In part 2 of the ODAC postmortem (see Part 1 on Aveo Oncology [AVEO] here), I will share my thoughts on the outcome of the afternoon session which reviewed Delcath’s (DCTH) Melblez Kit for the treatment of unresectable, hepatic dominant metastatic ocular melanoma
Unlike the tone of the morning’s discussion, I really felt that both FDA and ODAC seemed more open and collaborative, willing to work with the sponsor on getting this NDA through the review process. The panel seemed sympathetic to the fact that there are no approved therapies for this orphan population. When evaluating the clinical efficacy, they agreed that there was benefit in hepatic progression free survival; however they were concerned that this did not translate to a similar magnitude of improvement in progression free survival or overall survival. Unfortunately, there were too many insurmountable safety issues with the Melblez Kit to warrant a positive committee vote.
The major safety concerns can be summarized as:
1. Toxic death rate of 7% across pooled safety database coupled with a trend towards detrimental overall survival
I want to remind readers that in a PropThink article written prior to the release of the FDA briefing materials, I had referred to slide 46 of the current investor presentation citing an equal number of deaths at 30 days in the phase 3 study for patients treated on the Melblez arm as those receiving best available care (BAC): 3/44 vs. 3/49 respectively. Sadly, when FDA pooled safety data across 3 available studies looking at all patients who had received treatment with Melblez, substantial and severe toxicity was noted. Toxic deaths attributed to Melblez included hepatic failure, gastrointestinal hemorrhage, sepsis and myelosuppression. Other severe adverse events included severe hypotension during the treatment procedure, prolonged marrow suppression (occurring in >80% of patients), gastrointestinal perforations and ulcerations (due to retrograde reflux of melphalan via the gastrointestinal branches of the hepatic artery) and severe elevations in transaminases/electrolyte disturbances. In the pivotal study, 4 patients randomized to Melblez died to treatment related adverse events with no deaths due to adverse reactions occurring in the BAC arm. The ODAC panelists commented that in evaluating the submission, “first do no harm” and that the significance of these toxicities could not be overlooked. While there was no statistically significant difference in overall survival between the two pivotal study arms, the hazard ratio was 1.35- again leading to FDA questioning the detrimental trend for survival. Similar to the discussions in the morning session, the ODAC chair Sekeres brought up the concept of how he would counsel patients on the risk of the procedure given the high death rate and potential 35% trend towards worse overall survival.
2. Significant cardiovascular adverse events associated with the procedure
Percutaneous hepatic perfusion is a highly invasive procedure which is associated with episodes of severe hypotension leading to significant drops in mean arterial pressure. Given the duration of the procedure of ~3 hours, a sustained drop in MAP lead to organ under-perfusion and ischemia. The clinical consequences included cerebral infarctions, silent myocardial infarctions, troponin elevations, and acute renal failure in some patients. The panelists were very concerned about the impact on patient quality of life if the procedure resulted in a stroke. Unfortunately the sponsor failed to collect adequate patient reported quality of life outcomes during the study and was not able to present any robust counter argument in favor of therapy.
3. Systemic adverse events despite organ isolation
The Melblez procedure aims to deliver high concentrations of melphalan to the liver on a local basis. However, patients randomized to the procedure exhibited prolonged and severe levels of myelosuppression, notably neutropenia, similar to the type of myelosuppression that would be observed if they had been treated with very high systemic doses of the drug. The panel were concerned that local drug delivery should limit systemic side effects, yet these were occurring to fatal degrees. The sponsor shared pharmacokinetic sampling data taken before and after the filtering procedure demonstrating that plasma levels after the filter were lower, however they failed to provide any explanation for the basis of the severe myelosuppression leaving significant residual safety concerns.
4. Leakage of melphalan into the G-I tract
Two deaths had occurred in the study from gastric perforation and gastrointestinal hemorrhage. Both were thought to occur due to melphalan reflux via the tributaries of the hepatic artery. While the sponsor described in detail approaches related to pre-procedure angiography and vessel embolization to prevent such reflex, the committee’s concerns were not sufficiently addressed.
Other notes:
While Richard Pazdur was not as vocal during the afternoon session as he was in the AVEO review, he did point out that the committee should carefully consider the merits of “radiographic end points” when evaluating efficacy. Another committee member had pointed out that these were subject to open-label investigator bias, especially in the BAC arm. He voiced concerns that investigators may be rushing to get patients crossed over to the experimental treatment arm by liberally detecting radiographic progression too early. All in all, the safety concerns outweighed any clinical benefits, and the lack of collection of patient reported outcomes hurt the sponsor significantly during the deliberations as they were unable to credibly argue what the clinical benefits actually were.
Where does Delcath go from here?
Similar to the AVEO scenario, a complete response letter is inevitable by the PDUFA goal date of September 13, 2013 and further clinical studies will be required if the sponsor wishes to pursue US approval. Based on the very limited patient pool in metastatic ocular melanoma, they may wish to continue with the expanded access program for those patients and instead pursue approval in an alternate patient population.
