Identifying Airways Associated with Hyper-Reactivity in Asthma:

1
Identifying Airways Associated with
Hyper-Reactivity in Asthma Interfacing PET
imaging and Dynamic Mechanical Function with 3D
Airway Models N.T. Tgavalekos1, R.S. Harris2, M.
F. Vidal Melo2, G. Musch2, T. Winkler2, J. G.
Venegas2 and K.R. Lutchen1 1Department of
Biomedical Engineering, Boston University,
Boston, MA 2Dept. of Anesthesia and Medicine,
Massachusetts General Hospital, Boston, MA
IMAGE FUNCTIONAL MODELING
GOAL
PATIENT SPECIFIC METHODS
PET Tracer Retention
3D Model Tracer Retention
Lung Mechanics
It has been proposed that the primary location of
heterogeneous constriction resides in the
peripheral airways. Imaging studies, confirm
that ventilation distribution and, hence, airway
constriction in asthmatics is heterogeneous.
However, imaging technology cannot probe small
airways (lt 1mm.) in humans.
Figure 6 Example subject during an asthma
attack. Left reveals, in red, portions of the PET
data that retained the initial tracer at the end
of the washout period (ie., hypo-ventilated
alveoli). Right shows (in red) the terminal
airways in the corresponding model slice assumed
to be un-ventilated.
Figure 7 3D Model reconstruction. Left
represents healthy model with all terminal
branches well ventilated. Right shows the same
model, but with hypo-ventilating alveolar regions
(from PET data, Fig. 6) shown in red.
The overall goal is to advance a new paradigm
called Image-Functional Modeling (IFM), which
synthesizes patient specific structure-function
measures from imaging techniques with dynamic
lung mechanics into 3D computational models. We
desire to identify which alterations in specific
anatomic structures and locations can and cannot
simultaneously alter mechanical and ventilation
lung function in a manner consistent with both
imaging and mechanical data.
Figure 3Lung resistance (RL) and elastance (EL)
represent the mechanical load to breathing.
Experimental studies (top panel) have shown that
as asthma severity worsens, the frequency
dependence and mean level of RL and EL increases.
Morphometric modeling studies (bottom panel)
suggest that the increased frequency dependence
is a result of heterogeneous broncho
constriction in the peripheral airways.
ANATOMICALLY-BASED MODELING
PET Imaging
Structure
Ventilation Spectrum
Case 1
Case 2
AsthmaticPost Challenge
FIGURE 1 A 3D asymmetric conducting airway tree
structure described by Tawhai et al is created on
a personalized basis via MRI and airway tree
space filling algorithms.
over
Figure 4 The PET imaging technique for human
subjects requires a bolus of 13NN labeled saline
solution to be injected intravenously. The 13NN
allows diffuses into the alveoli. The rate at
which the tracer washes out is proportional to
local ventilation. Gas trapping is identified
as areas where 13NN delivered to the alveoli has
not washed out during breathing
normal
under
Function
PET Scan Mask
3D Model-Based Mask
apex
Figure 8 Top The computational model is used to
identify airway constriction patterns that are
consistent with the measured ventilation defects
shown in Figure 6 simultaneous with dynamic R
and E. Bottom Predicted specific ventilation
from Case 1 and 2 vs. measured from PET Case 1
Airways 4 mm and below constricted
heterogeneously so that ventilation defects are
localized in the exactanatomic locations as
identified in the PET image data. Simulation does
not predict well the measured mechanics (top
right). Case 2 Airways .6 mm and below
constricted heterogeneously as in Case 1.
Simulation predicts the best matchwith the
measured mechanics .
Ventilation Distribution
base
SUMMARY
FIGURE 5 Left PET scan mask defines the lung
field for each subject. Using anatomic markers
such as the heart, we approximated the
corresponding 3D Model-based mask (right). The
mask from the model only shows the terminal
airways that actually lie in the same plane as
the corresponding PET image mask for that slice.
To maintain consistency with both the PET images
and mechanical data, closures and constriction of
airways needs to occur small airways
,particularly those below the resolution of
imaging methods for humans (i.e. lt1mm.). Using
the IFM approach we expect that patient specific
models may be developed to aid in interpreting
structural conditions associated with lung
function impairment and evaluating the functional
impact of specific disease management strategies.
FIGURE 2 We model each individual airway as a
compliant tube. The airway terminate onto
viscoelastic alveolar tissue elements. Whole
lung RL and EL can be simulated from the network
combination of all pathways. Acinus ventilation
distribution can also be predicted from the
network analysis.