Biomedical Engineering Reference
In-Depth Information
bringing fresh perspective to the clinicians. Biologists help modelers identify working
hypotheses around which to build their models. While explicitly writing these out and
''translating'' them to code, we can evaluate what is and is not truly known in cancer
biology. Lastly, while developing the model and numerical algorithms, assessing the
expected clinical and experimental data helps in choosing the modeling approach; the
model may expose needs for additional experimental measurements.
Data generation, model calibration, and early testing Modelers and clinicians
jointly plan studies and choose which clinical data to gather (pathology, radiology,
case histories, etc.). Biologists and modelers jointly plan experiments to supplement
the clinical data and inform the model's constitutive relations. These data are inte-
grated into the model with the help of statisticians, image processing specialists, and
others. Early simulations help test and refine the data, model, and calibration.
Simulation, analysis, validation, and feedbacks The calibration procedure is applied
to simulate cancer in individual patients. The simulation data are postprocessed,
yielding quantitative predictions that we validate for each patient. This quantitative
focus allows us to assess and improve our underlying biological hypotheses. If the
predictions are accurate, trials may be planned to assess the model's ability to assist
individual treatment decisions. The modelers, clinicians, and biologists jointly identify
future refinements and experiments. They also jointly select new modeling foci as
suggested by both clinical needs and model-derived insights.
5.2.1 Application of Integrative Modeling to Breast Cancer
This approach guides our work on breast cancer. We have built a team that now
includes oncologists, pathologists, radiologists, biologists and modelers [
63
], and we
are continuing to recruit complementary expertise (e.g., in analytical pathology, tissue
bioengineering, etc.). We have jointly identified that patient-specific predictions of
progression from in situ to invasive carcinoma would be of immense clinical value,
and would naturally build upon our increasingly accurate in situ models. To that end,
we are developing key modeling technologies, such as improved BM and
ECM mechanics [
20
] and multiscale matrix metalloproteinase transport-reaction
kinetics [
19
]. Early modeling results will help guide future experimental design.
Given the critical role of tissue necrosis in DCIS progression, we are developing
next-generation models of intracellular fluid transport, solid synthesis, and dystro-
phic calcification to more accurately describe individual cell volume and composi-
tion changes during these processes [
64
], based upon in vitro measurements we are
currently gathering [
68
]. By this approach, it should soon be possible to accurately
simulate common pathology stains based upon each cell agent's composition. This,
in turn, should make possible new and innovative quantitative comparisons to patient
pathology, better refinement of the otherwise nigh-unmeasurable necrosis time
scales, and ultimately more accurate predictions of clinical progression.
The interested reader can find up-to-date information on these efforts (including
frequent news postings, animations, tutorials, simulation data, and software) at
MathCancer.org. We also encourage the interested reader to visit the newly-
Search WWH ::
Custom Search