FDA’s ODAC Votes Yes to Requiring Randomized Data for Future PI3Ki Approvals in Hematologic Cancers

The FDA’s Oncologic Drugs Advisory Committee has voted 16 to 0 that future approvals of PI3 kinase (PI3K) inhibitors be supported by randomized data given observed toxicities in the drug class, research showing a detriment in overall survival (OS), a narrow range between effective, and toxic doses. One voter abstained from the vote.¹

During the April 21 ODAC meeting, committee members were tasked with discussing the toxicities of PI3K inhibitors and whether randomized studies are needed to describe the benefit-risk ratio of this class of drugs in patients with hematologic malignancies.

Understanding the PI3K Drug Class

Nicole Gormley, MD, director, Division of Hematologic Malignancies II (DHM II), Office of Oncologic Diseases, Office of New Drugs, Center for Drug Evaluation and Research, FDA, who gave the opening presentation at the meeting, explained the evidence needed to support 4 PI3K inhibitor approvals.

“It is important to note that drugs granted accelerated approval or regular approval must meet the same statutory requirements for safety and effectiveness. For safety, there must be sufficient information to determine that the drug is safe for you, under the conditions prescribed, recommended, or suggested and labeling or effectiveness. There must be substantial evidence of effectiveness that allows for the conclusion that the drug will have the effect it purports or is represented to have under the conditions of use prescribed in labeling,” Gormely explained.

Although in hematologic malignancies, PI3K inhibitors offer durable overall response rates (ORRs) as well as improvement in progression-free survival (PFS). Gormely highlighted that the OS data in trial of patients with chronic lymphocytic leukemia (CLL) or indolent non-Hodgkin lymphoma have demonstrated worse OS in the PI3K inhibitor arm in comparison with the control arm.^1,2

“While the trials show a favorable impact on efficacy end points, just as progression-free survival or overall response rate, there have been higher rates of death, and the overall survival results are concerning. It's also important to consider the patient population, those with CLL and indolent non-Hodgkin lymphoma. These diseases have a long natural history and progression isn't necessarily an indication for treatment. While these are serious and life-threatening diseases, and there's a need for continued development of products to treat relapsed or refractory disease, there are multiple therapeutics with established efficacy and safety,” Gormely said.

Study design for studies of PI3K inhibitors is also of concern within the FDA. Specifically, most clinical trials supporting FDA approvals in the drug class were single-arm studies.

“In a single-arm trial, the safety findings are challenging to interpret. Without a comparator, it can be challenging to attribute adverse events observed to the drug or to the underlying disease,” explained Gormley. “The efficacy can also be challenging to interpret. The responses observed may not translate into true clinical benefit comparisons to historical populations or cross trial comparisons are fraught with limitations.”

Toxicities observed with PI3K inhibitors are an ongoing concern, and these toxicities differ based on the mechanism of action of the PI3K inhibitor.

“The PI3 kinase inhibitors’ distinct mechanisms of action result in a differentiated safety profile depending on the isoform targeted. The delta and gamma isoforms are preferentially expressed on leukocytes, resulting in infections and immune-mediated toxicities. Infections may occur in part because of treatment-related cytopenias, but also because of the modulation of the immune system by the PI3 kinase inhibitors,” said Gormley.

Finally, dosing may explain the varying results observed with PI3K inhibitors studies.

“While the PI3 kinase inhibitors have a unique toxicity profile, and several trials have demonstrated concerning overall survival results, some of these findings may be related to poor dose optimization. The optimal dose that maximizes efficacy and minimizes safety may not have been identified. Across their class, there have been limited dose exploration,” Gormley noted.

For the reasons outlined, PI3K inhibitors have had both successes and failures in the development stage, leaving the market with just 4 FDA-approved inhibitors.

