Well, the aim of real-time data is ambitious and exciting. It does raise a question for me, of what it will take to make the data interpretable by a clinician (say, in an ER) — because the model is describing a pretty dynamic and complex system. Are there specific indicator variables that would provide quick insight to a clinician?
- Tie Bo Wu
- http://www.igert.org/profiles/5278
- Graduate Student
- Presenter’s IGERT
- University of California at Santa Barbara
- Michael Trogdon
- http://www.igert.org/profiles/5271
- Graduate Student
- Presenter’s IGERT
- University of California at Santa Barbara
Judges’
Choice
Choice
Judges’ Queries and Presenter’s Replies
Presentation Discussion
-
-
That is a fantastic question, in regards to what output should this model give to a clinician; I think there are many possibilities. One possibility that I think would have a significant impact (if we can have a fast enough model) would be “time til depletion”, so that ER physicians can approximate what stage of the injury the patient is in. Another possibility would be to build in a tool that determines the best intervention strategies based on what the model forecasts. These are just things that I see being possible and useful but we work with many trauma physicians that can provide further insight when the time comes.
Thanks for that question.
-
Further posting is closed as the event has ended.
Jon Kellar
Faculty: Project Co-PI
How does the microfluid cell mimic the surface of a blood vessel? Has the cell surface been treated in any way?
Tie Bo Wu
That is a very good question and I can only speak for the simulation, I’ve asked our collaborators and I will add to this response when they respond.
For this mode we assume that the undamaged portions of the inner walls of a blood vessel do not influence any chemical reactions in the coagulation cascade so we simply treat the other walls of the channel as non-reactive boundaries with a no-slip boundary condition.
Tie Bo Wu
The microfluidic chip mimics physical and biochemical characteristics of a blood vessel experiencing coagulation. First, the microchannel is fabricated to have dimensions on the order of a midsize blood vessel. Second, the coagulation reaction is studied in the channel in blood flowing at physiological shear rates. Third, we pattern coagulation initiation factors and endothelial cells onto the walls of the channel to simulate how coagulation initiates at surfaces and interfaces of a torn blood vessel. Together these components form a novel and more realistic in vitro system for studying coagulation.
-Faye Fong
Marc Porter
Faculty: Project PI
Does mechanical compliance (blood vessel swelling and elongating, etc.) need to be included in the model?
Tie Bo Wu
That is a very good point, thank you for your input. Constriction and dilation of the blood vessel have a significant impact on the flow rates and are therefore important to the coagulation process. It is not in the current model but might be an addition in future versions to fully capture what happens to a specific blood vessel during coagulopathic conditions.
Adriane Ludwick
Faculty: Project Co-PI
Can or will your model be extended to non-accident life-threatening diseases such as a pulmonary embolism and/or arteriosclerosis? Please elaborate on how this can or will be done OR why it can not or will not be done.
Tie Bo Wu
Thank you for your interest. This model can be applicable to thrombus related conditions such as a pulmonary embolism (if initiated by formation of a thrombus) with the addition of one main feature. If we implement an accurate way to model the growth of a thrombus, (a possible way would be to relate it to the concentration of fibrin) we could implement a moving boundary to the problem. In this situation we would solve for a new flow profile at every time step and keep the chemical equations more-or-less the same.
We do plan on adding these capabilities to the model and are already working on adding the moving boundary capability. However we have not decided on the best way to proceed with modelling the growth of the thrombus but that is something that we and our collaborators are trying to figure out.
Peter Gannett
Faculty: Project Co-PI
Your fit to the experimental data is pretty good. However, there is a significant deviation with respect to the lag time for thrombin production and final thrombin production. What elements of your model are responsible for these deviations and how might they be ‘fixed.’
Tie Bo Wu
That is a very good question, this is something that we have discussed with our collaborators at length. This discrepancy may be due to an issue with the bonding of the lipidated tissue factor to the microfluidic channel and it is possible that the flow itself could be sweeping away tissue factor causing thrombin generation to be delayed. But there are adjustments in the model that could be made to speed-up or slow down the production of thrombin. Since human plasma is a variable substance, the model uses an accepted “normal” initial concentration for the input. Therefore it is possible that the deviations could be caused by a difference in initial concentrations. Significant progress has been made to change the procedure and architecture of the experiments but we have not yet received new results to compare to the model.
Antal Jakli
Faculty: Project Co-PI
In what units of your shear rates are given (what does it mean 20 shear, 100 shear, ..?
Tie Bo Wu
That’s a good question, I’m sorry I didn’t make it more clear and I meant for it to be short for 20 shear rate because I felt that 20 per second would be less clear. In the literature shear rate is the most common way of describing fluid flow in blood vessels because it takes into consideration the dimensions of the channel and it is in units of s^-1. Shear rate scales linearly with average velocity and in our simulations, 20 shear rate corresponds to an average velocity of 0.333 mm/s, 100 shear corresponds with 1.667 mm/s etc…