Companies ignoring the advice of their independent data monitoring committees (IDMCs) do so at their own peril, but Argos Therapeutics Inc. might have a more compelling case than most for opting to continue the phase III ADAPT study testing rocapuldencel-T in combination with Sutent (sunitinib, Pfizer Inc.) in metastatic renal cell carcinoma (mRCC), despite a disappointing interim analysis in February, in which the IDMC recommended the study be stopped for futility.

As principal investigator Robert Figlin succinctly put it, "I personally would like to see more data – longer-term data – before deciding whether this therapy did or didn't do what we'd hoped."

Rocapuldencel-T (formerly AGS-003) is an individualized immunotherapy based on the company's Arcelis technology. It is designed to work by capturing mutated and variant antigens specific to each patient's tumor and inducing an immune response targeting the patient's tumor antigens. The ADAPT study, which is being conducted under a special protocol assessment, is evaluating rocapuldencel-T plus multitargeted tyrosine kinase inhibitor Sutent vs. Sutent alone in patients newly diagnosed with mRCC. The trial opened in January 2013 and completed enrollment in mid-2015, randomizing 462 patients 2-to-1 to the combo treatment or Sutent alone.

The first two interim analyses went off without a hitch. But during the third analysis, reported in February, the IDMC concluded that rocapuldencel-T was unlikely to show a statistically significant improvement in overall survival (OS) in the intent-to-treat (ITT) population, the study's endpoint. Those results included 75 percent of the targeted 290 events needed for the primary analysis and OS assessment, and the median OS estimated using the Kaplan-Meier method was 27.7 months (95 percent confidence interval [CI]; 23, 35.9) for the combination arm compared to the 32.4 months (95 percent CI; 0.83, 1.46) for the control arm. That was greater than the pre-defined futility boundary for the final interim analysis of 0.98. (See BioWorld Today, Feb. 23, 2017.)

The IDMC charter, established at study launch, was "based on our assumptions of that time regarding survival expectations and using the standard methodology," Argos President and CEO Jeff Abbey explained during a conference call late Tuesday afternoon. At that time, he said, the company assumed a 16-month survival time in the control arm, 22 months for the combo arm, a hazard ratio of 0.708 and a power of 80 percent, which resulted in 290 events needed for primary analysis of OS.

When the IDMC met in February, 75 percent of that number, or 217 events, had occurred, Abbey said, pointing out that was less than half the number of total patients enrolled. In the ITT population, 307 were randomized to the rocapuldencel-T/Sutent arm; of those, 39 never received any vaccinations with rocapuldencel-T.

The trial design called for patients to receive six weeks of Sutent, with those in the combo arm then receiving rocapuldencel-T at weeks six, nine, 12, 15, 18 and 24. After the 24th week, vaccinations would continue on a quarterly basis.

At the time of analysis, Abbey said, about 67 patients in the combination arm were still receiving rocapuldencel-T treatment. The median follow-up was 20 months (ranging from 0.4 months to 47.7 months) and the median number of rocapuldencel-T doses was eight.

Also at interim analysis, more than half of patients from each arm – 51 percent in the combo arm and 55 percent in the control arm – were censored for survival. Roughly 98 percent of those subjects were alive as of the Feb. 3 cut-off date for interim analysis.

Argos is basing its rationale to continue ADAPT on those censored patients and the limited median follow-up, combined with rocapuldencel-T's overall safety data and its mechanism of action.

"Of course, we believe, as most do, with an active immunotherapy there is the potential for a delayed treatment effect, otherwise known as the tail of the curve effect, where the separation between the two curves, between the control arm and experimental arm, may not occur for an extended period of time," Abbey said. "Thus, an interim analysis may not capture the potential benefit of the active immunotherapy in the experimental arm due to an insufficient period of time for long-term follow-up."

He added that "may be particularly true in the case of our therapy," which is designed to induce long-term memory immune responses. In looking at the data, Argos noted a lack of statistical correlation between immune response and survival until measurements taken after the seventh dose, or roughly week 48.