Successes and Failures With PI3K Inhibitors

Idealisib (Zydelig) was granted accelerated approval for the treatment of relapsed CLL, follicular lymphoma (FL), and small lymphocytic lymphoma (SLL) in 2014.³ The agent was evaluated in a single-arm phase 2 study (Study 101-09; NCT01282424). In the study, idelalisib demonstrated a response rate of 57% with a 12.5-month median duration of response (DOR).⁴

The safety profile of the agents was also considered acceptable for patients with indolent non-Hodgkin's lymphoma who were heavily pretreated, but the box warning for the agent notes fatal and serious toxicities: hepatic, severe diarrhea, colitis, pneumonitis, and intestinal perforation.³

In 2017, the FDA granted accelerate approval to copanlisib (Aliqopa), for the treatment of adult patients with relapsed follicular lymphoma who have received at least 2 prior systemic therapies.⁵ The approval was supported by efficacy results from an open-label, single-arm, multicenter, phase 2 trial (CHRONOS-1; NCT01660451). In the study, copanlisib achieved an ORR of 58.7% (95% CI, 48.6%-68.2%) with a 12.2-month median DOR.¹

The agent also showed a manageable safety profile with a low incidence of severe toxicity. The FDA did not list a black box warning for the drugs.

A year later, duvelisib (Copiktra) was granted regular approval by the FDA for the treatment of adult patients with relapsed or refractory CLL or SLL after at least 2 prior therapies. Data from a randomized, multicenter, open-label trial (NCT02004522) supported the approval. Duvelisib extended PFS, compared with ofatumumab, with a median PFS of 16.4 months vs 9.1 months, respectively (HR, 0.40; standard error, 0.2). Further, the ORR was higher with duvelisib at 78% vs 39%.

Duvelisib was approved with box warnings for fatal and/or serious infections, diarrhea or colitis, cutaneous reactions, and pneumonitis and warnings for neutropenia and hepatotoxicity.

Umbralisib (Ukoniq) was granted accelerated approval by the FDA in 2021, for 2 indications. The first indication was for adult patients with relapsed or refractory marginal zone lymphoma (MZL) who have received at least 1 prior anti-CD20-based regime, and the second was for adult patients with relapsed or refractory FL who have received at least 3 prior lines of systemic therapy. Findings from 2 single-arm cohorts of an open-label, multicenter, multi-cohort trial, UTX-TGR-205 (NCT02793583), supported the FDA approval.

In the MZL cohort, the ORR observed was 49% (95% CI, 37.0%-61.6%) with the median DOR not reached (NR) (95% CI, 9.3 to NR). In the FL arm, the ORR was 3% (95% CI, 33.6%-52.2%) with a median DOR of 11.1 months (95% CI, 8.3-16.4).

The prescribing information for umbralisib warns about adverse reactions including infections, neutropenia, diarrhea and non-infectious colitis, hepatotoxicity, and severe cutaneous reactions.

As a class, multiple randomized trials have revealed that evaluating a PI3K inhibitor as monotherapy or in combination in patients with CLL or non-Hodgkin lymphoma demonstrates concerning overall survival (OS) when compared with results from the control arm, according to Nicholas Richardson, DO, MPH, clinical team leader DHM II, Office of Oncologic Diseases, Office of New Drugs, Center for Drug Evaluation and Research, FDA. While OS information from trials comes early and typically represent low numbers of events, a pattern observed across multiple randomized trials has repeatedly shown a favorable impact of efficacy end points, including PFS or ORR, indicating that the OS concerns are a primary safety concern.¹

Within randomized trials, PI3K inhibitors have shown higher rates of grade 3 or higher, fatal, and serious adverse events (AEs). Specifically, grade 3 or higher AEs were reported in 71% of patients given idelalisib, 85% with copanlisib, 84% with duvelisib, and 51% with umbralisib. Differences in safety are driven by PI3K-associated toxicities.

Examining toxicity concerns with PI3K inhibitors, Richardson stated: "Given the toxicity concerns with the PI3K inhibitor class, optimized dosing is warranted. The PI3K inhibitors exhibit a narrow range between an effective and toxic dose. Across the class, there has been limited dose exploration.”

"For each approved PI3K inhibitor, there are exposure response relationships for safety, primarily for PI3K-associated toxicities. Conversely, exposure-response relationships for efficacy have generally not been observed. Despite the need to balance efficacy along with safety and tolerability. There has been insufficient dose exploration as monotherapy and in combination for these agents, Richardson added.

