Computerized Tomographic Assessment for Phenotyping Asthma

Structural changes in the airways and parenchyma contribute to airflow obstruction in various lung diseases including asthma, chronic obstructive lung disease (COPD), bronchiectasis and cystic fibrosis. Disorders are virtually always distributed heterogeneously in the lungs. The standard evaluation of these structural changes has been based on histopathology but requires surgical or autopsy samples of lung tissues and cross-sectional design. Computed tomography (CT) has been regarded as an excellent clinical tool to assess the morphological change of the lung with various diseases in high resolution. CT can provide much information on airway and parenchymal changes to understand the pathogenesis of diseases, assess different contributions of disease components to the loss of lung function, and assess changes over time.

Asthma is a chronic inflammatory lung disorder. It involves the small and medium-sized airways causing party reversible airflow obstruction, typically assessed with a pulmonary function test. CT findings of asthma include thickening of the bronchial wall, narrowing of the bronchial lumen, areas of decreased lung attenuation on inspiration CT scans, and air trapping on expiration CT scans.¹ CT is not recommended in diagnosing or managing asthma except for assessing complications or comorbidities. However, CT has been used as one of the essential elements of clinical research because it provides valuable structural information on the whole airways and parenchyma in asthmatics. For assessing airway wall changes, qualitative and quantitative evaluation has been used, showing a significant correlation between airway wall thickening and a decline of forced expiratory volume in one second (FEV1).² For assessing the collapse of the airway lumen in the respiration cycle, combined inspiration and expiration CT scanning have been used.² Subjective assessment of emphysema, hyperinflation, a mosaic pattern of lung attenuation, and air trapping on expiration CT can provide information on the parenchymal change. CT density of lung parenchyma can be calculated to quantify emphysema or hyperinflation. The extent and severity of air trapping can be quantified using an image registration algorithm to match the density change of corresponding lung tissues on both inspiration and expiration CT scans.³ Based on CT information of airway remodeling and parenchyma, it is possible to subgroup asthma patients matching with distinctive clinical phenotypes.⁴ Furthermore, using inhaling contrast agents such as Xenon gas combined with the specialized imaging method of dual-energy CT technology, the actual distribution of inhaling can be directly visualized and quantified.⁵ CT is important to link the structural changes and the resulting functional losses of the lung in asthma.

Asthma is a heterogeneous airway disorder characterized by distinct phenotypes with diverse etiologies, natural histories, and treatment responses. In this regard, subgrouping asthmatic patients is an essential issue of asthma research to better understand the disease, leading to personalized management of the individual cases. A paper published in the current issue of Allergy, Asthma & Immunology Research reported the relationship between CT findings and lung function trajectory in patients with asthma.⁶ This paper follows the previous paper of dividing 1,679 asthmatics into 5 subgroups of different lung function trajectories over 1 year, using clinical, demographic, and inflammatory factors.⁷ Fifty-nine asthmatics with available CT images after treatment were included in the study, and 5 different subgroups can be extracted similar to the previous studies. Two chest radiologists conducted a subjective assessment of the absence or presence of 8 different imaging findings on CT, including emphysema, bronchial wall thickening, bronchiectasis, anthracofibrosis, fibrotic bands, mosaic attenuation on inspiration, air trapping on expiration, and centrilobular nodules. In addition, they divided the lung into 6 areas and assessed the finding separately to evaluate the extent semi-quantitatively. They found that the Tr5 group with lower baseline FEV1 and persistent decline of FEV1 after treatment has a larger extent of emphysema and airway wall thickening than the Tr4 group with lower baseline FEV1 and normalized over time. This is an important observation linking both clinical and morphological subtypes of asthma. It also contains clinical usefulness in predicting the treatment response of asthmatics with lower baseline FEV1.

These results, however, warrant further investigation and validation. Several parameters evaluated in this study, such as anthracofibrosis, fibrotic band, and bronchiectasis, may not be the component of asthma but those of combined other diseases such as sequelae of previous infection. It needs to be clarified if these elements should be assessed in the assessment of asthmatics. The criteria for bronchial wall thickening, one of the important findings of this study, seems too strict because these criteria are from the study of bronchiectasis and will depict the airways of severe wall thickening. Of particular interest is the finding of emphysema. According to the criteria for emphysema in this study, the hyperinflation of the lung can be excluded. Considering that the authors scored the emphysema only when its extent is larger than 30% of the covered area, the extent of emphysema in Tr5 seems larger than 15% of the whole lung. As emphysema is the major component of COPD, there should be further studies if the Tr5 groups include the patients with asthma-COPD overlap.⁸ As discussed by the authors, the small number of cases is another limitation. Dividing 59 cases into 5 subgroups may raise an issue of reproducibility. The most critical point of this paper is the subjective assessment of disease findings. To generalize the results of this study, the reproducibility of the image assessment should be evaluated. Nevertheless, the authors didn’t assess the intra- and inter-reader agreement for the evaluation. The authors argued that the proposed subjective semi-quantitative assessment is simple and easy to be used clinically. However, dividing the lung into 6 regions, assessing the presence or absence of the 8 different imaging features separately, and counting the involved areas of each finding may not be done routinely in a clinical setting.

With the advance of computerized imaging processing methods along with the adoption of deep learning technology, fully automated assessment of airway and lung parenchyma has become possible.⁹ All parameters of airway remodeling, such as lumen volume, lumen diameter, wall area, and wall thickening, can be measured in every airway branch extracted automatically. Measured values can be summed and averaged according to the branch levels or regional areas such as lobar or segmental regions. The extent of emphysema, lung volume, and air trapping can also be assessed automatically. However, routine CT quantification in asthma clinics awaits further improvements.¹⁰ Establishment of the normal value of each measure is essential considering normal physiological variation by age, sex, ethnicity, and location within the lung. In addition, measurement variations caused by the imaging protocols, degree of lung inflation, and analysis software should be understood and controlled. Integration of routine workflow without additional resources and time is critical, which has become more realistic nowadays. Most importantly, real-world validation is needed to examine if adding CT assessment improves patient outcomes.

Using CT assessment in managing asthmatics is not a clinical recommendation at present. However, with the recent advancement of quantitative CT assessment of airways and parenchyma, the present study by Kim and colleagues⁶ shows the potential of using CT assessment to phenotype asthmatics for better management in the future. Further investigation with quantitative analysis in a larger number of cases is awaited.