5

Treatment

___________________

¹ This list is the rapporteurs’ summary of points made by the individual speakers identified, and the statements have not been endorsed or verified by the National Academies of Sciences, Engineering, and Medicine. They are not intended to reflect a consensus among workshop participants.

The fifth session of the workshop featured a firsthand account from the father of a traumatic brain injury (TBI) survivor of her injury and the long-term ramifications on her and her relatives’ lives. The session then explored innovative clinical trial designs to increase learning efficiency. The objectives of the session included (1) describing recent advances in clinical trial designs that enhance efficiency in evaluating therapeutics for heterogeneous medical conditions such as TBI, (2) presenting an evolving evidence base for pharmacological and nonpharmacological therapies in TBI—including examples of ongoing clinical research—and future directions, and (3) highlighting challenges in translating academic research to therapeutic interventions, including regulatory and industry barriers. Ramon Diaz-Arrastia, professor at the University of Pennsylvania, moderated the session.

Diaz-Arrastia remarked that treatment is often considered to be the holy grail in TBI medicine; this perspective can divert focus away from developing

new diagnostics that will not necessarily affect treatment. However, Diaz-Arrastia contended that the value that accurate diagnosis and prognosis brings to patients is itself of critical importance. Development and widespread buy-in and implementation of both diagnostic and therapeutic strategies will involve the use of biomarkers to identify and enroll the patients best suited to participate in a given clinical trial. Prognostic or predictive biomarkers could aid in the selection process, he noted, and could also inform the particular therapy that specific patients are more likely to benefit from. Pharmacodynamic biomarkers could facilitate determinations of optimal dose, timing, and duration of therapies. Diaz-Arrastia maintained that developing the suite of biomarkers needed in the TBI field will be an iterative process.

A LIVED EXPERIENCE PERSPECTIVE

E. Wesley Ely, parent of a family member with TBI and professor and Grant W. Liddle endowed chair in medicine at Vanderbilt University School of Medicine, described how his daughter sustained TBI as a child and the decades of ramifications she and her family have experienced as a result. Ely authored a work of narrative nonfiction about these hardships, and he read a passage about the injury event in which he and his wife were at the pool with their three daughters.

Ely’s daughter was taken by ambulance to the Vanderbilt Emergency Department where scans revealed that she had fractures on both sides of her skull that extended through her foramen magnum (the opening in the base of the skull through which the spinal cord passes). After several days in a neuro intensive care unit (ICU), doctors told Ely that his daughter was fine and would fully recover. Noting that he is an intensivist and his wife is a pathologist, Ely remarked that for years providers—including those in trauma and neurosurgical services—assured them that their daughter was

okay. She was afraid of falling each time she climbed into her small bed, but no follow-up visits for TBI were recommended, said Ely.

Going on to study acquired brain injury after critical illness as a physician scientist, Ely recalled being perplexed 25 years ago by patients who faced persisting issues after being discharged from the ICU, including unemployment, broken relationships, and suicidal ideation. Studying postintensive care syndrome and Alzheimer’s disease and related dementia, he described being unaware that his daughter could be experiencing chronic effects from her TBI, despite his lingering concerns. He attempted to request positron emission tomography, single-photon emission computed tomography, or functional magnetic resonance imaging (MRI), but said he was unable to convince providers to order these diagnostics for her.

Ely described how the post-TBI challenges his daughter faced placed significant stress on both her and the family, reflecting that she had strived and struggled to find her place in the world. Ely described witnessing his daughter’s hardships as a brutal experience for a father, commenting that from his perspective the precarious situations, decision-making choices, personal struggles, and job, mental, and physical health challenges she encountered took a toll on the entire family. His marriage suffered and he spent years suppressing his pain, focusing instead on a need to control his life and career. In the past 5 years, he noted that he has become active in Al-Anon—a spiritual but nonreligious program for people with a loved one contending with addiction—and reflected that his involvement with the Forum on Traumatic Brain Injury led to his understanding that his daughter has chronic TBI. At a forum meeting shortly after its creation, he realized the long-term effects from his daughter’s TBI and broke into tears in the meeting room. He recounted his shock at realizing that so many years had passed without understanding the full entity that his daughter and family were contending with. Ely stated his hope that forum efforts will address long-term—not just acute—TBI and consider the ways that patients, families, spouses, and others suffer in facing the challenges of TBI. He remarked that people directly affected by TBI must be given a voice, because in the absence of their perspective, everyone loses.