A post-hoc subgroup analysis looking at the first third of patients enrolled in the study – those for whom the longest follow-up data were available – showed an estimated median OS of 30.1 months for the combination group vs. 22.2 months for the control arm, with a hazard ratio of 0.88.

MEETING WITH THE FDA

The ADAPT trial also looked at a pre-specified exploratory endpoint, defined as an increase in CD8-positive/CD28-positive/CD45RA-negative memory T cells of at least two standard deviations from the subject-specific baselines to specific time points after dosing with rocapuldencel-T. A post-hoc analysis of 83 patients assessed for immune responses after seven doses of rocapuldencel-T showed an increase in memory T cells, which was associated with improved survival.

Data also showed that memory T cells increased only after administration of rocapuldencel-T, with statistical significance. "What this tells us is that it's unlikely that the appearance of the T cells, which correlate with clinical benefit, is coming from some source other than rocapuldencel-T," said Charles Nicolette, chief scientific officer and vice president of R&D.

It's also a sign that multiple vaccinations are needed to increase memory T cells. "For those of us in kidney cancer research," said investigator Figlin, "what we're really looking for is durable remissions," with low toxicity. "My own bias is that we're not really going to know the contribution of the vaccine . . . until we follow these patients for a longer period of time. That's why I've encouraged the company not to abandon" the ADAPT study.

Not everyone was sold, however. Analyst Mike King, of JMP Securities, appeared to have a difficult time believing the IDMC didn't take into account some of those other findings before making its recommendation to terminate the study.

"How many studies have been terminated by an IDMC and then somehow proved to demonstrate a benefit once the event rate passed 75 percent?" he asked on the conference call. "I'm not aware of any, but maybe I'm wrong."

Abbey responded by noting the difference between demonstrating benefit according to the predefined statistical analysis plan vs. demonstrating benefit by showing survival according to the tail of the curve effect. "I think that's a different question," he said.

Argos plans to sit down with the FDA next month and will talk about that tail of the curve effect, as well as the immune response demonstrated in patients that correlated with survival benefit, Abbey said.

While admitting he was going back a few years, trial investigator Figlin pointed out that Proleukin (aldesleukin), a recombinant IL-2 product from Chiron Corp. (later acquired by Novartis AG) – a product some consider to be the first immunotherapy – gained FDA approval, without progression-free survival (PFS) or OS data.

Data from the 1992 approval, however, showed complete responses and durable responses. Those results showed a "tail of the curve effect, so the issue we're at least hypothetically raising is [can] a vaccine not having an effect on an ITT population still" show benefit in a subpopulation of patients who have memory T-cell responses. "This might not have been the proper [trial] design to deal with this question," he acknowledged, but "we have to work with what we have."

That said, Argos is being realistic about what the FDA might require going forward.

Abbey said it's unlikely the agency will accept the ADAPT trial alone as the basis for approval, assuming the final data are positive. But the FDA might accept the study results "as part of a package that may include a follow-up study yet to be designed," he said.

A follow-up study likely would include new treatments that have emerged in mRCC in the years since Argos started ADAPT, notably the PD-1 inhibitor Opdivo (nivolumab, Bristol-Myers Squibb Co.), approved for mRCC in 2015. That possibility prompted another analyst to ask why Argos didn't simply shut down ADAPT and move into a new study right away.

In response, Abbey stressed the need for a "full view of efficacy. I think we want to see that and the FDA is going to want to see that."

Secondary endpoints in ADAPT are evaluating PFS, objective response rate and disease control rate. Argos also is looking at biomarker data in the ADAPT study, to determine which patients did well and whether there is a way to identify those patients at baseline for the next trial.

"I think there's still a lot to learn and we think it makes sense to continue that and learn from it," he added.

Shares of Argos (NASDAQ:ARGS), which had fallen more than 90 percent since the IDMC's recommendation in February, regained some ground Wednesday, closing at 60 cents, up 21 cents, or 53.4 percent. Shares changed hands at nearly 13 times the normal volume.