Introduction

Asthma is a heterogeneous disease characterized by episodic airflow obstruction, airway remodeling, and variable response to therapy. Treatment of asthma is generally tailored to the severity of symptoms and adjusted based on the level of self-reported symptom control ¹. Approximately 50% of asthmatics remain poorly controlled ² with up to 10% experiencing recurrent exacerbations. Recent clinical studies suggest that poor disease control is due to airflow heterogeneity ¹, associated with luminal inflammation and airway remodeling as the disease progresses. Yet, the means to comprehensively evaluate and longitudinally track the dynamic airflow heterogeneity remains unknown. Hyperpolarized ³He MRI (HP ³He MRI) offers the capability to depict dynamic airflow changes with segmental resolution. In this study, we evaluated variable airflow heterogeneity responses at the lobar level in asthmatics following bronchodilator (BD) therapy.

Methods

Patients with a clinical diagnosis of asthma (n=36, age=25.6±5.8) and healthy controls (n=8, age=22.1±5.7) were enrolled for HP ³He MRI using a 1.5-T scanner (Siemens Healthcare, Erlangen, Germany). All subjects underwent ³He ventilation MRI accompanied by matching ¹H thoracic cavity imaging at baseline and after BD. ³He MRI was acquired using a fast low-angle shot sequence with 1.95x1.95x10 mm³ (coronal) or 1.64x1.64x10 mm³ (axial) spatial resolution. The ¹H anatomy scan was acquired using a Turbo spin-echo (TSE) sequence. To evaluate therapy response, the baseline and post-treatment ventilation scans were normalized by 1) co-registering the baseline and post-treatment ³He/¹H scans, 2) correcting the aligned ³He images for RF-induced inhomogeneity³, 3) normalizing the intensity range of the aligned ³He scans using histogram matching⁴, and 4) normalizing the ventilation scan intensity histograms to a range of 0 to 1 by rescaling to the top percentile. Ventilation defect percentage (VDP) was first derived from each of the normalized ³He scans using the binning method ⁵. For each normalized ³He scan, local entropy was quantified by calculating the minimum number of bits needed to encode the local intensity histogram. The lobar analysis was performed to determine the entropy in each lobe⁶. Regions with essentially no ventilation also exhibit low entropy given the minimal local intensity variations. Patients with high VDP may thus yield a small averaged global entropy, similar to a healthy subject. Therefore, we introduced a new index – weighted entropy – quantified by multiplying global or lobar VDP ratios and the global or lobar entropy values, respectively. Differences in global weighted entropy, lobar weighted entropy, and response to BD between healthy and asthmatics were performed employing the Mann Whitney test.

Results

At baseline, asthmatics showed significantly higher VDP compared to controls (p = 0.0038), yet no significant lobar VDP difference was found between healthy controls and asthmatics. Employing the weighted entropy index, asthmatics showed significantly higher global weighted entropy than controls (Healthy: 9.1 ± 1.1, Asthma: 25.2 ± 22.6, p = 0.002, Figure 1). Asthmatics also showed significantly higher lobar weighted entropy in the left lower lobe (p = 0.02) and the right lower lobe (p = 0.02) (Figure 1). After BD, asthmatics demonstrated significant global weighted entropy response compared to healthy subjects, with the most significant improvement in the left lung (Figure 2). Figure 3 presents three representative asthmatics that each belongs to a certain weighted entropy group: 1) greater than 15 but lower than 25; 2) greater than 25 but lower than 75, and 3) greater than 75. Figure 3A shows a 35-year-old female with FEV1_%pred=39.2, FVC_%pred = 88.5, FEF_2575%pred=6.6 at baseline. After BD, FEV1_%pred=90.5, FVC_%pred = 124.45, FEF_2575%pred=53.8. VDP decreased from 25.4% to 9.8%. The global weighted entropy was 113.1 at baseline and decreased to 45.8 after BD. The lobar analysis further demonstrated that all lobes contributed to the global weighted entropy decrease as shown in figure 4A, while the left lung and the right lower lobe improved the most. Figure 3B shows 28-year-old male with FEV1_%pred=86, FVC_%pred = 96, FEF_2575%pred=64 at baseline. After bronchodilator, FEV1_%pred=97, FVC_%pred = 104, FEF_2575%pred=80. VDP decreased from 11.4% to 10.7%. The global weighted entropy was 53.5 at baseline and decreased to 48.7 after therapy. The lobar analysis showed that all lobes have contributed to the weighted entropy decrease, except the left lower lobe (figure 4B). Figure 3C shows a 24-year-old male with FEV1_%pred=95.6, FVC_%pred = 102.9, and FEF_2575%pred = 81.7 at baseline. After bronchodilator, FEV1_%pred=117.3, FVC_%pred = 119.6, and FEF_2575%pred=133.6. VDP increased from 6% to 7.5%. The global weighted entropy was 26.8 at baseline and increased to 33.7 after therapy. The lobar analysis showed that the left lower lobe and the right lower lobe contributed to the increased weighted entropy (figure 4C).

Conclusions and Discussion

We found that variable ³He ventilation distribution changes were observed globally in patients with asthma, with the most significant improvement in the left lung. In asthmatics with global improvement in weighted entropy, different lobes may perform differently in airflow distribution. These findings support that airflow obstruction in asthma has a heterogeneous distribution in the lung. Further investigation is needed to better define the various endotypes in asthma and determine how various weighted entropy values relate to the different levels of asthma control.

Acknowledgements

This work is supported by National Heart, Lung, and Blood Institute of the National Institutes of Health (grant no. RO1 HL66479, R01-CA172595, and S10-OD018079).