Major areas of research:

  • Quantitative Imaging characterization of COPD (Airway Disease): Our research has resulted in the novel description of Expiratory Central Airway Collapse (ECAC) as a smoking-related lung disease that is independent of underlying emphysema and airflow obstruction (JAMA 2016). Extending imaging techniques to small airway disease, we quantified the complex branching and remodeling of airways using fractal analysis. This Airway Fractal Dimension (AFD) is associated with respiratory morbidity and lung function decline, offers prognostic information that is in addition to traditional CT measures of airway wall thickness, and can be used to estimate mortality risk (J Clin Invest 2018). We developed a novel metric to identify subclinical small airway disease in regions of the lung that appear normal using traditional density thresholds for emphysema and gas trapping. This metric, the “Normal Density E/I ratio” is a measure of subthreshold gas trapping, represents mild small airway disease, and is associated with respiratory morbidity (AJRCCM 2018). More recently, we showed that the surface area to volume ratio (SA/V) of airways can be used to classify airway disease in COPD into predominant airway narrowing and predominant airway loss phenotypes (AJRCCM 2020).
  • Quantitative Imaging characterization of COPD (Parenchymal Disease): Our imaging research has also advanced our understanding of the mechanics of lung affected by emphysema, and resulted in the derivation of imaging biomarkers to predict lung function decline. Using image registration to match inspiratory and expiratory images voxel-to-voxel, we calculated the Jacobian determinant, a measure of lung biomechanics. Given the disagreement between spirometric impairment and structural disease on CT in many individuals, we showed that CT-derived measures of lung mechanics improve the link between quantitative CT and spirometry (Acad Radiol 2016). We went on to show that the Jacobian determinant, a measure of local lung expansion and contraction, is associated with quality of life, the BODE index, and mortality, thus offering additional prognostic information beyond traditional measures of lung function and single-volume CT metrics (Thorax 2017). With regional measurement of the Jacobian determinant, we showed that areas of normal-appearing lung within 2 mm of emphysematous voxels are mechanically influenced by the emphysematous areas, and termed this the “Mechanically Affected Lung within 2 mm” or “MAL2” (AJRCCM 2018). MAL2 predicts lung function decline, and can thus serve as an imaging biomarker for disease progression.
  • Early Diagnosis of COPD via Imaging: We used deep learning to phenotype COPD into predominant emphysema versus predominant airway disease using spirometry data points (JCI Insight 2020).
  • Early Diagnosis of COPD via Spirometry: We developed several novel new metrics for airflow obstruction that are more sensitive for the detection of early/mild airflow obstruction than traditional spirometry metrics. These include Parameter D, Transition Point and Transition Distance.
  • Deep Learning: We have applied deep learning for the segmentation of the lungs and airways, to phenotype COPD into predominant airway and emphysema subtypes using spirometry data, to phenotype small versus large airway disease, and to diagnose COPD early.

Ongoing Projects:

Structural Determinants of Disease Progression in COPD
PI: Surya Bhatt and Arie Nakhmani
Funding: NHLBI R01HL151421

Deep Learning and Fluid Dynamics Based Phenotyping of Expiratory Central Airway Collapse
PI: Surya Bhatt
Funding: NIBIB R21EB027891

Lung Mechanics in Active Smokers With and Without COPD
PI: Sandeep Bodduluri
Funding: ATS Foundation Research Program