CLINICAL TRIAL DESIGNS TO ACCELERATE PROGRESS IN THE TREATMENT OF TBI

Roger Lewis, professor at the University of California, Los Angeles, explored the benefits of innovative clinical trial designs before outlining various design types and examples of studies using innovative designs (such as those discussed in Berry et al., 2015; Woodcock and LaVange,

2017; Butler et al., 2018, Kidwell and Almirall, 2023, and other sources).² Underscoring the hardships described by Ely, he stated his hope that more effective therapies for the treatment of TBI will be developed and that this will lead to changed narratives for the patients and their families in the minutes, days, weeks, and decades after injury. Similar to many complex therapeutic areas, numerous challenges face later stage development (i.e., phases II and III³) and evaluation of new treatments for TBI. He remarked that many of the concepts pertinent to acute TBI treatments also apply to diagnostics, rehabilitation, and long-term therapy.

Highlighting the complexity and heterogeneity of the TBI patient population and injuries as key challenges, Lewis noted that most TBI treatments are not effective and fail when tested in clinical trials. Effective therapies are likely to be characterized by (1) combinations of treatments, (2) sequences of treatments, in which decisions within the sequence are dependent on individual patient responses, and (3) therapies matched to specific responding subsets of patients identified by biomarkers, imaging patterns, electroencephalography (EEG), clinical phenotypes, or other groupings. He added that significant heterogeneity of treatment effect is likely, in which a treatment that is effective for one patient subgroup is not effective, or could potentially even cause harmful effects, for another patient subgroup. Progress in TBI treatment development has likely been slowed by a disconnection between typical trial design—which tends to be overly simple with questions distilled down to a single focus—and care delivery and characteristics of the patient population, both of which are highly complex, Lewis asserted.

Bridging this disconnection, marked progress is evident across therapeutic areas in accelerating the evaluation of therapies through innovative clinical trial design, said Lewis. He described that little is unique about the clinical trial design challenges in the field of TBI in comparison to other fields. The statistical models are agnostic to disease and operate the same across diseases, yet the diffusion of innovation advances in the field of TBI has been slow. In bringing innovative clinical trial design across

___________________

² Lewis noted that his academic and employment relationships include the County of Los Angeles Department of Health Services, David Geffen School of Medicine at UCLA, and Lundquist Institute for Biomedical Innovation, as well as Berry Consultants LLC, which works in adaptive trial design. He shared that he is a special government employee with FDA (currently inactive) and the lead statistical editor for JAMA (Journal of the American Medical Association).

³ In clinical research, phase I studies involve a small number of participants and aim to understand the safety of a treatment and help identify dosage, while phase II studies explore its efficacy and side effects. Phase III studies are late-stage trials to test efficacy in larger numbers of people (see https://www.fda.gov/patients/drug-development-process/step-3-clinical-research#Clinical_Research_Phase_Studies; accessed August 10, 2024).

areas, Lewis is often asked whether a certain type of trial has been run on a specific disease, as if the disease determines the performance of a clinical trial design. Another common question is whether the U.S. Food and Drug Administration (FDA) approved a drug or device evaluated in the trial, given that when FDA denies clearance, this is typically because of the product’s failure to improve outcomes, Lewis said.

Innovative Trial Goals

Innovative trial design can be used to accomplish multiple goals, Lewis stated. For example, researchers want to reduce the risk of failure, with failure defined as a trial that ends without yielding a clear answer to the primary motivating question. In such cases, researchers may experience “anticipated regret”—anticipating that they will wish at a trial’s completion that a different design decision had been made regarding an area of uncertainty. For instance, when researchers are not confident about the dose, patient population, or proper outcome measure to use in a test, and the option they select does not yield satisfactory answers, they may anticipate making a choice that they later regret. These areas of anticipated regret are productive areas to address through innovative trial design.

Another goal of innovative trial design is efficient failure at the individual therapeutic level. He explained that if an individual therapy is not going to prove effective, realizing this as quickly as possible enables researchers to divert energy toward therapies currently offering promise. Efficiency should be pursued via multiple paths, said Lewis. Structural efficiencies can be achieved through trial design, such as the use of shared control groups or factorial trials that simultaneously evaluate multiple therapies. Operational efficiencies include platform trials that involve multiple arms for various treatments, rather than running separate trials for each treatment being evaluated. Inferential efficiencies involve making better use of the available data and can include adaptive designs and response-adaptive randomization. He explained that response-adaptive randomization randomizes patients preferentially to the treatments that appear to be performing better at a point in the trial where superiority has not yet been demonstrated. Lewis noted that mirroring clinical care is an additional innovative clinical trial goal; this could include evaluating sequences of therapies driven by patient response, thereby embracing the need for personalized therapies that are based on individual patient characteristics.

Examples of Innovative Trial Designs

Lewis outlined example trial designs that carry the potential to accelerate development in TBI once promising therapies are available to test.

