Calling the hazard ratio of 0.35 "remarkable," FDA reviewers nevertheless express concern about the number of censored events and the difference in percentage of censored events between the two groups. The FDA’s analysis employed different patterns of censoring than that of AstraZeneca, with the result being that 74.5% of vandetanib patients were censored, compared with 59% in the placebo arm.
"Censoring is primarily due to the lack of correlation between investigator and IRC-read scans," the agency says. "Only 88/174 (50.6%) scans were read as progression by both the investigator and IRC."
Clinical Trial Conduct Under Review
In addition to the censoring issue, the agency hints at concerns about study conduct that could cast a shadow on the efficacy findings. Protocol violations occurred in 21.2% of vandetanib-treated patients and 27% of subjects in the placebo arm. "These included the absence of measurable disease by investigator in 14 patients, the absence of a confirmed histological diagnosis in one patient and one patient who received the incorrect treatment," the FDA says. "Inspection of the clinical sites has resulted in several observations that are under review."
The phase II trials assessed response rates (complete response plus partial response). In one study that evaluated an initial daily dose of 300 mg vandetanib, investigator and IRC-assessed response rates were 20% and 16.7%, respectively. A second study evaluated a 100 mg dose, and the investigator response rate was15.8%.
The phase II study response rates were considerably lower than those seen in the phase III study, where the investigator and IRC response rates for vandetanib were 39% and 44.6%, respectively.
Adverse Event Concerns Could Tip Risk/Benefit
Dose reductions and interruptions due to adverse events were a prominent issue in all three studies. In the phase III trial, 35.9% of vandetanib patients required a dose reduction and 47.2% required a dose interruption.
Serious adverse events occurred in 30.7% of vandetanib-treated patients in the phase III trial, with the most common events being diarrhea, pneumonia, and hypertension.
The agency expresses particular concerns about AE signals from the 3,000-patient vandetanib safety database relating to QT interval prolongation, cerebrovascular events, interstitial lung disease/pneumonia, and serious skin conditions.
"Vandetanib at a dose of 300 mg is associated with a substantial (mean effect 35 ms) and concentration-dependent prolongation in QTc," the FDA says. "This increase in mean QTc does not lessen over time and the half-life of vandetanib (19 days) makes this prolongation in QTc interval particularly problematic. In addition to QTc prolongation, the majority of the severe adverse events seen with both EGFR and VEGFR inhibitors have been reported with vandetanib. This includes Stevens-Johnson syndrome, some ischemic arterial events, and interstitial lung disease."
The safety profile is raised in the FDA’s one voting question for ODAC: "Given the substantial toxicity observed with vandetanib and the long natural history of the disease, is the risk-benefit profile acceptable for the proposed indication or should the indication be limited to patients with progressive, symptomatic medullary thyroid cancer?" The agency also seeks the committee’s views on other subgroups that may be appropriate for treatment.
On the issue of dosing, AstraZeneca proposes "dosing to tolerance," starting with an initial daily dose of 300 mg, with reductions as necessary to manage side effects and avoid discontinuation in most patients "to maintain the maximum possible derived clinical benefit and engagement with molecular targets over many months to years." However, the FDA is asking ODAC whether additional doses should be explored as a post-marketing study requirement to determine the optimal dose and potential study designs for accomplishing this.
Used with permission from "The Pink Sheet." Internal Medicine News Digital Network and "The Pink Sheet" are owned by Elsevier.