a recurrent brassy or barking cough. However, a more comprehensive diagnostic approach should be used if a chronic wet cough and/or sputum suggests a chronic endobronchial infection. This should include a CT scan and flexible bronchoscopy, as an MRI scan cannot detect the high likelihood of coexisting lower respiratory tract pathology.

The study limitations included its retrospective nature, small cohort size and only performing flexible bronchoscopy on 13/24 children. In addition, the CT technique used was susceptible to patient cooperation, unlike newer CT techniques like 4D‐CT scans. The physician performing the flexible bronchoscopy was also aware of the radiological findings.

Further research on diagnosing tracheomalacia requires larger cross‐sectional or longitudinal observational studies, to strengthen the data, enhance its clinical impact and provide more comprehensive insights.

The lack of a gold standard and standardised diagnostic criteria are major obstacles in accurately assessing diagnostic methods for tracheomalacia in children. However, MRI scans provide an attractive option, due to their advantages and superior performance.

AUTHOR CONTRIBUTIONS

Spyridon Prountzos: Investigation; writing – original draft; project administration; conceptualization; formal analysis. Konstantinos Douros: Investigation; writing – review and editing; conceptualization; methodology; software; data curation; validation. Dafni Moriki: Visualization; formal analysis; project administration. Argyro Mazioti: Visualization; formal analysis. Vasiliki Papaevangelou: Resources. Nikolaos Kelekis: Resources. Efthymia Alexopoulou: Conceptualization; visualization; funding acquisition; supervision; validation; methodology.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.