Adaptive designs have prespecified rules for using incoming data to modify trial characteristics, such as randomization ratios, sample size, or enrollment criteria. Trial characteristics are modified to mitigate the risk of failure, to improve treatment of patients within the trial, and to achieve statistical efficiency to reach a conclusion as early as possible, whatever the conclusion may be. He noted that time to information is critical in an adaptive trial, and an outcome measure must become rapidly available to provide stimulus for the trial’s response. For many acute and medium-term TBI treatments, the outcomes are known fairly early because differences in patient groups and recovery trajectories are at focus, rather than the final outcome that will ultimately be achieved by the patient 6 months or 10 years in the future, he explained. An adaptive enrichment trial is a highly specific type of adaptive trial in which the initial structure of the trial involves broad inclusion criteria that then narrow to a responsive subpopulation once a subpopulation emerges, said Lewis.

Adaptive Trial Designs

Presenting recent examples of adaptive trial designs in practice, Lewis summarized two trials in ischemic and hemorrhagic stroke. In a trial evaluating thrombectomy for patients with large vessel occlusion stroke,⁴ researchers were uncertain whether a size limit in the core volume of a stroke applied to therapy response (Nogueira et al., 2018). Rather than guessing, investigators designed a trial to enroll patients with a broad range of core stroke sizes. They then followed the data to determine whether only a smaller subpopulation was responding to the treatment, at which point they would narrow the trial. Researchers found that the therapy worked broadly across the patient population, and the trial adaptation of narrowing the population was never needed because the full population saw benefit.

This result contrasts with that of a trial published days before this workshop that examined minimally invasive surgical technique for the evacuation of intracerebral hemorrhage (ICH) (Pradilla et al., 2024). In this trial, investigators were uncertain whether the therapy would work equally well in anterior basal ganglia locations and in lobar locations for ICH. As the trial progressed, the algorithm found that the benefit of the therapy seemed to be located only in those patients with lobar locations of ICH. Instead of potentially ending with a failed trial, researchers ceased enrolling patients with a basal ganglia location and went on to demonstrate tremendous benefit in the responding subpopulation of patients with lobar

___________________

⁴ This was the Diffusion-Weighted Imaging or Computed Tomography Perfusion Assessment with Clinical Mismatch in the Triage of Wake-Up and Late Presenting Strokes Undergoing Neurointervention with Trevo (DAWN trial).

locations. Lewis emphasized that simply running the trial with a very large set of inclusion criteria could have resulted in failure to identify treatment benefit.

Platform Trial Designs

In addition to adaptive designs, innovative trial designs include platform trials, said Lewis. A platform trial is an experimental infrastructure intended to investigate or compare the outcomes associated with multiple treatments in which available treatments change during the course of the trial. Thus, treatments may be introduced, they may graduate owing to success, or treatments may leave the trial because of the futility of further investigation. Lewis likened the platform trial to a sports stadium hosting multiple games in that the stadium does not have to be rebuilt for each match. Treatments enter and exit the platform and are evaluated against one another and against the control or standard of care, thereby generating tremendous operational efficiency, he explained.

A multifactorial platform trial is designed to examine combination therapies such as those used routinely during the treatment of TBI patients during initial resuscitation. Lewis explained that a platform trial might begin with the investigation of three agents: A, B, and C versus control (see Figure 5-1). Researchers may stop testing one or more agents if they find the treatment harmful or futile. Should one of the experimental arms perform better than the control, the trial is not ended and restarted from the beginning; instead, the better therapy now becomes the standard of care, and the trial can continue.

Many people believe that platform trials are only used in oncology, but this design has been used in trials for a variety of conditions, including COVID-19 and amyotrophic lateral sclerosis, Lewis stated. He commented that the fear generated by the COVID-19 pandemic made people more receptive to innovation in clinical trial design, which accelerated use of these approaches during the pandemic. The National Institute of Neurological Disorders and Stroke has initiated a major effort to launch a multicenter adaptive platform trial in acute ischemic stroke.

Platform trials have multiple potential efficiencies that can be categorized as structural efficiencies, adaptations, and statistical approaches. He specified that because multiple therapies are used for patients during the acute postinjury period of TBI, multiple treatments should be evaluated in each patient enrolled in TBI trials. In a multifactorial platform trial (see Figure 5-2), researchers wanting to compare neuroprotectants—or hemostatic treatments, in the case of hemorrhage indicated by computed tomography (CT)—would conduct factorial randomization in which the patient would be randomized to one option within each of the domains of

therapy to provide an individual patient regimen. He explained that this approach allows the individual patient to yield information on how best to control hemorrhage and on the effectiveness of the neuroprotectant, thereby fundamentally accelerating the speed with which both therapeutic areas can be evaluated. Lewis outlined that an international research team success-

fully used a single multifactorial platform to test COVID-19 treatment by comparing results on interleukin-6 antagonists, anticoagulation in seriously ill patients, anticoagulation in noncritically ill patients, and in other areas (Goligher et al., 2021; Gordon et al., 2021; McQuilten et al., 2023).

