Summary
- Cassava's pivotal Phase 3 studies roll on and are expected to be fully enrolled by the end of 2023 - an admirable achievement.
- Meanwhile, the latest Phase 2 Open Label knocked the share price.
- The data releases from these studies have gotten progressively less encouraging - that may be because they dealt with milder Alzheimer's patients first.
- Cassava's thesis around Simufilam has been subject to much criticism - but we can seemingly make a case the drug may work in patients with early-stage dementia or AD.
- The next catalyst will arrive in Q3 when Cognition Maintenance Study data is announced. The more transparent the data, the better for everyone concerned.
Investment Overview
I first covered Cassava Sciences ( SAVA ) for Seeking Alpha over 3 years ago, in January 2020 , shortly after the company reported positive data from a Phase 2a study of its new drug, PTI-125, now known as Simufilam. I wrote:
PTI-125 is certainly a radically different approach to treating Alzheimer's, a disease which has no known cure. It is a departure from currently available treatments such as cholinesterase and beta amyloid inhibitors.
The theory goes that PTI-125 targets Filamin A ("FLNA"), a "scaffolding" protein which brings different proteins together and ensures they are functioning and interacting as they should do. A toxic form of collapsed FLNA is often found within the brains of Alzheimer's patients, disrupting their interactions, which leads to neurodegeneration and brain inflammation. PTI-125 is a small molecule drug that restores the normal shape of FLNA in the brain.
Cassava, its drug Simufilam, its shareholders and its management, and those looking to "short sell" the stock (borrowing shares at an inflated price and selling them, then waiting for the price to decline before purchasing stock and returning it to the broker, realising a substantial profit in the process) - have been on a roller-coaster ride ever since.
A Brief History Of Simufilam
To briefly summarise, there was a failed Phase 2b study later in 2020, followed by the re-evaluation of that data in a different lab (a lab that happened to be run by long term Cassava collaborator Dr Hoau-Yan Wang) that came to a different conclusion - that the data were actually positive, achieving statistically significant improvements across a panel of validated biomarkers.
There was the initiation of a Phase 2 open label (no placebo arm) study which showed , after 6 months of treatment, that patients' cognition improved by 1.6 points on the "gold standard" Alzheimer's Disease Assessment Scale-Cognitive subscale ("ADAS-COG) scoring scale. Simufilam also improved dementia-related behavior, such as anxiety, delusions and agitation, by 1.3 points on the Neuropsychiatric Inventory.
The results were unprecedented - patients' cognition actually seemed to be getting better when using Simufilam, as opposed to declining at a slower pace than placebo - something that no drug had been able to demonstrate decisively in more than 20 years.
Biogen's Aduhelm had snuck over the FDA approval line in June 2020 based on some somewhat dubious data that suggested a mild improvement over placebo in cognitive decline, but its efficacy / safety profile was simply not persuasive enough to persuade the Centers for Medicaid and Medicare to provide reimbursement in a real-world setting.
Meanwhile, Cassava's share price began to sky-rocket - it became the best performing stock across any sector of the stock market, rising to a high of >$120 and up >8,000% prior to the announcement of 9-month data from the Phase 2 Open Label.
The Backlash Arrives
The 9-month open label data suggested that cognition scores, based on ADAS-Cog 11 scoring, had now risen by 3 points overall . Bear in mind that a meta-analysis of 20k patients has reported an average 5.5 point decline in placebo groups over a 12 month period.
On the heels of this seemingly groundbreaking data came the backlash. A Citizen's Petition was filed with the FDA by a law firm apparently representing a group of whistleblowers and scientists who may also have been short-sellers of Cassava stock.
The petition raised concerns over whether Cassava may have manipulated some of its early research data in order to secure an Investigational New Drug ("IND") approval from the FDA, allowing the company to conduct in-human clinical studies.
