The FDA’s Oncologic Drugs Advisory Committee (ODAC) made a clean sweep of its 2-day meeting, deciding that the PI3K inhibitor duvelisib (Copiktra) does not offer a favorable risk-benefit ratio for patients with chronic lymphocytic leukemia (CLL) or small lymphocytic lymphoma (SLL).
By an 8-4 vote, the panel came down on the side of the FDA staff perspective that duvelisib had a potentially detrimental effect on survival in a randomized trial that supported Secura Bio’s approval application. Analysis of 5-year overall survival (OS) from the randomized DUO trial showed a hazard ratio of 1.09 versus the ofatumumab (Arzerra) control arm in an intention-to-treat analysis and 1.06 in the indicated population (relapsed/refractory disease and two or more prior lines of therapy).
Adjusted analyses yielded similar results. Although not statistically significant, the suggestion of a survival detriment, along with evidence of high rates of toxicity, swayed the panel’s overall opinion.
The duvelisib outcome followed negative votes for the risk-benefit profile of the lung cancer drug candidate poziotinib and the myeloma treatment melphalan flufenamide (Pepaxto) during the first day of the meeting.
‘Playing With Fire’
The outcome continued a string of bad news for the PI3K inhibitor drug class. At a meeting in April, the ODAC panel reviewed data showing a consistent trend toward an adverse impact on OS across the entire class, which an FDA report characterized as “unprecedented in the field of oncology.”
“I think with this drug and with this class of drugs, we are playing with fire,” said Mikkael Sekeres, MD, of the University of Miami’s Sylvester Comprehensive Cancer Center, following the duvelisib vote. “This drug had modest activity with significant toxicity, as did other members of this class, and was compared to a drug that we would no longer use in this setting. This drug [duvelisib] itself we would no longer use in this setting, as patients received other drugs, such as BTK [Bruton’s tyrosine kinase] inhibitors and BCL-2 inhibitors, for which they would have been disqualified from the study.”
“So, we’re left with a drug that has substantial toxicities and a questionable indication today,” he added.
Other ODAC members struggled with the data and with their decision, even when they voted no.
“If you have an indolent disease, and overall survival obviously is very difficult to figure out in that setting, then you become more compelled by the overall response rate and progression-free survival,” said Christopher Lieu, MD, of the University of Colorado Cancer Center in Aurora. “The progression-free survival benefit here is compelling … and we want more drugs in this setting.”
“If you have a disease where survival is measured in weeks to months, the bar you set for toxicity and what you’re expecting out of the therapy is pretty low,” he added. “But the flip side is also true. If you have an indolent disease, what is the cost to our patients that we’re going to expect out of a therapy in a setting where we’re not sure that it improves overall survival?”
Same Data, Different Perspectives
Similar to Thursday’s session on poziotinib, the FDA staff and the sponsor, Secura Bio, had sharply differing views on duvelisib’s risks and benefits. Noting that an updated survival analysis from the DUO trial showed a trend favoring the control arm, the FDA staff report concluded that the “current benefit-risk of duvelisib in patients with relapsed or refractory CLL or SLL is not favorable.”
Looking at the same data, Secura Bio representatives said the final survival analysis “does not support the conclusion of a detriment in OS in patients treated with duvelisib and did not identify any new safety concerns.” They pointed to a huge imbalance in crossover treatment: 90 from the ofatumumab control arm to duvelisib versus nine from duvelisib to ofatumumab.
The FDA staff countered with adjusted analyses that accounted for the crossover disparity and still showed a trend toward excess mortality in the duvelisib arm. They also noted that several patients died of treatment-related adverse events after switching from ofatumumab to duvelisib.
A number of other factors confounded data assessment. Only 14% of the study population came from the U.S., and Black patients accounted for only 1% of the total, raising questions about the applicability of the findings. Ofatumumab is not widely available nor routinely used in the management of CLL/SLL.
The randomized trial excluded patients previously treated with BTK inhibitors, and BCL-2 inhibitors were not yet available. Those two classes currently represent first- and second-line treatment for CLL/SLL. Whether duvelisib, or any PI3K inhibitor, offers a benefit after progression on those two drug classes is unknown.
Several ODAC panelists expressed confusion and frustration with the data on survival and toxicity. Panelist David Harrington, PhD, a biostatistician at Dana-Farber Cancer Institute in Boston, offered a bottom-line assessment of the data.
“It is incumbent upon the sponsor to establish that there was a favorable risk-benefit profile,” he said. “Given the current context — the data about this class, extended follow-up on this study — I don’t think they’ve done that.”
Approved 4 years ago, duvelisib was hailed as “an important addition to the evolving treatment paradigm for patients with CLL/SLL and follicular lymphoma (FL).” In the phase III, multicenter, randomized DUO trial of CLL/SLL, patients who received the PI3K inhibitor had a 7-month improvement in median progression-free survival (PFS) versus ofatumumab and a two-fold improvement in objective response rate.
Duvelisib had more toxicity versus ofatumumab, but appeared to be less toxic than idelalisib (Zydelig), the first approved drug in the PI3K inhibitor class. OS data, a secondary endpoint in DUO, were immature at the time of the approval.
The FL approval was based on the phase II single-arm DYNAMO trial, which showed a 43% response rate in rituximab (Rituxan)-refractory patients. The accelerated approval came with the stipulation that the sponsor (at the time, Verastem) would conduct a confirmatory trial.
Late last year, Secura Bio voluntarily withdrew duvelisib’s FL indication, citing the “current treatment landscape for FL patients in the U.S. and the logistics, cost and timing of the post-marketing requirements.”
At the April meeting, ODAC put the entire PI3K inhibitor class on notice about safety concerns, as data reviewed by the committee showed an excess of fatal events, as well as high rates of severe and overall toxicity in patients with indolent lymphomas, normally associated with a prolonged clinical course and survival. The panel agreed that future approvals for PI3K inhibitors should be contingent on survival data.
“Certainly, PFS, at least for most of us who voted yes, didn’t appear to be a reasonable endpoint for this class of agents,” ODAC chair Jorge Garcia, MD, of Case Western Reserve University in Cleveland, said at the time.
Then, in late June, the FDA singled out duvelisib in a warning about an increased mortality risk among patients treated with the drug in the DUO trial. FDA also announced plans for an ODAC meeting in September to discuss the updated OS data from DUO.
Prior to the duvelisib vote, Garcia said, “It has been obviously a complex discussion, and I predict that it’s not going to be an easy vote.”