Sequential Multiple Assignment Randomized Trial Design

Sequential multiple assignment randomized trials (SMART) are specifically designed to evaluate sequences of therapies in which the treatment choice at each randomization point is dependent on the patient’s response to prior therapy, Lewis explained. In a SMART, a patient is randomized between available options for initial management. Some patients will respond well to the therapies they are assigned to, whereas others will not. The patients that fail to respond to their assigned therapy are randomized between alternative secondary or second-line treatments. He highlighted a SMART design that rigorously evaluates the overall effectiveness of four separate treatment algorithms (Kidwell and Almirall, 2023). Lewis stated that this design could be applied to address variability in patient response to treatments received over time to objectively evaluate the best sequence of therapies.

Benefits and Common Pitfalls

Lewis presented a table of the various types of adaptive and platform trials and their potential benefits in terms of how well each type meets specific innovative trial design goals (see Table 5-1).

Lewis explained that each approach has strengths, and the use of an individual therapy in an innovative trial often improves the participant’s prospect of benefit. These approaches could reduce the risk of failure, although the degree to which risk is reduced depends on whether failure is defined as pertaining to the patient population or to the entity being evaluated, he specified. Each design improves inferential efficiency (i.e., the average number of patients required to definitively answer a therapeutic question), but the size of gains varies between approaches. Lewis explained that these designs offer the prospect of operational efficiencies, and the SMART design is particularly useful in evaluating the longitudinal trajectories of care influenced by patient response.

Lewis cautioned against falling prey to common pitfalls in trial design. He highlighted the tendency to believe that trial design for one’s focus area is special. From a clinical design perspective, nothing is special about TBI, he contended. Additionally, investigators should avoid mistaking careful thinking for actual information. For example, a company unsure about the optimal dose, population, outcome, or responding subset for a therapy might consult with experts and then misconstrue expert opinion as fact.

Instead, researchers should design studies that yield data that enable identification of correct options. This approach mitigates the risk of failure associated with the uncertainty that exists at the time of trial design. Lewis encouraged investigators to avoid blaming prior failures on trial designs or selection of wrong outcome measures. Typically, trials fail because the treatment does not work, he remarked. Lastly, he underscored that phase II trials should be designed with phase III considerations in mind. For instance, a phase II trial design can aid in determining the predictive probability of phase III success. All too often, the design of phase II trials has little to do with the structure of the phase III program, said Lewis.

The design of clinical trials in TBI can be improved by learning and incorporating progress from therapeutic development in other fields, Lewis summarized. Adaptive trials, platform trials, and SMART can yield substantial operational and inferential efficiencies and improve patient-centeredness, he asserted. Underscoring that particular trial designs have tremendous potential to accelerate the evaluation of TBI therapies, he stated that researchers should match the specific innovative trial design to the specific threats to a particular evaluation program’s success. Lewis emphasized that success of a trial is characterized by efficiently and definitively

answering the motivating question, regardless of whether that answer is positive or negative.

PANEL ON PHARMACOLOGICAL AND NONPHARMACOLOGICAL TREATMENT

Introducing a panel of academic investigators with entrepreneurial interests, Diaz-Arrastia noted challenges engaging large pharmaceutical companies in TBI research and development, and shared that the panel features speakers from small biotech companies to share progress, barriers, and limitations they have experienced in their work and highlight opportunities for collaboration moving forward.

Targeting the Apolipoprotein E Receptor

Daniel Laskowitz, professor of neurology and therapeutic lead of neuroscience medicine at the Clinical Research Institute at Duke University, said that he has dedicated his career to acute brain injury as an attending physician in the neuro-ICU and in preclinical research. Stating his frustration at the slow pace of therapeutic development, he recalled the absence of nonsurgical treatments to offer TBI patients in the 1990s, when he began practice. This lack of therapeutics necessitated challenging conversations with patients and caregivers, and he described being haunted by the decisions families faced in determining whether to withdraw care from their children or continue life support. These outcomes motivated Laskowitz to create a translational laboratory to develop new therapies and clinically relevant models. Three decades later, TBI treatment remains limited after surgical stabilization, he said. Despite increased insight into basic mechanisms of secondary tissue injury, translation of new pharmacological therapies has been slow, said Laskowitz.

Early in his career, Laskowitz explored the genetics of recovery following brain injury. His time spent in the ICU demonstrated that two patients could share the same injury, demographics, and mechanism of injury, yet one would be discharged from the hospital and the other would die from severe cerebral edema. This variability in response to injury motivated him to examine genetic influences on recovery, with a specific focus on the polymorphisms of apolipoprotein E (apoE), a protein involved in the metabolism of fats that is associated with Alzheimer’s disease and that influences inflammation, excitotoxic injury, and outcome. Since it does not cross the blood–brain barrier, apoE is not a therapeutic that can be administered peripherally, he clarified. Laskowitz and colleagues have spent approximately 30 years working to move from signal transduction to cell culture in an effort to make the

protein receptor region a druggable target. They developed a product that has progressed through phase I and is currently in phase II trials.