Furthermore, it accused Dr Wang's lab of doctoring the Phase 2b data to suggest the study had been successful after an independent lab had declared it a failure. Scientists began to weigh in, suggesting that Cassava's Filamin-A thesis lacked merit, and querying why no other company or researchers had pursued this thesis - implying they hadn't because it was a flawed thesis.
If the goal of the petition was to drag down Cassava's share price - it worked. By the beginning of January 2022,, Cassava's stock price had fallen back below $40 - by May 2022, it had fallen <$20.
Cassava Handles Its Detractors Amid FDA Support For Pivotal Studies
It has to be said that Cassava management has handled the situation reasonably well. The company formally responded to the Citizen's Petition, addressing most of the points raised in a press release, and later launched its own lawsuit against "certain individuals who executed a "short and distort" campaign against the Company."
Cassava has denied claims made in a July 2022 Reuters article that the Department of Justice ("DoJ") had launched a criminal probe into management's activities, and also dealt with accusations that management would realise massive financial windfalls if they could keep Cassava's share price artificially high. In a statement, management commented:
It is worth nothing that no officer or director of Cassava Sciences has sold Company stock in over a decade. In contrast, short sellers, some of whom have been pursuing an unprecedented attack on Cassava Sciences, have reportedly made over $100,000,000 shorting SAVA stock.
Meanwhile, Cassava apparently retained support of the FDA, who declined to follow the Citizen's Petition's demands (e.g. that all studies of Simufilam be halted) and agreed protocols with Cassava for 2 pivotal Phase 3 studies - named rethinkALZ and refocusALZ - details are shown below.
Cassava's twin pivotal Simufilam studies (Cassava investor presentation)
In a press release earlier this month Cassava updated on the progress of these 2 studies. Apparently, 953 patients have been enrolled across both studies and both studies are now >50% enrolled. Cassava says it is "actively recruiting Alzheimer's patients in over 100 clinical sites in the United States, Canada, Puerto Rico, South Korea and Australia", and adds that its goal is to achieve full enrolment for both studies before the end of the year.
Whatever your stance on Cassava may be, this can be considered good progress, although it does not necessarily mean we will see any data from rethink or refocus this year, or possibly even next year. The critical data readouts may not arrive until 2025.
More Open Label Data Announced - Efficacy Apparently Demonstrated In Mild-to-Moderate Patients
In late January, Cassava announced full data from its Phase 2 Open Label Study, from >200 patients with mild-to-moderate Alzheimer's disease, defined as a Mini Mental State Examination score of 16-26 (out of 30). The release begins by presenting some headline figures from the entire study population:
Top-line Results - mean scores, baseline to month 12 ( lower is better, except for MMSE) :
- ADAS-Cog11 scores changed from 19.1 (±9.2) to 19.6 (±13.3)
- MMSE scores changed from 21.5 (±3.6) to 20.2 (±6.4)
- NPI10 scores changed from 3.2 (±4.6) to 2.9 (±4.6)
- GDS scores changed from 1.8 (±1.8) to 1.4 (±1.9)
The figure that likely jumps out is the ADAS-Cog11 score - the overall mean patient score no longer reflects an improvement, rather a decline of 0.5 points.
So, was Cassava pulling the wool over investors eyes with its early data - suggesting patient's cognition was improving, as opposed to declining?
Not quite - as SA Contributor Lane Simonian has postulated , it seems Cassava began its study by treating patients showing the mildest degree of cognitive decline first, hence the apparently excellent 3.2 point improvement in ADAS-COG scoring amongst the first 50 patients treated after 12 months.
When Cassava reported data from the first 100 patients, the overall improvement fell to 1.5 points, and as above, when data from all 216 patients were collected, there was an overall mean 0.5 point decline. Amongst the first 50 patients, 68% improved their ADAS-COG score, and amongst the first 100 patients, 63% improved their score.
In the final analysis, Cassava says that 47% of patients improved their scores with an average improvement of 4.7 points, whilst 23% declined <5 points, with an average decline of 2.5 points.
