Investing in Brain Cancer - Part 1
Scope of the Challenge
Series Executive Summary
Many investors (especially institutional capital) view brain cancer as a No Fly Zone. The prevailing view is that investments in the area are tantamount to lighting a pile of money on fire and watching it burn. Why?
In a Series of posts on MissionGBM over the next two months, we will examine the question and unpack the various perspectives. The Series will be organized as follows:
Part 1 – Scope of the Challenge (below in this post)
We invite the reader to follow the Series, and to ponder the challenges and opportunities that exist. The ability to craft a solution begins with understanding the problem.
Multiple Overlapping Challenges – All Difficult
While nothing is more complex than human biology, the brain cancer therapeutic area (both primary and metastatic brain cancer) presents several challenges that are especially difficult.
The Brain is a Protected Environment. When we invest to support research and companies focused on ultimately producing new therapeutic agents, we begin by assessing the barriers to delivering a medicine or medical device action to the site of therapeutic need. Because the brain is such a vital organ, nature has evolved multiple, overlapping structures and mechanisms to protect it from everyday insults like infections and toxins. While these features are very effective at protecting the brain, they also function as strong barriers to delivering a therapeutic agent to the brain. The Blood Brain Barrier (BBB) is a neurological drug developer’s nemesis as it is must be transited in order to get a drug to the site of therapeutic action. The BBB is particularly vexing for macromolecular agents like proteins, gene therapies and cell therapies. In fact, the past several decades of Alzheimer’s Disease research is largely a case study of the BBB’s ability to reject all but a tiny fraction of antibodies and other HMW drugs administered systemically. Even for LMW agents, transporting sufficient drug across the BBB and keeping it at the therapeutic site for the necessary duration to affect clinical benefit can be challenging.
Brain Cancers Can Be Quite Heterogeneous. Primary High Grade Gliomas (HGG) are a notoriously heterogenous clinical presentation. Single cell analysis and sequencing of resected tumor tissue typically shows a spectrum of molecular and genetic profiles associated with the cancer cells. Applying therapeutic pressure to one cell profile often serves to provide a competitive survival advantage to other cells with differing profiles. Thus, monotherapies and single target approaches have little chance of being effective, which is what the clinical literature illustrates. Thus, treatment strategies need to emphasize combination approaches that can simultaneously hit multiple anti-cancer mechanisms. We will expand on this topic in Part 2 of this Series.
The situation is a bit different with Low Grade Gliomas (LGGs) and Brain Metastases from a primary tumor site outside of the brain (Brain Mets). These two categories can display more homogeneous tumor profiles with a manageable handful of therapeutic targets.
Brain Cancers Often Lack Validated or Druggable Therapeutic Targets. In addition to being heterogeneous and resident behind the BBB, HGGs tend to have (i) low Tumor Mutational Burden; (ii) Micro-Satellite stability; (iii) high degree of Immunosuppression in the Tumor Micro-Environment (TME); (iv) targets which are considered to be “undruggable” (e.g. transcription factors, dynamic biomolecular structures lacking druggable features); and (v) driver oncogenes (EGFR, CDK4/6, MDM2/4, etc.) that primarily exist on extra-chromosomal DNA (ecDNA) and are highly dynamic and fungible.
Again, the situation can be somewhat different with LGGs and Brain Mets.
HGGs are Aggressive and Invasive. HGGs can grow very quickly and aggressively invade surrounding brain tissue. When this happens, the clinical situation changes rapidly. Clinical timelines can be short, which serves as a real challenge to enrolling patients in clinical studies. Short clinical timelines also have a knock-on effect with respect to HGG clinical trials. The field tends to skew Clinical Heavy, but Science Lite. There are simply far too many clinical trials in HGGs for which there is little or no scientific rationale.
We get it: “The patients can’t wait. Something must be done!” We encounter this mantra every day at MissionGBM.
Nonetheless, we do have to question the validity (or ethics?) of subjecting HGG patients to trials in which (i) the therapeutic agent has demonstrated little or no treatment effect in pre-clinical studies; (ii) the agent is known to produce Serious Adverse Events (SAEs; Grade 3+); and (iii) data is lacking to show that the agent gets across the BBB in sufficient concentration or duration to yield any therapeutic benefit.
Pre-Clinical Models are Inadequate and Do Not Translate Well. Pre-clinical models need to be reproducible and translatable to the human clinical situation in order to have real utility. At present, this is not the case because the models rely on implantation of unimodal cell lines or patient-derived xenografts to create model brain tumors. The simplified model brain tumors that are produced are not sufficiently heterogenous or representative of human HGG presentations. Monotherapies often appear to work in animal models, but fail miserably in human studies due to the relative homogeneity of the model tumor among other factors. We have all cured cancer in mice! Some of the better models involve spontaneous brain tumor development in dogs (French Bulldogs, in particular), but the spontaneous and unpredictable occurrence of the brain tumors means that it is hard to plan or resource pre-clinical studies.
