COPD Imaging: New Tools to Tackle an Old Problem?

Keywords :

• COPD
• CT emphysema
• perfusion

There is a growing appreciation of the staggering global burden of COPD, but for a number of reasons, there are yet no treatments that significantly reduce mortality rates or alter the disease course. This lack of progress is due in part to the fact that COPD is typically diagnosed when the disease is well-established because the respiratory system, like most human systems, is over-engineered for day-to-day tasks. Moreover there is a myriad of underlying COPD phenotypes that demand targeted therapy and new measurement tools are needed to accurately identify and quantify these phenotypes. Although COPD phenotypes beyond those related to emphysema and airways disease have been identified (1–7), it has been extremely challenging to develop and validate their measurement and develop targeted therapies.

A number of studies have recently shown that certain COPD patient phenotypes show superior response to certain treatments. For example, there is preliminary evidence that greater short-term bronchodilator responsiveness is observed in COPD patients with specific ADRB2 polymorphisms (3). Another study showed that COPD patients with chronic bronchitis reported greater benefit in terms of pulmonary function, symptoms and exacerbations specific to roflumilast therapy (8). Moreover, ECLIPSE identified a frequent-exacerbation phenotype (4) in COPD patients, regardless of GOLD grade. Clearly, measurement tools that identify patient phenotypes or subgroups that respond to therapy, or patients that require more aggressive monitoring, have the potential to improve patient outcomes and quality of life. However, the development of more targeted, phenotype-specific therapies requires a better understanding of the underlying structural and functional changes that lead to airflow limitation, COPD symptoms and exacerbations.

Imaging-based COPD phenotypes have the advantage that they can provide quantitative, regional and independent measurements of underlying disease mechanisms, regardless of FEV₁ (11, 12). For example, the National Emphysema Treatment Trial (NETT) investigated mortality following lung-volume-reduction surgery in patients with severe emphysema. Although there were no survival benefits for all COPD patients when pooled (5), subgroup analyses revealed dramatic long-term mortality improvements in patients with upper-lobe emphysema quantified using computed tomography (CT), and worse survival in patients with non-upper-lobe predominant emphysema (5). As an alternative to thoracic surgery, bronchoscopic lung-volume-≠reduction approaches can be implemented using a variety of interventions (9–11), including endobronchial valves (12).

Although there is much lower morbidity associated with bronchoscopic versus surgical interventions, the Endobronchial Valve for Emphysema Palliation Trial (VENT) study reported similar pitfalls to the surgical approach (6, 7). In all patients studied, VENT reported very modest improvements in lung function, exercise tolerance and quality of life, but subgroup analysis revealed that patients with complete fissures (measured using CT) showed the greatest improvement (6, 7). Although the reasons for these stark ≠differences in response are not completely understood, it is very clear that COPD is ≠heterogeneous and patients respond in different ways to therapy, and this is likely dependent on underlying pathologies. Better prospective COPD patient phenotyping holds the promise for the development and testing of new treatments and perhaps identification of patients with early disease. In this regard, phenotyping addresses two important unanswered questions in COPD: Can better patient phenotyping lead to phenotype-specific treatment and ultimately better outcomes? Can earlier diagnosis and intervention change the natural history of COPD? In order to begin to answer these questions, we need better COPD measurement tools.

In the current issue of the journal, two teams of COPD researchers aimed to address these challenges using X-ray computed tomography (CT) of the airways and parenchyma (13) and perfusion measurements using magnetic resonance imaging (MRI) (14). In a relatively large study of 1,138 COPD patients, Hoesein and colleagues (13) exploited CT phenotype measurements to better understand their relative contributions to overall pulmonary function. Emphysema appears as regions of low attenuation on CT images and can be quantified by visual inspection or CT density thresholds (15). Airways disease can also be evaluated using 3D reconstruction and segmentation of airway wall features (16). However, the resolution limits of CT practically mean that only airways up to the 5th or 6th generation can be visualized and quantified, so that the smallest airways cannot be directly measured.

