The decision marks the first time the FDA has relied on a single externally controlled trial to support a potential approval in oncology.
In a 14 to 6 vote, the FDA’s Oncologic Drug Advisory Committee (ODAC) agrees that eflornithine (DFMO) shows sufficient evidence to reduce the risk of relapse in pediatric patients with high-risk neuroblastoma (HRBN) who are in remission and have completed multi-agent, multi-modality therapy. The decision is a potentially precedent-setting one.1
Study 3b design NMTRC003/3b (NCT02395666), referred to as study 3b, which supported the new drug applicationfor eflornithine, did not have its own control arm and instead presented external control data from study ANBL0032 (NCT00026312), which evaluated isotretinoin with or without dinutuximab, aldesleukin, and sargramostim following stem cell transplant in patients with neuroblastoma.
“The FDA has not previously relied on a single externally controlled trial [ECT] to support an approval in oncology. Given this context and in the setting of this unique clinical trial, we are seeking additional feedback from the advisory committee.” Elizabeth Duke, MD, clinical reviewer from the FDA’s Office of Oncologic Diseases, said.
“Because these results indicated the possibility to reduce the number of patients we see relapse and die, and because of the availability of [ANBL]0032 as a uniquely optimal external control, we solidified our decision to pursue registration with study 3b,” Giselle Sholler, MD, division chief of pediatric hematology, oncology, and bone marrow transplant at Penn State Children’s Hospital, said on behalf of the applicant.
DFMO is an ornithine decarboxylase (ODC). ODC plays a critical role in polyamine synthesis and a transcriptional target of MYCN, and inhibiting this process suppresses tumor growth and formation.1,2 Study 3b is a multicenter, open-label, single-agent study. Patients received oral DFMO for 2 years and were followed for event-free survival (EFS) and overall survival (OS) until death or for 5 years after DFMO treatment. The primary end point was a comparison of 2-year EFS between upfront remission patients from study 3b and the ANBL0032 rate in patients with upfront remission. Results from study 3b estimated EFS was 85% (95% CI: 0.763, 0.904) compared to the historical control rate of 70%.1
“The high unmet medical need for patients with neuroblastoma, the specific external control data source, and results of the propensity score-matched analysis impacted FDA’s willingness to consider an external control design in this circumstance,” Diana Bradford, MD, cross disciplinary team leader for the FDA’s Office of Oncologic Diseases, said in the FDA’s opening remarks.
Duke presented the appropriateness of the proposed ECT. Strengths included the natural history established by prior clinical trials, the FDA-verified data source of ANBL0032, similar follow-up protocols, and comparable definitions and ascertainment of endpoints. Duke noted the limitations as differing enrollment timelines, uncertainty in treatment effect due to external control trial design, and potential missing covariates in the respective analyses.
There were also several non-matched clinical characteristics between study 3b and ANBL0032. These included patient demographics like ethnicity and social determinants, as well as treatment-related characteristics including days from diagnosis to transplant and transplant regimen.
Arup Sinha, PhD, statistics reviewer from the FDA’s Office of Biostatistics, presented several sensitivity analyses assessing study design and data, unmeasured confounding variables, and alternative statistical approaches. He concluded that the applicant’s primary analyses were consistent with the FDA’s sensitivity analyses.
“The FDA has not previously relied on a single ECT as the primary source of evidence in oncology. However, the ECT has specific strengths due to the provenance of the external control data,” Sinha said.
Study 3b was initiated in 2012. US WorldMeds’ sponsor briefing document states that the goal was to design a phase 3 study informed by the findings from phase 2. However, preliminary outcomes from study 3b were better than investigators expected. In 2018, the FDA recommended that US WorldMeds provide patient-level data from study ANBL0032. In 2020, the FDA granted DFMO a breakthrough treatment designation based on the propensity score that matched external control data from ANBL0032.
The applicant, US WorldMeds, submitted a NDA (215500) for DFMO tablets in 2022. The applicant brought forth several experts to discuss the unmet needs of HRNB treatment, DFMO’s efficacy, and clinical perspective. Kristen Gullo, vice president of development and regulatory affairs for US WorldMeds, began the applicant presentations and provided information on DFMO’s safety.
“DFMO was generally well tolerated with few patients requiring dose modification and even fewer requiring discontinuation,” Gullo said.
DFMO’s safety was studied in a population of 311 patients with HRBN. Out of these patients, 216 (69.5%) received treatment for at least 1 year, with a median duration of exposure of 2 years. Grade 3 or 4 adverse events (AEs) were reported in 43.7% of the population and included hypoacusis (11.9%), increased alanine aminotransferase (11.9%), increased aspartate aminotransferase (6.8%), pyrexia (4.2%), anemia (3.9%), and decreased neutrophil count (3.5%). Dose modification was required in 11.9% of patients, and 5.1% of patients discontinued the drug. Serious AEs were reported in 16% of patients, and there were no deaths due to AEs.
DFMO has a risk of significant toxicities, with ototoxicity being the most established. Researchers reported that 12.9% of patients experienced hearing loss that worsened at least 1 grade from baseline, and 12.2% experienced hearing loss that worsened to grade 3 (i.e. the point where intervention such as hearing aids were required).1
Among patients whose doses were modified, or drug administration was discontinued, 63% had hearing that improved or returned to baseline. In her presentation, Gullo noted that less than 5% of patients discontinued the drug, and only 2% discontinued the drug due to hearing loss.
“This supports the FDA’s overall conclusion that risks of DFMO therapy can be managed with recommendations proposed for product labeling,” Gullo said.
Thomas Clinch, senior director of biometrics and clinical development of US WorldMeds, presented information on DFMO’s efficacy and how well study ANBL0032 served as an external control. In his presentation, Clinch highlighted the similar demographics, disease characteristics, therapeutic administration methods, follow-up, and event definitions between studies 3b and ANBL0032.
Committee members expressed concerns about voting yes in a potentially precedent-setting situation.
“It is very complicated. I don’t feel that a randomized trial would be impossible. In fact, I think it’s quite feasible,” committee member Brian Weiss, MD, said during the meeting’s discussion. “I have a lot of patients who were not on the trial who also had very similar stories to the moving testimony that people gave of their child or their own results on DFMO. And that’s why we have to do a randomized trial.”
“Neuroblastoma is weird…and we don’t completely understand that,” Weiss added. Weiss was among the 6 committee members who voted no.
“There are circumstances in which we could avoid this. There are cases where there are sufficient patients that we could do a randomized control quickly. I felt in this setting that the evidence shows DFMO is effective…The preclinical studies are also compelling,” committee member Shahab Ashgharzadeh, MD, who voted yes, said after the vote.
1. US WorldMeds. Eflornithine (DFMO) tablets to reduce the risk of relapse in pediatric patients with high-risk neuroblastoma who have completed multiagent, multimodality therapy. https://www.fda.gov/media/172661/download
2. US Food and Drug Administration. FDA Briefing Document: NDA 215500. https://www.fda.gov/media/172659/download