'Data mining' and 'p-value hacking': Cel-Sci CEO tries to spin cancer trial failure into a win

Cel-Sci's stock price plummeted 45%, losing more than $11 per share and dropping to $13.69 after the news was issued June 28. (Andrii Yalanskyi/Getty Images)

Cel-Sci CEO Geert Kersten begged investors in 2018 to hold on until phase 3 data could prove the struggling biotech’s cancer drug worked. Fast forward to 2021: The results are in, and they are unlikely to inspire the rally needed to save the company.

The biotech's stock price plummeted 45%, losing more than $11 per share and dropping to $13.69 after the news was issued June 28. The tumble has continued, with the stock sitting at about $8.38 apiece Wednesday.

Kersten has turned to his keyboard once again in an effort to spin the results for Multikine into a win.

“Some false assertions and misrepresentations have been made and published by parties who either did not understand the protocol and statistical analysis or had ulterior motives pertaining to our stock price,” Kersten wrote in a letter to stockholders published Wednesday.

RELATED: Cel-Sci's 11-year cancer drug odyssey ends in failure, sinking stock, but plots FDA filing anyway

Here’s what the results said: Multikine failed to improve overall survival when added to standard of care treatment in the overall population tested in the phase 3 head and neck cancer study. Cel-Sci wanted to see at least a 10-percentage point improvement on this metric when Multikine was added to standard of care, comprising surgery, radiotherapy and chemotherapy. This was the main goal of the trial, meaning the test was unsuccessful.

But Cel-Sci plans to file the drug for FDA approval anyway, focusing on a narrow subset of patients who did not receive the chemotherapy part of their treatment. The study showed a 14-percentage point improvement in overall survival in this group.

This is where Kersten believes the “confusion” arises from. By confusion, he means an investor revolt that has turned into at least one class action lawsuit from shareholders alleging the company’s officers and/or directors engaged in securities fraud or unlawful business practices.

According to Kersten, the data spurred accusations that the company had engaged in “data mining” and “p-value hacking.” Cel-Sci was accused of sitting on the data for a year before landing on the subgroup analysis as a way to move the therapy forward anyway. Questions arose as to whether Cel-Sci could even use the subgroup data for a filing because they were not prespecified in the study’s initial endpoints.

Kersten cited an unnamed statistician to support his case that no mining or hacking had occurred. He claimed the subgroup analysis was prespecified, meaning the company can use these data to file for FDA approval. He also said the p-value data were “in fact very strong.”

RELATED: Cel-Sci CEO paints 'promising future' in letter to shareholders

The p-value measures the probability that the effect observed in the trial is a result of the experimental treatment and not merely chance. The lower the number, the better it is for the study. Cel-Sci reported that the overall survival results for Multikine plus standard of care without chemotherapy showed a p-value of 0.0236.

Kersten complained that short sellers had targeted the company before the results were issued and therefore had an interest in seeing the shares plummet.

“But now, the data clearly shows that Multikine extends life in 40% of newly diagnosed advanced primary head and neck cancer patients,” Kersten wrote. He said that if Multikine is cleared by the FDA, tens of thousands of patients with head and neck cancer may live longer. “What kind of person continues to sell short and attack a company that can make this kind of difference in the lives of cancer patients?”

Back in 2018, Kersten’s letter pointed to this data release as a “key inflection point” that could right the ship after several rocky years. He promised the biotech had a "promising future" ahead. 

In reality, Cel-Sci had just made a pitch to the New York Stock Exchange to try to hang on to its listing, stockholders had an equity deficit and the company reported net losses on the five prior fiscal years. The company was also emerging victorious from a lawsuit with a contract research organization that was involved in a phase 3 study for head and neck cancer that ran from 2011 to 2013. The legal action had shaken the confidence of investment funds and analysts, but Kersten said at the time that the win should put the hesitation to bed.

RELATED: Cel-Sci wins 4.5-year-long arbitration case against CRO for breach of contract  

Despite all of those events, Kersten believed the company was way undervalued at the time.

In a December 2020 (PDF) 10-K filing, Cel-Sci calculated its net losses at $384 million since inception, with $30.2 million clocked for fiscal year 2020. The company has not commercialized any products and receives revenue only through the sale of securities. This is expected to continue "for the foreseeable future."

"Multikine is the only product candidate in late-stage clinical development, and Cel-Sci’s business currently depends heavily on its successful development, regulatory approval and commercialization," the filing stated.

The FDA has not cleared a new drug for this type of advanced head and neck cancer in decades. Cel-Sci will now take the data to the FDA for a meeting ahead of a regulatory filing for patients to receive Multikine followed by surgery and radiation—but not chemotherapy.

Cel-Sci clearly hopes the subgroup will push Multikine—and the company—to a new inflection point.