Indirect CT measurements of small airway function made using expiratory CT images (17) is based on the premise that small airway dysfunction leads to regional differences in emptying rates, manifesting as low attenuation regions that reflect air trapping. Hoesein and co-workers determined the relative contribution of these underlying COPD phenotypes to pulmonary function measurements. They showed that emphysema measurements were most strongly associated with DLCO and FEV₁/FVC, while airway wall thickness and air trapping were most strongly associated with FEV₁ and RV. Although previous investigations have tackled these issues (18–20), this study uniquely teases out the role of three different CT phenotypes and their dominant contributions to COPD functional abnormalities.

Another important contribution in the current issue of the journal (14) investigates the use of MRI pulmonary perfusion measurements in very early or mild disease in 51 smokers. A previous study evaluated CT perfusion measurements in smokers with normal spirometry, but with evidence of very mild centralobular emphysema (21). This previous work showed the presence of greater pulmonary perfusion heterogeneity in the smokers with mild emphysema than smokers without emphysema. Prior to this, there was some experimental evidence that smoking induced pulmonary vascular remodeling and this may precede emphysema (1), but this finding was discordant with the previous understanding that vascular changes are secondary to emphysema in COPD patients. Taken together, these previous findings point to the importance of identifying pulmonary vascular abnormalities in smokers with or without emphysema as an early disease marker.

Pulmonary MRI is technically challenging and lacks the spatial resolution of CT but has the advantage that it does not use ionizing radiation, allowing for serial and longitudinally imaging to be performed without regard to radiation risk. Xia and colleagues (14), exploited these advantages and investigated MRI pulmonary perfusion measurements in 13 smokers at high risk of COPD, but with normal spirometry, and in 38 smokers with GOLD I-IV COPD. They report that all smokers in the high-risk group had MRI perfusion defects, but less than half of them showed CT evidence of emphysema. Interesting, there were a greater proportion of subjects with perfusion defects than CT evidence of emphysema in the mild and moderate COPD groups as well, although these differences were not statistically significant. For all smokers, there were stronger FEV₁ correlations with MRI as compared to CT measurements. The authors finally concluded that perfusion MRI measurements may be more sensitive than CT to early or mild disease abnormalities in smokers at high risk of but without diagnostic criteria for COPD.

These findings have the potential to have an impact on our understanding of early COPD and to improve its detection. However, a better understanding of the etiology and clinical meaning of MRI perfusion defects is still needed. For example, MRI perfusion defects in patients without obvious emphysema may be explained by the insensitivity of CT to very early or mild emphysema or, as the authors speculate, there may be a relationship with small airway remodeling. Regardless, it will be important to demonstrate that MR perfusion defects are reproducible and direct comparison to CT perfusion measurements will strengthen these findings. Furthermore, these subjects should be monitored longitudinally to confirm that perfusion defects precede airflow limitation and COPD.

One aspect that deserves mention is the exclusion of female smokers in both studies. The growing prevalence and mortality rates associated with COPD in women are alarming (22), and studies investigating these CT and MRI phenotypes in women are warranted. Perhaps we need to re-think smoking history requirements in large-scale studies because this typically excludes women and they may be at an increased risk of developing COPD with lower exposures (23). It has also been demonstrated that women experience a greater loss of lung function per pack-year smoked (24, 25) and are also more likely to experience severe, early onset COPD (26). These findings suggest that there are differences in the natural history of COPD between the sexes, and that sex differences may exist for the underlying disease phenotypes that lead to airflow limitation. Investigating potential sex-differences in phenotype contributions to airflow limitation and other pulmonary function measurements will improve our understanding of the impact of COPD in women. Moreover, more information regarding the role of vascular disease in early COPD in women is also warranted and studies using MRI perfusion imaging as reported here (14) may be the key to improving this understanding.

The thread that runs through both of these important contributions is that there is an unmet need for measurement tools that allow for better COPD patient phenotyping –in particular for patients in early or mild stages. There is still much for us to learn from the use of existing CT tools (13) and from novel, more sensitive MRI tools (14), both demonstrated in the current issue.

Nevertheless, the findings in both manuscripts urge us to continue to pull the threads together to weave a robust tapestry of phenotyping approaches that will accelerate the development of better treatments and outcomes in COPD patients.

Declaration of Interest Statement

The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.