Regulatory Framework is Ill-Suited for HGG Therapeutics Development. Modern clinical science and regulatory framework is justifiably built upon a foundation of Randomized Control Trials (RCTs). The cleanest possible RCT is one that tests a single therapeutic agent in the Treatment Arm versus a Placebo (or SoC baseline) in the Control Arm within Subjects who are randomly assigned to each Arm. Further, Gold Standard trials also employ a blinding protocol so that Subjects, Caregivers, Investigators and Sponsors do not know which Subjects are receiving the Treatment under test and which Subject are receiving Placebo. Worldwide regulatory agencies strongly encourage RCTs, and Sponsors know it. Diverge from this paradigm at your peril.
In addition, Endpoints matter in clinical trials, both from a trial design and the regulatory point of view. Oncology trials have a variety of Endpoints with median Progression Free Survival (mPFS) and median Overall Survival (mOS) being two of the most frequently occurring. At present, the majority of regulatory agencies are insistent upon an mOS Endpoint for HGG (and other brain cancer) trials, which poses a real challenge for RCTs in HGGs. Why? It is a simple math problem.
HGGs are rare diseases. Each year in the US, there are approximately 25,000 cases of HGGs diagnosed (about 12-15,000 are GBM). The cases are scattered over several dozen major brain tumor centers as well as even more community hospitals. Identifying Subjects (i) that meet the trial Inclusion and Exclusion Criteria; (ii) are willing to consent to enroll in a trial (often with lengthy travel involved); and (iii) will not break protocol or withdraw during the trial is incredibly difficult. Add on top the aggressive clinical timelines of HGGs and the competition from multiple competing trials (many of which have no scientific justification), and the number of available Subjects winnows rapidly. But that is just the warm up band.
The BIG issues are:
Current regulatory framework does not accommodate Combination trials very well. HGGs (like most cancers) respond best to Combination trials involving multiple points of attack on tumor drivers with multiple therapeutic agents. Monotherapies rarely work against an mOS Endpoint, particularly in HGGs. In order to rigorously evaluate each therapeutic agent, a factorial trial design testing the Combination versus the Component Parts is needed. However, the number of Subjects required for such testing mushrooms into a total number that cannot be accommodated with the limited number of consenting and eligible Subjects available in the HGG cohort. Expanding the trials on a global basis does very little to mitigate the problem, and serves to magnify quality control, logistics and resource (human, material, capital) issues. Adaptive trial designs also struggle to handle Combination protocols. The obvious proposal is to allow the trial framework to be more flexible to encompass de novo Combination trials in which the combination of therapeutic agents is approvable on a Combination label without a requirement for factorial design in the pivotal trial(s). To affect such a transition will require regulatory agencies to flex away from a long history and cultural tradition that emphasizes the current RCT factorial approach. The matter gets discussed at just about every regulatory agency HGG workshop, but little progress towards flexibility has been observed to date. Talk about “The patients can’t wait!”
Current techniques to assess Treatment Response do not translate well to an approvable Endpoint of mOS. Solid tumor oncology trials tend to measure Treatment Responses via a mixture of imaging techniques and tissue biopsies against a pre-specified set of Objective Response (OR) criteria (RANO in brain cancers; RECIST in most solid tumors). Many therapies for non-brain cancers have been approved on the basis of mPFS instead of the more rigorous standard of mOS. However, mPFS is very difficult to reliably assess in brain cancers due to the modest sensitivity and reproducibility of brain imaging techniques, and significant discordance among multiple Neuro-Radiology expert panels reading the same blinded images. Moreover, serial biopsies of brain tumors are not considered feasible. Regulatory agencies know this, and tend to insist upon mOS as the approvable Endpoint. While mOS is definitive, it takes longer to assess, which means more capital is required to conduct the trial, and more Subjects need to be recruited and maintained on the study in order to satisfy the trial statistical power calculations in view of the inevitable drop out of Subjects during the trial.
For the interested reader, the textbook example of a regulatory agency accepting mPFS as the approvable Endpoint in an HGG indication only to later discover that the drug has no significant effect on mOS is the bevacizumab (Avastin) case. This case involved a single arm, open label design. The bevacizumab case made a material contribution to the coining of the term “Pseudo-Response”, which is used to describe apparent ORs on imaging that are actually temporary artifacts caused by the action of the agent on the tumor vasculature and contrast agent distribution (see here and here). Oof! Regulatory agencies vigorously frown upon therapeutic agents that get approved in the structured environment of a pivotal clinical trial only to later reveal no benefit in real world use or followup confirmatory trials. Especially true when later data indicates that the technique used to evaluate the Endpoint proved to be misleading or flawed.