Laskowitz has collaborated with senior investigators who are developing promising therapies in TBI that aim to protect against excitotoxicity or neuroinflammation. Acknowledging that advancing from cell culture to animal modeling and through FDA review and approval processes is necessarily a time-intensive process, he stated that several common pressure points can hinder progress. For example, intellectual property (IP) issues can hinder translation, he said, noting that academics typically approach research through a process of obtaining grants and publishing research, but public disclosure can preclude potential commercialization opportunities.

Additionally, generating interest in funding therapeutics is challenging, and Laskowitz remarked that the absence of large pharmaceutical company representation at this workshop reflects that challenge. After many efforts to move research forward by appealing directly to pharmaceutical companies, he discovered that the regulatory pathway is not merely a process, but rather it is a strategy to stimulate commercial interest in a product. He remarked that the Clinical and Translational Science Award grant at Duke University provided Laskowitz with access to a regulatory expert who explained the importance of identifying the therapeutic outlet. Assisting Laskowitz in mapping out a strategy, the expert explained how the therapeutic outlet feeds into FDA Investigational New Drug studies and drives the ultimate therapy forward.

These commercialization considerations reflect a broad skill set involving communication with companies, IP issues, and regulatory navigation that is not traditionally taught in medical school, said Laskowitz. Academics can use resources to fill such knowledge gaps and advance stalled, yet promising, developing therapies. He noted that he is working on an orphan drug that cleared phase I and has stalled at phase II owing to lack of funding. Laskowitz remarked that TBI is a challenging area for fundraising because it is considered high risk, and companies are often not interested in the area. In contrast, a checkpoint inhibitor in oncology might be more to likely to draw interest from numerous pharmaceutical companies, but this will not be the case for TBI until risk is minimized, he stated.

For many small companies, waiting several years while attempting to obtain National Institutes of Health (NIH) grant funding is not a highly viable commercialization strategy, said Laskowitz. Stating his gratitude for the infrastructure created via efforts such as the Transforming Research and Clinical Knowledge in TBI study, he credited such research for allowing drugs he is developing to reach phase II trials. Laskowitz added that TBI trials are incredibly challenging and require an infrastructure of harmonized MRI protocols and biomarker protocols that may not be readily available to many academic investigators and small businesses.

Selective Calpain-2 Inhibitors

Michel Baudry, founder and chief scientific advisor at NeurAegis and professor at the Western University of Health Sciences, stated that he has been researching the enzyme calpain for 45 years. In 1986, his postdoctoral advisor started Cortex Pharmaceuticals to develop calpain inhibitors for the treatment of neurodegeneration, but this effort did not yield effective therapies. Four decades of research have led Baudry to understand that previous trials of calpain inhibitors failed because selective calpain-2 inhibitors are needed. In 2016, he started a company, NeurAegis, to develop selective calpain-2 inhibitors as neuroprotective drugs. Baudry explained that selective calpain-2 inhibitors block a pathway that leads to neuronal damage and death. A blood biomarker has been identified that reflects brain activation of this pathway, making it possible to test whether the drug is working as intended within 24 hours of administration. The blood biomarker remains elevated for several days, which enables researchers to determine the appropriate dose, the time window for initiating treatment, and the treatment duration. Baudry was optimistic that clinical trials would begin in 2025.

Baudry outlined several challenges he has faced in developing this therapy. Given that research on preventing neurodegeneration with calpain inhibitors began in 1986, convincing those within and beyond the scientific community that his company has identified a novel, unique pathway that leads to neuronal damage and can be blocked by selective calpain-2 inhibitors has been incredibly difficult. He was unsuccessful in obtaining NIH support, but he secured funding for his preclinical work from the U.S. Department of Defense (DoD). He remarked that pharmaceutical companies and venture capitalists have lost interest in supporting TBI work because all potential TBI treatments have failed. Describing the struggle of securing funding for clinical trials for a novel drug treatment that he believes will have a significant effect, Baudry stated that innovative trial design could be a mechanism for decreasing the cost of clinical trials while increasing the likelihood of success. Increased financial support for TBI treatment would also enhance the probability of finding an effective treatment, he concluded.

Nonpharmacological Therapies and Sex Differences in TBI

Maheen Mausoof Adamson, clinical professor at the Stanford School of Medicine, director of research at the Women’s Operational Military Exposure Network Center of Excellence, and senior scientist for rehabilitation services at the U.S. Department of Veterans Affairs (VA) Palo Alto Healthcare System, described working on a repetitive transcranial magnetic stimulation (rTMS) TBI VA grant approximately 20 years ago

and the difficulty she encountered in obtaining FDA approval for its use in mild and moderate TBI patients, an approval process that required EEG and MRI evidence to proceed. Noting progress in the field, she stated that the Stanford neuromodulation therapy trial is currently demonstrating approximately 90 percent remission from major depressive disorder, and this technique is now also being used in TBI patients in the VA system. This accelerated treatment uses functional neuronavigation, and Adamson noted that she is starting a multisite trial with this therapy with David Brody of the Uniformed Services Health Science University.

