The topic of the tenth U.S. Department of Defense (DoD) International State-of-the-Science Meeting (SoSM) on Blast Injury Research was “Toward a Unified Multiscale Computational Model of the Human Body's Immediate Responses to Blast-Related Trauma.” The meeting was held August 16–17, 2022, at the RAND Corporation office in Arlington, Virginia, and more than 60 scientists, clinicians, and military leaders provided scientific overviews, presentations, and posters describing new and emerging science in the field. Before the meeting, a conference planning committee consulted on the literature review and research questions and served as a peer review panel for submitted abstracts. Five leading scientists and clinicians in related fields were invited to serve on an expert panel, lead working groups, and develop overall recommendations.
State-of-the-Science Meeting Summary
The proceedings and findings from the meeting were intended to help address the objectives of the SoSM, which were to
- assess the state of the science of unified multiscale modeling of the human body's responses to blast exposure
- identify major barriers and knowledge gaps impeding progress in the field, along with opportunities for investment in future research
- identify additional opportunities for collaborative action (both government and public-private) that could accelerate progress
- provide recommendations to advance preclinical and clinical research, determine key policy gaps, and identify areas to advance production development.
These proceedings will be of particular interest to scientists, clinicians, military personnel, and policymakers working in areas related to military medicine and health, blast injuries, and, of course, burn injury.
The tenth SoSM was the latest meeting in a series established in 2009 under the authority of the DoD Executive Agent for Blast Injury Research. The meeting and these conference proceedings were sponsored by the U.S. Army Medical Research and Development Command and the DoD Blast Injury Research Coordinating Office. The series aims to identify knowledge gaps in blast injury research; ensure that DoD medical research programs address existing gaps; foster collaboration between scientists, clinicians, and engineers in blast injury–related fields; promote information sharing on the latest research; and identify immediate short- and long-term actions to prevent, mitigate, and treat blast injuries.
Working Group Questions and Answers
Working groups developed responses to four questions, or topics of discussion, designed in advance of the SoSM to address the four meeting objectives. Because the fourth working group question was focused on recommendations, the working group's response to that question is folded into the section titled “Expert Panel Recommendations.”
What Is the State of the Science of Unified Multiscale Modeling of the Human Body's Responses to Blast Exposure?
In each of the working groups, the first response to this question was a question in and of itself: Does the research community need one unified multiscale model of the human body, or does it suffice to have low-fidelity models that are representative of the population with multiscale component models available to researchers as needed? The consensus across each of the working groups was that the state of the science for the indefinite future lies in utilizing multiscale component models on an as-needed basis, particularly given the existence of many high-fidelity component models (for components such as the head, lung, and ear). Although many of these models exist, there is a dire lack of interoperability between said models, posing a significant barrier to progress. Furthermore, although multiscale modeling is emerging, it remains rudimentary at best, one of the reasons being that multiscale finite element models (FEMs) are associated with a prohibitively high computational cost.
Another discussion point was the fact that the primary consideration of human body modeling lies in the body's mechanical response to a blast event. The scientific understanding of the body's mechanical response (also known in various disciplines as biodynamics, kinematics, and/or external body loading) is far more advanced and sophisticated than the understanding of the body's pathological response (also known as tissue biomechanics or internal body loading) to a blast event. Although the current default method for simulating responses at multiple scales is hierarchical modeling, particularly at the component level, subject-matter experts consider even hierarchical modeling to be relatively unsophisticated. For example, the precise formulation of interface conditions between the scales remain a matter of approximations and guesswork. Concurrent modeling schemes that incorporate multiple scales into the same model are nonexistent.
What Are the Major Barriers and Knowledge Gaps Impeding Progress in the Field?
There are various barriers and knowledge gaps impeding progress in the computational modeling of the human body's immediate response(s) to blast exposure:
- There is a lack of diversity in human models because most models represent the average human male. Anatomical variability has a tremendous effect on injury risk and designing personal protective equipment (PPE), among other things, which means that developing models that are increasingly representative of the population is critical to predicting response and subsequent injury. Models beyond the average human male subject are being developed, and this is a space that is rapidly improving for models of the brain, in particular; however, it remains a critical gap.
- The initial conditions (including the stress state), as well as boundary conditions, are often idealized or unknown. For example, tissues are typically studied in the relaxed state even though, in reality, they are in a stressed state before impact. Active muscle creates active loading within the body, which creates an initial level of stress and strain; however, either we do not know this loading or, in cases in which it is known, it is difficult to represent computationally. The idealization of initial and boundary conditions affects researchers' ability to accurately translate results to real-world applications.
- Although the high-fidelity macroscale models of brain biomechanics have been well established, microscale mechanobiology models of injury-sensitive structures (e.g., neurons, glia, axons, synapses, blood barriers) are lacking. Such models are needed to identify potential targets for both acute and chronic neurotherapeutics. It also remains unclear how to specify microscale model initial and boundary conditions from macroscale simulation results.
- There remains a tremendous focus on the impact of the overpressure wave, despite the fact that other forms of energy that cause injury exist.
- As mentioned previously, there is a lack of interoperability between existing high-fidelity component models. These models are not unifiable, meaning that it is not possible for researchers to unify individual models as needed to create a multiscale model of the human body or a portion of the human body. There are many reasons that interoperability of models is impeded, including the use of different solver platforms, different licensing rights, and varying classification levels.
