Biomedical Engineering Reference
In-Depth Information
Box 2. Validation Strategy ISM-ABM-CMM
The multiscale model can be validated by first validating each of the mod-
ules (ISM, ABM, and CMM) individually (Steps 1 and 2), and then vali-
dating the integrity of the unified multiscale model (Step 3):
Step 1: Disconnect each of the modules (ISM, ABM, and CMM) from one
another and compare outputs from each module to independent experimental
data collected (or reported in the literature) at that level of scale. For
example, to validate the ISM, one can perform a set of simulations to acquire
an in silico dose-response curve relating SMC proliferation rates to PDGF-
BB concentration. Then compare this predicted dose-response curve to
experimentally measured proliferation rates from PDGF-BB dose-response
studies. This will ensure that the internal structure of the ISM accurately
reproduces SMC proliferation in response to PDGF-BB.
Step 2: Disconnect each of the modules (ISM, ABM, and CMM) from one
another and compare those outputs that each module has in common with
one another. This will ensure that the modules are internally consistent with
one another. For example, both the CMM and the ABM will predict the
thickness of an atherosclerotic plaque, but it is important to check that the
predictions from both modules are congruent (and minimize the residual
differences between ABM and CMM predictions), given that the parameters
and ''rules'' governing the ABM and CMM will be derived from different
sources (Fig.
5
).
Step 3: Validate the unified multiscale model by comparing its outputs (i.e.,
outputs generated by the integrated ISM, ABM, and CMM) to independent
experimental data. Quantitative predictions of the multiscale model at time
t [ t
0
can be compared with identical outcome metrics collected from the
quantitative analyses of hypertensive patients or ApoE -/- plaques. Agree-
ment between prediction and experiment will suggest that the multiscale
model is valid for that range of parameters
The hope in pursuing this type of multiscale modeling is that one day we will be
able to confidently theorize about new drug or knock-out treatments and cause-
effect relationships. For example, if the angiotensin II (ANG-II) receptor type II
(AT-2) is blocked, ANG-II binds to the type I receptor (AT-1) of ECs. Binding of
AT-1 activates the tyrosine kinase and downstream proteins (mitogen-activated
protein kinase (MAPK), Janus kinase (JNK), and signal transducer and activator of
transcription (STAT)) leading to increased intracellular calcium, activation of the
L-type calcium channel, and consequently arterial constriction. Activation of
MAPK also stimulates fibroblast and SMC migration and proliferation via syn-
thesis of platelet derived growth factor and tissue growth factor-b. These growth
factors as well as increased aldosterone all serve to facilitate extracellular matrix
production in a particular collagen, which leads to increased wall stiffening or
pulse wave velocity. Stiffened arteries not only require a larger pressure to distend;
flow propagation is impaired due to inadequate elastic recoil. Thus over time the
Search WWH ::
Custom Search