Return 2012 MSM Working Group discussions
Discussion on Translational Issues at MSM Meeting October 23, 2012
attending: Bill Lytton, Marc Garbey, Rhonda Dzakpasu, Mark Kramer, Tommaso Mansi, Donna Lochner
Discussion centered on making the connections closer to the clinical community. It was noted that involvement of industry might be one route for making such connections. Donna spoke on how FDA works closely with industry to evaluate products and how such involvement might make it easier to provide the level of detail that would be required for certification of a simulation software product for clinical use. Note was also made that many participants are interested in Clinical and Translational issues but also participate in other working groups that focus on their particular method or organ system of interest. It was suggested that successive rounds of WG discussions would allow participants to participate to two different WGs and encourage both vertical and horizontal analysis: e.g. biomechanics as more specialized vertical integration and data sharing horizontal).
These were the topics we had meant to discuss but didn't get to -- all still open for further discussion here as well:
1. Another (nonmonotonic) "scale" to think about: in vitro, in vivo normal, in vivo pathological, human normal, patient.
The traditional translation of a medical discovery is up from the Petri dish to the animal model to the clinical application. Prior to this translation, one often follows a dual process to discovery: we have some ideas from clinical observations that we then test with hypotheses in vitro or in vivo. During the process of moving up and down in these scales we encounter incommensurability -- e.g. mouse models of human pathology that are imperfect due to species differences. Such model-coupling failures could be addressed by explicit systems biology simulation.
This raises several further questions:
A. Hypothesis-driven research: how MSM can be used to 1. refine the hypothesis for experimental validation; 2. define a related hypothesis (a corollary) that can be more readily tested.
B. When can a virtual experiment be considered a reasonable way to test a clinical hypothesis?
C. How can we use models to optimize the experimental plan, design clinical trials, accelerate translation, reduce risk of therapeutic failure?
D. What are common bottlenecks in the MSM approach?
E. Effective strategies to ensure integration and consistency within and between models and model parts.
F. How to go from MSM demonstration of mechanism to clinically-relevant measures (imaging, heatmaps, physiologic variability, etc.) and back.
2. How can we start make connections and contacts among the specific medical specialties: Marc Garbey will discuss the example of surgery. other fields seem ripe for MSMization: Oncology, Radiation Oncology, Radiology, Neurology, Internal Medicine, others?
What role do the various specialty organizations have as gateways to their specialties? How can we tailor talks and reviews for delivery to clinical audiences
3. How to piggyback MSM as the genome and proteome begin to be introduced to clinical practice. How to get clinical and research genomacists interested in MSM.
How will new, very-large datasets regarding an individual patient, particularly from next-generation sequencing, translate into a clinical decision? This will be coming very fast in the cancer field. Many mutations will turn out to be null, or will be secondary to either disease progression or treatment. This is where the mechanistic level understanding will be particularly valuable.
This is a concern that is already encountered in idiopathic (sporadic) epilepsy - while population statistics offer a measure of risk, the reality is that the amount of genetic variation in an individual's genome confounds personalized risk prediction. In our studies [Klassen et al] we have observed known causative mutations for severe myoclonic epilepsy of infancy (SMEI) in normal neurological controls. MSM is about the only way we'll be able to translate genome -> phenome by integrating mutation(s) function(s) mechanistically.