Adamson is particularly interested in sex differences in brain injury and other health conditions. Conducting a matched-sample analysis to examine behavioral symptomology differences within the polytrauma population, Adamson found that women experienced more psychiatric symptoms, more vertigo, and—despite being more educated—more unemployment than their male counterparts. Prior research has reported that MRI-based analysis of cross-sectional cortical thickness in men and women with and without brain injury revealed that women’s cortical thickness did not return to original size to the same degree as that of men. She emphasized that these differences are unrelated to treatment. Noting presentations during the workshop on TBI related to blast exposure, sports, and falls, Adamson underscored that TBI also occurs within the polytrauma population and among those experiencing violence and sexual trauma, an issue that is understudied. Thus, associated post-traumatic stress disorder, depression, and other symptoms can manifest differently in women than in men. She questioned whether researchers have considered how to integrate such aspects into treatment protocols.

In 2020, Adamson coauthored a review paper of nine studies of rTMS TBI treatment, many with relatively small sample sizes (Oberman et al., 2020). These findings highlight the need for personalized, precision treatment for people with TBI, depression, and pain. She described cortical differences between men and women and stated that recent research has revealed that women who sustain TBI in the luteal phase of their menstrual cycle, when progesterone is high, have worse outcomes than women in their follicle phase. Furthermore, the realizable electric field in women shows differences in the cortex and in gray and white matter from that in men. Adamson emphasized that female hormones play a role in mechanisms of TBI and they must be taken into account during any treatment offered.

Efforts are needed to incorporate considerations of these factors into studies and increase sample sizes to optimize treatment for women, particularly those who have survived intense environmental exposures, said Adamson. She emphasized that the short duration of Stanford neuromodulation therapy trials is important to making participation feasible for female participants, given that 80–90 percent of caregivers worldwide are women

and dedicating time to attending multiple and sequential appointments can be difficult. On the other hand, shorter and more innovative therapy trials can be more expensive—and thus, less accessible in a different way—as a result of the higher costs of the cutting-edge technologies being deployed (e.g., fMRI scans to enhance localization of rTMS neurostimulation sites). Adamson pointed out a knowledge gap among research scientists about specialized nomenclature in industry biomedical development relating to up-scaling from clinical trials to commercialization. She stated that additional education in industry development concepts and approaches (i.e., minimum viable product) could help researchers understand what’s needed to commercialize new therapies.

Biomarkers and Biologics

William Haskins, chief executive officer and cofounder of Gryphon Bio and Owl Therapeutics, shared that 2 decades ago he applied his Ph.D. in proteomics to biomarker discovery in TBI before shifting his focus to drug development. Approximately 6 years ago, he and Kevin Wang founded Gryphon Bio to apply their combined learnings on developing novel diagnostics and therapeutics to improve patients’ lives. Drawing on his perspective as a chemist, designing the best drugs possible for TBI patients is incredibly important, Haskins said. Dramatic improvements in diagnostic and therapeutic technologies for drug development have been made over the past 20 years and include advances in biomarker discovery and testing platforms and new therapeutic modalities, including biologics such as monoclonal, bispecific, and trispecific antibodies.

During his career, Haskins has led drug development teams for antibodies, bispecific antibodies, antibody drug conjugates, and gene therapies. Attention to repurposing drugs approved for other indications for applicability to TBI is warranted, he said, and development of drugs specifically designed for TBI is also needed. Exploring the application of artificial intelligence (AI) to drug design, Haskins noted that Google DeepMind developed the AlphaFold AI system to predict three-dimensional structures of proteins. Generative AI and large language models are now being used to virtually dock libraries of small molecules to these theoretical protein structures and select lead compounds. Haskins expressed his excitement about current opportunities to bring new therapeutic modalities to TBI and, combined with new diagnostic tools for blood biomarker testing, improve the safety and efficacy of novel medicines for TBI patients.

DISCUSSION

Approaches to TBI Heterogeneity and Inclusion Criteria

Uzma Samadani, founder of Oculogica and neurosurgeon at the Minneapolis Veterans Administration Medical Center, expressed hope that innovative trial design will be an effective mechanism to address the heterogeneity of TBI and avoid grouping patients with varied pathophysiology together, a practice that contributes to failed trials. She remarked that clinical trials often classify brain injury with Glasgow Coma Scale (GCS) scores and perhaps positive CT scans and use the Glasgow Outcome Scale-Extended (GOSE) to measure outcomes. Samadani asked how to facilitate a shift toward more refined inclusion criteria and outcome measures. Laskowitz agreed that GCS scores have limited usefulness in stratifying patients and commented that a new series of trials will include multiple inclusion criteria, including neurologic score, initial radiographic data, and a tissue biomarker for the degree of tissue injury.