- Similarly, the verification, validation, and certification of models must be more rigorously established. The limitations of models and subsequent applicability to real-world scenarios must be made transparent to all potential users so that models can become fully interoperable.
- Although animal models have been used to study brain injury neuropathology and cognitive responses for years, they have not necessarily provided a pathway for the translation of successful preclinical results to efficacious treatment for humans. A complementary use of humanized in vitro brain-on-chip platforms should be accelerated. It should be easier to develop corresponding mathematical models of such devices than to develop such models of the in vivo human brain.
- There is a distinct need to bridge the gap between animal and human models. There is a high dependency on animal models because of an inability to conduct a variety of needed experiments on humans combined with the lack of a correlation between cadavers and living humans. Cadaver models do not take into consideration critical components, such as blood flow, the inflation and deflation of the lungs, and the like, making live animal models that much more important.
- A fundamental gap exists between modelers and clinical researchers because the scientific community is failing to use model-driven experimental design, resulting in a lack of a connection between simulation results and warfighter performance. Because of a lack of collaboration between necessary parties, it remains difficult to correlate actual physical injury into the computational space—a gap that consistently came up throughout the SoSM. There is a need to correlate the probability of effect (derived from computational and experimental work) with clinical endpoints. Barriers to successfully accomplishing this include a lack of quantitative biomarkers to define these clinical endpoints and poor data collection. Improvements must be made to environmental and physiologic sensors to better facilitate data collection and collect model-appropriate measures in real time, allowing the scientific community to first create models and theories before collecting the requisite data to verify the model.
- A significant barrier to progress is a lack of sufficient knowledge integration and data sharing between various communities. More specifically, there is the perception that the military communities must do a better job of supporting the research and scientific communities. The military has access to meaningful datasets at both the individual and population levels, yet gaining access to these samples and data is challenging. Creating access to these data, whether through interfaces, de-identification processes, or other means, would allow the scientific community to develop more-realistic models that would then yield more-applicable, more-relevant results. Such data sharing would also more readily allow for the reuse of models; users within specific communities could simply utilize tools that already exist, making slight modifications to what has already been created as opposed to reinventing models from scratch.
- Although the focus of this SoSM was on the immediate effects of a blast on the human body, it is worth noting that there is a lack of knowledge on the temporal effects of blast exposure—particularly, those that arise from repetitive loading (e.g., deterioration, adaptability). It remains unknown how repeated blasts affect tissue over time. Given how many service members are exposed to repetitive blasts in a nondeployed setting, determining the effects of repetitive exposure is critical. Similarly, correlations between mechanical insult and the biophysical series of cascading events are unknown. For this reason, current predictive capability is limited to acute or primary mechanical response (injury).
What Opportunities Exist for Investment in Future Research, as Well as Collaborative Action (Both Government and Public-Private), That Could Accelerate Progress?
There are ample opportunities for investment in future research and collaborative action, all of which could accelerate progress in the field. The working groups outlined the following opportunities for future research:
- Future research should emphasize the development of composite, multiscale, multi-physics models, as well as reduced versions that can be used in rapid field responses.
- The modeling and research communities need further support not only for the development of models but also for maintenance.
- In an effort to make simulation results more realistic and relevant to warfighter scenarios, there is a need to create engineered tissues for model parameterization, testing, and validation. This would allow the community to move away from using cadavers and animals and instead use engineered tissues as more-representative surrogates.
- There is a need to integrate mechanistic and data-driven models, allowing for the creation of a digital twin. This effort could first be piloted with a pig before a digital twin for a human is developed.
- Future research should focus on the development of a robust capability to collect, clean, and prepare data for use in mechanistic models. This effort should allow the modeling community to more efficiently identify appropriate data sources while simultaneously ensuring that all data are ethically sourced.
- A long-term commitment of resources (e.g., funding, personnel, and infrastructure) is necessary to create sustainable solutions, and interdisciplinary collaboration must be encouraged and, when possible, required.
Opportunities for collaborative action include the following:
- An artificial intelligence (AI)–based system should be developed that accesses both DoD and civilian databases to better facilitate data sharing across all sectors. Such a system could also be used for classification arbitration and for determination of privileges to access private information using privacy-preserving algorithms.
- There should exist a central DoD organization that coordinates and sponsors all blast-related activities across the services to marshal resources and coordinate disparate, yet concurrent, efforts. This joint organization would also be responsible for issuing a road map for progress, including tangible, measurable milestones. Additionally, this organization would be responsible for hosting an annual meeting for all stakeholders in blast research, providing a space for thoughtful, focused collaboration.
Expert Panel Recommendations
Following the SoSM, there was a closed session with the expert panel to develop final recommendations for future research and suggestions for policy priorities. Working from the literature review, the SoSM presentations, and the working group findings, the expert panel developed final conference findings with proposed directions for future research. We took notes during the closed session, held on August 18, 2022, at the RAND office in Arlington, Virginia, and subsequently synthesized them in Table 1. Although many of the responses to the working group questions listed above make for compelling recommendations in and of themselves, the expert panel specifically highlighted the recommendations below as being of critical importance moving forward. However, it should be noted that the recommendations below do not take away from the importance of many of the previously mentioned recommendations related to research and other domains.
Table 1. Expert Panel Recommendations by Domain
|Domain||Expert Panel Recommendation|
|Data collection and sharing||
|Ethics and diversity||