CT-based Biomarker Provides Unique Signature for Diagnosis of COPD Phenotypes and Disease Progression

1] Department of Radiology, University of Michigan, Ann Arbor, Michigan, USA. [2] Center for Molecular Imaging, University of Michigan, Ann Arbor, Michigan, USA.
Nature medicine (Impact Factor: 27.36). 10/2012; 18(11). DOI: 10.1038/nm.2971
Source: PubMed


Chronic obstructive pulmonary disease (COPD) is increasingly being recognized as a highly heterogeneous disorder, composed of varying pathobiology. Accurate detection of COPD subtypes by image biomarkers is urgently needed to enable individualized treatment, thus improving patient outcome. We adapted the parametric response map (PRM), a voxel-wise image analysis technique, for assessing COPD phenotype. We analyzed whole-lung computed tomography (CT) scans acquired at inspiration and expiration of 194 individuals with COPD from the COPDGene study. PRM identified the extent of functional small airways disease (fSAD) and emphysema as well as provided CT-based evidence that supports the concept that fSAD precedes emphysema with increasing COPD severity. PRM is a versatile imaging biomarker capable of diagnosing disease extent and phenotype while providing detailed spatial information of disease distribution and location. PRM's ability to differentiate between specific COPD phenotypes will allow for more accurate diagnosis of individual patients, complementing standard clinical techniques.

Download full-text


Available from: Meilan K Han, Oct 03, 2015
34 Reads
  • Source
    • "The impact of alterations during disease, such as airway narrowing or changes in tissue properties, on regional ventilation and tissue stresses are not well understood. For example, one hypothesis is that airway disease may precede emphysema [12]. The computational lung model could be applied to investigate the impact of airway narrowing and tissue stiffness during obstructive lung diseases on tissue stresses, alveoli pressure and ventilation. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This work is motivated by the modelling of ventilation and deformation in the lung for understanding the biomechanics of respiratory diseases. The main contribution is the derivation and implementation of a lung model that tightly couples a poroelastic model of lung parenchyma to an airway fluid network. The poroelastic model approximates the porous structure of lung parenchyma using a continuum model that allows us to naturally model changes in physiology by spatially varying material parameters, whilst conserving mass and momentum within the tissue. The proposed model will also take advantage of realistic deformation boundary conditions obtained from image registration, to drive the simulation. A finite element method is presented to discretize the equations in a monolithic way to ensure convergence of the nonlinear problem. To demonstrate the coupling between the poroelastic medium and the network flow model numerical simulations on a realistic lung geometry are presented. These numerical simulations are able to reproduce global physiological realistic measurements. We also investigate the effect of airway constriction and tissue weakening on the ventilation, tissue stress and alveolar pressure distribution and highlight the interdependence of ventilation and deformation.
  • Source
    • "More complex pattern recognition algorithms can be used to identify the appearance of different diseases [6]. In particular, a method involving parametric response maps derived from CT data acquired at full inspiration and full expiration has been described, which is useful in distinguishing between emphysema and functional small airways disease [7]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Background Determination of regional lung air volume has several clinical applications. This study investigates the use of mid-tidal breathing CT scans to provide regional lung volume data. Methods Low resolution CT scans of the thorax were obtained during tidal breathing in 11 healthy control male subjects, each on two separate occasions. A 3D map of air volume was derived, and total lung volume calculated. The regional distribution of air volume from centre to periphery of the lung was analysed using a radial transform and also using one dimensional profiles in three orthogonal directions. Results The total air volumes for the right and left lungs were 1035 +/− 280 ml and 864 +/− 315 ml, respectively (mean and SD). The corresponding fractional air volume concentrations (FAVC) were 0.680 +/− 0.044 and 0.658 +/− 0.062. All differences between the right and left lung were highly significant (p < 0.0001). The coefficients of variation of repeated measurement of right and left lung air volumes and FAVC were 6.5% and 6.9% and 2.5% and 3.6%, respectively. FAVC correlated significantly with lung space volume (r = 0.78) (p < 0.005). FAVC increased from the centre towards the periphery of the lung. Central to peripheral ratios were significantly higher for the right (0.100 +/− 0.007 SD) than the left (0.089 +/− 0.013 SD) (p < 0.0001). Conclusion A technique for measuring the distribution of air volume in the lung at mid-tidal breathing is described. Mean values and reproducibility are described for healthy male control subjects. Fractional air volume concentration is shown to increase with lung size.
    BMC Medical Imaging 07/2014; 14(1):25. DOI:10.1186/1471-2342-14-25 · 1.31 Impact Factor
  • Source
    • "Applications in medical imaging range from motion compensation to intra-operative fusion of different modalities. In particular, non-linear registration methods are able to capture complex deformations with high accuracy , enabling advanced diagnosis and treatment [2]. Many of these methods, however, exhibit long processing times or require special hardware such as GPUs. "
    [Show abstract] [Hide abstract]
    ABSTRACT: We present a novel parallelized formulation for fast non-linear image registration. By carefully analyzing the mathematical structure of the intensity independent Normalized Gradient Fields distance measure, we obtain a scalable, parallel algo-rithm that combines fast registration and high accuracy to an attractive package. Based on an initial formulation as an opti-mization problem, we derive a per pixel parallel formulation that drastically reduces computational overhead. The method was evaluated on ten publicly available 4DCT lung datasets, achieving an average registration error of only 0.94 mm at a runtime of about 20 s. By omitting the finest level, we obtain a speedup to 6.56 s with a moderate increase of registration error to 1.00 mm. In addition our algorithm shows excellent scalability on a multi-core system.
    IEEE International Symposium on Biomedical Imaging: From Nano to Macro, Beijing, China; 04/2014
Show more