Lewis highlighted an opportunity regarding biomarkers intended to be predictive of therapeutic response. Using an example from stroke research, he described the use of imaging biomarkers available at baseline that indicated the initial ischemic stroke volume or hemorrhage location. Given that basic imaging is often performed before making determinations about treatment, it is possible to capture imaging biomarkers naturally during clinical care. He clarified that an experimental setting is not required to qualify a predictive biomarker. For instance, researchers could enroll a broad population of patients with TBI and perform baseline biomarker assessment. Even if the biomarker assay requires time to run in a research laboratory, patients could be treated in the meantime, and researchers could meld the outcome and response data with the biomarker data once it becomes available. Measuring the biomarker at baseline aids in assessing the effectiveness of the therapy and assessing the heterogeneity of the treatment effect. Even if a biomarker does not yet have feasible clinical application because of the time required to obtain results, it can still be evaluated. Lewis explained that a biomarker can be qualified as a predictive biomarker before discovering how to make the result actionable in a clinical setting.

Baudry emphasized that multiple mechanisms are known to be engaged after TBI, including neuronal death, axonal damage, inflammation, and a number of molecular events. Heterogeneous outcomes do not negate that a core set of mechanisms are generally involved after TBI. The target he and his colleagues have identified, selective calpain-2, is useful because it becomes engaged very early after trauma, remains activated for several days, and a blood biomarker that reflects brain activation of this pathway is identified, he said. The ability to monitor the level of this blood biomarker

enables his team to determine whether a selected drug works as intended. He anticipates being able to determine both the dose required for complete prevention of biomarker increase and the duration needed to maintain prevention of pathway activation. In animals, treatment lasting 1 week post-trauma results in optimal neuroprotection, he noted. Potentially, a different mechanism, such as brain inflammation, could become predominant after 1 week, in which case an anti-inflammatory drug would be important to prevent brain inflammation. Baudry said that use of blood biomarkers will be important in determining TBI treatments moving forward.

Payer Requirements for Trial Design

Leslie Wise, chief executive officer at EvidenceMatters, asked how a biomarker that is not yet standard of care—even one cleared by FDA—can be used to determine treatment that will then be reimbursed by payers. In her experience, payers want to see results from classic randomized controlled trials (RCT) to prove efficacy and legitimize a causal relationship, even if the trials are conducted with small populations. Stating her appreciation for adaptive trial designs, Wise responded that payers first want results from RCTs before expanding to adaptive designs.

Lewis clarified that every innovative trial design he discussed is an RCT; they are all randomized designs. Lewis replied that the rules of inference and determining causality do not differ between (a) a traditional single intervention, (b) randomized, fixed sample size trials, and (c) adaptive trials. Rules are in place for fixed sample and adaptive trial design types regarding randomization, baseline measurements, and accounting for various causal pathways to patient outcomes. He noted that a traditional approach to clinical trials often involves a 2-year startup period, concluding a trial, and waiting 2 years before starting the next trial, and that this process is inherently inefficient. Adaptive trial designs provide mechanisms to mitigate some of these sources of inefficiency by allowing modifications to ongoing trial procedures based on interim data analysis (FDA, 2019; Kaizer et al., 2023).⁵ While some innovative trial designs have greater inferential subtlety in terms of how they achieve efficiencies, the rules for demonstrating causality remain constant, he remarked.

No therapeutic trials in TBI have succeeded, and extraordinary clinical need exists, which should fuel a desire throughout the field to be as

___________________

⁵ In 2019, the FDA provided guidance on the use of adaptive designs in clinical trials for drugs and biologics (FDA, 2019). These adaptive trial mechanisms may include terminating a trial before completion due to futility or efficacy, re-estimating the sample size needed to achieve statistical power, enriching the target population of enrolled participants, selecting across multiple treatment arms, revising allocation ratios used for randomization, and identifying the most appropriate endpoint measure (Kaizer et al., 2023).

efficient as possible at each stage of development, Lewis said. He added that methods of increasing efficiency are also available during the early phase of development. The basic ground rules of the evidence and the statistical characteristics of that evidence to demonstrate patient population and estimated treatment effect are known, he reiterated. Tremendous regulatory history around evaluating diagnostic and therapeutic products is established, Lewis added.

Haskins remarked that Gryphon Bio approaches trials as multistep processes. Currently, his team is developing biomarkers in an effort to improve the odds of clinical trials succeeding. He stated that they are not designing biomarkers to be used as companion diagnostics that would require payer approval once the patient is treated; that would constitute a third step after biomarker identification and testing the therapeutic. Haskins commented that different questions are at play in these steps. Laskowitz maintained that researchers need to differentiate a biomarker to ascertain the effect or optimal dose of a therapeutic in the context of a clinical trial versus for a companion diagnostic, such as one indicating a large vessel occlusion needing thrombectomy.