In its press release Cassava shares 2 tables, the first showing that versus historical placebo arm data collected from prior Alzheimer's studies, treatment with Simufilam results in a markedly smaller decline in ADAS-COG scoring - in patients with mild and moderate forms of the disease .
model of historical placebo arm declines vs simufilam - mild to moderate patients (Cassava press release)
Cassava shares a second table showing that in early and mild disease only , patients treated with Simufilam show a markedly lower decline than historical placebo arms.
As ever, however, these results almost raise more questions than they answer. Cassava does not share any p-values confirming a statistically result, and includes the following disclaimer in its press release:
Data results from our open-label safety study do not constitute, and should not be interpreted as, regulatory evidence of safety or efficacy for simufilam in Alzheimer's disease. Rigorous evidence for drug safety and efficacy is derived from one or more large, randomized, placebo-controlled studies. The open-label design and size of this study may introduce clinical or statistical bias or may generate results that may not fully distinguish between drug effects and random variation.
It is possible to make the argument that treating patients showing signs of early cognitive decline with Simufilam may be a good idea, based on the levels of improvement shown, but then again, these early stage patients may not go on to develop full-blown Alzheimer's, and the study results may just be a reflection of the fact levels and rates of cognitive decline fluctuate regardless of the treatment. Alzheimer's tests are highly subjective, after all.
Ultimately, it could be said that early results from this open label study created a raging debate about the efficacy or otherwise of Simufilam that resulted in incredible peaks and troughs in the company's share price, even while this study was never powered or intended to fundamentally answer such a question.
The Shifting Sands Of The Treatment Landscape - Updates On Other Projects
As most followers of Alzheimer's treatments will likely already know, the dominant thesis in the field has been the removal of amyloid beta - a kind of plaque that accumulates in the brains of patients with Alzheimer's.
That such plaque builds up is undeniable, but is it true to say that removing it can help improve a patient's condition, or rather stabilise it by slowing their rate of cognitive decline? Before answering that question it is worth pointing out that the process of amyloid beta removal is fraught with safety concerns.
Biogen ( BIIB ) and its Japanese Pharma partner Eisai won approval for a second Alzheimer's drug (after Aduhelm) in January, based on a study that - after more than 2 decades of trying - finally seemed to establish a direct correlation between amyloid beta removal and a slowing of cognitive decline in Alzheimer's patients.
As I wrote in a post on Biogen back in September last year:
The headline figure was that patients in the Lecanemab arm of the clinical study - which enrolled 1,795 people with early onset Alzheimer's - saw a 27% greater reduction in CDR-SB scoring - which assesses cognitive and functional performance in six areas: memory, orientation, judgment and problem solving, community affairs, home and hobbies, and personal care - vs. placebo after 18 months.
That represents a statistically significant treatment difference in the score change of -0.45 (p=0.00005) in the analysis of Intent-to-treat ("ITT") population.
Just like its predecessor Aduhelm, however, Leqembi can lead to brain swelling and brain bleeding in some patients - 3 patients in the pivotal trial mentioned above reportedly died from brain hemorrhages.
The Alzheimer's community now faces a wait to see if, this time, the CMS is prepared to provide reimbursement for a drug that shows marginal efficacy but comes with a worrying safety profile.
Meanwhile, earlier this month, Eli Lilly's ( LLY ) AD drug Donanemab - which has been progressing through a potentially pivotal Phase 2 study - was handed a Complete Response Letter ("CRL") from the FDA in response to the Pharma's attempt to secure an accelerated approval.
The FDA told Lilly that it wanted to see data from at least 100 patients who had received a minimum 12m of treatment to make sure the safety profile is satisfactory. Lilly's data included >100 patients, but many had retired early from the study after satisfactory clearance of their amyloid beta.
Since Donanemab has a similar mechanism of action and target to Leqembi, the agency wants to be confident instances of amyloid-related imaging abnormalities ("ARIAs") - or brain swelling - are manageably low and not life-threatening.