Although the PI3K inhibitor approvals for patients with indolent non-Hodgkin lymphomas were based on findings from single-arm trials, there is requirement for confirmatory trials to verify clinical benefit which led to the FL and CLL indications for idelalisib to be withdrawn as well as the FL indication for duvelisib. This was due to feasibility concerns for the confirmatory trials in addition to a changing treatment landscape.

In April 2022, the FL and MCL indications for umbralisib were voluntarily withdrawn based on concerns from a randomized trial. The reasons for the withdrawal highlight the limitations of single-arm trials for development and registration for PI3K inhibitors.

Randomized trials remain the preferred treatment approach as they demonstrate causal effects of a treatment, according to Richardson. Within these trials, one can compare groups of patients with respect to both known and unknown factors, which is impossible in other trial designs. Richardson concluded that there is a better chance of the trials being unbiased across randomized groups and investigators are able to assess time-to-event end points.

Voting on the Future Development of PI3K Inhibitors

Based on the information provided by the FDA and their expert knowledge of the PI3K drug class, the voting members of the ODAC collectively discussed the question raised. Considering that there are multiple ways to alter trial design in ways to address the FDA’s concern regarding PI3K inhibitors, the voting members see a need for better end points in addition to bigger trials.

"As a lymphoma hematologist and a researcher focusing on understanding toxicity and tolerability, my perspective here is that no one is arguing that there's not a clear efficacy signal or that development of this drug class should be halted, and they shouldn't be available to our patients,” said Gita Thanrajasinem, MD, a hematologist/oncologist, and internist at the Mayo Clinic in Rochester, Minnesota. “The question is whether randomized data are warranted as a regulatory strategy here. And so, as a clinician, these aren't drugs that I'm reaching for initially, but they're ones that I would like to have available as options for my patients in later lines, usually after exhausting other available options. But for patients in later lines of therapy whose life expectancy is the most limited by their disease, the benefit-risk assessment is still very crucial and it's still ‘first do no harm’ for this precious population of patients.”

Some members consider the voting questions difficult to reconcile, and in response, Gromley stated; “If the cure for cancer is developed tomorrow, I think we could find ways to review this expeditiously. None of us has a crystal ball, but I think that it's really important that we don't make the same mistakes from yesterday and that we learn from our experience,” she said, hinting toward previous withdrawals in the class and failed trials after accelerated approvals.

In the end, the temporary voters Kieron M. Dunleavy, MD, Walter K. Kraft, MD, MS, FACP, Michele Nadeem-Baker, MS, Gita Thanrajasinem, MD, Jessie Lai-Sim Au, PharD, PhD, Andy J. Chen, MD, PhD, Christopher S. Coffey, PhD, MS, and Louis F. Deihl, MD voted in favor of requiring randomized data. One temporary voter, Anthony Sun, MD abstained. Further, all ODAC members voted yes to requiring randomized data.

_REFERENCES:

_{1. Oncologic Drugs Advisory Committee (ODAC) Meeting. April 21, 2022. FDA website. Accessed April 21, 2022.}

_{2. U.S. Food and Drug Administration approves Gilead’s Zydelig® (idelalisib) for relapsed chronic lymphocytic leukemia, follicular lymphoma and small lymphocytic lymphoma. News release. Gilead Sciences, Inc. July 23, 2014. Accessed April 21, 2022. https://bit.ly/3L53sVs} 

_{3. Gopal AK, Kahl BS, de Vos S, et al. PI3Kδ Inhibition by Idelalisib in Patients with Relapsed Indolent Lymphoma. N Engl J Med. 2014; 370(11): 1008–1018. doi: 10.1056/NEJMoa1314583}

_{4. FDA grants accelerated approval to copanlisib for relapsed follicular lymphoma. News release. FDA website. September 14, 2017. Accessed April 21, 2022. https://bit.ly/3rHrspY}

_{5 . Dreyling M, Santoro A, Mollica L, et al. Abstract CT149: Copanlisib in patients with relapsed or refractory indolent B-cell lymphoma: Primary results of the pivotal Chronos-1 study. Cancer Res. 2017;77 (Suppl 13): CT149. doi: 10.1158/1538-7445.AM2017-CT149}