Trial Considerations Regarding TBI Severity Level

Kristy Arbogast, R. Anderson Pew Distinguished Chair of Pediatrics at Children’s Hospital of Philadelphia, noted that most of the trials discussed so far were for situations in which people have moderate and severe TBI and asked if trials for people with milder forms of TBI involve unique challenges or opportunities. Baudry remarked that the selective calpain-2 pathway being studied is the same for both mild and severe TBI, and thus the selective calpain-2 inhibitor drug being developed is neuroprotective for all severity levels. Underscoring Ely’s discussion of his daughter’s symptoms and the length of time before they emerged, Adamson highlighted the need to determine how to acquire data from patients experiencing chronic symptoms, as no biomarker is yet available to indicate TBI years after injury. She remarked that GCS scores and self-reported, evidence-based forms on TBI can be used to this end. Researchers can also use addenda to gather data on more focused issues related to intimate partner violence, military sexual trauma, or other aspects. Adamson stated that pulling various components together and understanding the variables, including potentially social determinants of health, can make RCTs more powerful.

Lewis stated that mathematics does not vary from disease to disease. Likewise, the process of making inferences about treatment effects does not change for mild TBI versus severe TBI. However, time to information is very important in innovative trial design, he specified. Treatment intended to mitigate or prevent TBI ramifications that accrue over a decade is a scenario not unlike dietary interventions for preventive cardiology or vaccine

trials that feature lengthy time to information. Such situations offer fewer and different opportunities for efficiency, Lewis explained, than those in which the symptom targeted for intervention becomes available in a short time scale relative to the duration required to conduct the trial. Regarding the difficulty in enrolling some populations in trials, Lewis remarked that in these cases, as much information as possible should be collected from each person enrolled.

Frederick Korley, professor at the University of Michigan, noted that cardiology therapies were typically developed in more severe cases and then over time were tested for effectiveness on lower severity levels. However, given that TBI pathophysiologies can differ, Korley asked whether an initial focus on patients with more clear-cut evidence of structural injury and structural pathology could lead to more successful trial designs in the future. Lewis replied that the ideal selection of injury or illness severity is that which enables more efficient learning. He stated that this determination involves understanding the pathophysiology, the severity level most likely to demonstrate successful effect, and the response of the population in terms of the outcome measure. Remarking that the ideal severity level for initial research is not likely broadly generalizable from one disease or condition to another, Lewis said that this determination can be informed by the current work developing TBI outcome measures and biomarkers and in understanding patient population recovery trajectories.

Phase II Design and Endpoints

Korley commented on the difficulty in selecting the best outcome measure when designing phase II studies in TBI, noting the importance of using surrogate endpoints to show evidence of efficacy. For example, a choice may involve using a 1-month Glasgow Outcome Scale-Extended (GOSE) assessment, a 6-month GOSE assessment, a given biomarker test, and/or other measures. Lewis replied that the purpose of a phase II program is to determine whether and how to run a phase III program. A well-designed phase II program features multiple steps at which researchers should assess whether to continue or to stop testing a therapy, he explained. For example, early phase II may involve proof of concept and target engagement, and researchers should accept when a treatment fails and return to the proverbial drawing board rather than assume the trial design is to blame, Lewis stated. However, the best result of a phase II study is a quantitative prediction of probability of phase III success, and late phase II work necessitates enrolling the same patients with the same treatment and measuring the same outcome, he offered. Thus, the concept of phase II endpoints is fundamentally flawed from the perspective that the purpose of phase II is to indicate whether and how to run phase III, said Lewis. He remarked

that when a treatment fails when moving from phase II to phase III, this indicates flaws in the phase II design.

Lewis stated that early neurologic outcome is highly predictive of later-term neurologic outcome on a group level, though not on an individual level. However, many trial designs do not take advantage of these predictors. Providing an extreme example, Lewis remarked that death at day 1 is extremely predictive of death at 6 months. In designing clinical trials in which the registration endpoint is a 6-month neurologic outcome, the Modified Rankin Scale can be used at hospital discharge and at 1 month to reduce the uncertainty regarding whether the individual patient will return for the 6-month assessment. Emphasizing that the correlation between proximate endpoints and longer-term endpoints can be determined—but not assumed—Lewis remarked that using proximate endpoints can enable the phase II trial to efficiently indicate whether the phase III trial should be conducted.

Neuroinflammation Interventions

A participant commented that a major cause of secondary injury in TBI is neuroinflammation and asked about the use of nutritional interventions or neurosteroids to help address neuroinflammation. Diaz-Arrastia remarked on emerging research in this area, including studies on omega-3 fatty acids and B vitamins. The wide availability and lack of label claims for these therapies pose a challenge in recruitment and other aspects of research, he said. Noting the potential for deeper understanding of therapies to address neuroinflammation, Diaz-Arrastia stated that the postacute or chronic setting could be a focus of this research, as these interventions are perhaps more applicable to these populations experiencing ongoing symptoms.

This page intentionally left blank.