Meanwhile, the likes of Prothena Corporation ( PRTA ) Anavex Sciences ( AVXL ), BioVie ( BIVI ), and Annovis Bio ( ANVS ) are amongst a handful of companies claiming to have successful Alzheimer's therapies in development, with differentiated mechanisms of actions ranging to targeting inflammatory pathways to inhibition of the "SIGMAR" receptor. Anavex, whose candidate targets the latter, has even reported positive data from a >500 patient study, but similarly to Cassava, doubts around the integrity and presentation of the data have held the share price back.
Conclusion - Cassava Continues Its Clinical Progress As Data Hints At An Approval Shot In Early Stage Dementia/Alzheimer's
It is quite well known that prior to turning his attention to Simufilam, Cassava CEO Remi Barbier had failed on four occasions to secure FDA approval for Remoxy, a type of gel based opioid therapy designed to ensure users could not abuse it as they might a pill formulation.
With Simufilam, Barbier may have set himself just as difficult a task, given how hard it has proven to win approval for new therapies in Alzheimer's.
In my opinion Cassava has shown admirable qualities in sticking to its task of proving that Simufilam may work in the face of sky-rocketing share prices, massive collapses in valuation, intense speculation surrounding its every data set, Citizen's Petitions, criticism from the scientific community, and accusations of criminal behaviour.
Only Cassava management truly knows how close it has sailed to the wind at the present time, and whether any of the accusations leveled at it around data manipulation have any merit. The company has vociferously denied that it has broken any rules.
The next catalyst arrives when Cassava provides data from a Cognition Maintenance Study ("CMS") of ~100 patients who continue vs discontinue Simufilam, following 1-year of open-label treatment. This catalyst is scheduled to arrive in Q323, and given Cassava's operational efficiency, I am fairly confident it will.
What questions will that data answer? It may depend which 100 patients were chosen, and what level of dementia / Alzheimer's they are suffering from. Could Cassava opt to continue patients with mild dementia on Simufilam, and put patients with less mild forms of the disease in the placebo arm? If they did, we may find ourselves in another Open Label style situation where the initial data looks miraculous - until we understand what the patient breakdown was. Hopefully, Cassava will divided the placebo / treatment arms equally among patients so we can see an objective result.
As for whether Simufilam works - the jury is still out. The fact that Cassava may have teased the market by issuing exceptionally positive results from its Phase 2, knowing it was treating patients with the mildest declines first, may count as a black mark against the company, although it did not contravene any regulations so far as I am aware. And even so, the final results remain promising in their way.
What we know so far is firstly that Simufilam has a very clean safety profile. Some have concluded that this is because it is little more than a placebo drug. Secondly, we know that out of 206 patients evaluated in the Phase 2, 47% ultimately improved their ADAS-COG scoring after 12 months. We don't however, know if these results are statistically significant or how these same patients may have performed on placebo.
Finally, and most importantly, we know that the twin Phase 2 studies will almost certainly provide the answer as to whether Simufilam is an effective therapy for Alzheimer's patients or not.
If we were to speculate, we would likely conclude that the chances are it may not be an effective treatment for later stage Alzheimer's patients - unfortunately, the jury is out on Leqembi and Donanemab in that matter also.
We might conclude that it could help patients with very early stage Alzheimer's, or simply with the early stages of dementia, but as other SA contributors have pointed out, it may be very tough to secure an FDA approval on those grounds, or secure reimbursement even if approved.
At this stage, I fear failure for Cassava Sciences marginally more than I anticipate success. The company has ploughed a lonely furrow attempting to show its unique and much criticized thesis is valid, and let us hope they are ultimately proven correct. The more transparent the company can be when the CMS results arrive in Q323, the better it ought to be for everyone.
For further details see:
Cassava Sciences: Now We Wait - But In Hope, Or Expectation?