ImmunoCellular Therapeutics (NYSEMKT:IMUC) presented ICT-107 phase II results at the 2014 ASCO annual meeting and in a press release, expressed the company's intent to initiate a phase III trial to treat newly diagnosed gliobastoma (GBM) patients.
The company's CEO alluded that the phase III trial will only randomize a subpopulation of newly diagnosed GBM patients who are HLA-A2.
"The pre-specified subgroup analyses in our phase II trial indicate the potential value of a targeted population approach going forward, as the demonstration of treatment benefit in HLA-A2 patients presents a favorable case for selecting these patients for the population to be studied in phase III clinical testing," Andrew Gengos.
The phase III trial will probably be similar in design to the phase II study. Although the company has spoken to EU regulators, there has been no mention of establishing manufacturing in Europe, so it is unlikely that the phase III trial will include sites outside the US at the moment. As before, patients will be enrolled and then screened for HLA type. The main difference is that only HLA-A2 patients will be randomized at a 2:1 ratio to receive standard of care with or without ICT-107, respectively. The primary endpoint will most likely be overall survival (OS) again.
It is difficult to determine how many patients need to be enrolled for the phase III trial, since OS benefit in HLA-A2 patients is not known yet. Let's assume that ICT-107 improves median OS by 4.5 months. Recall that the phase II study was powered to show a 9 month improvement in median OS. So IMUC should randomize about 250 HLA-A2 patients in the phase III study. Of the 278 patients enrolled in the phase II study, 77 were HLA-A2 patients - roughly 28%. To randomize 250 HLA-A2 patients, about 900 patients will have to be enrolled.
In the phase II study, 278 patients were enrolled in about 1.5 years at 25 sites - an average of about 15 patients per month. To enroll 900 patients, it could take as many as 5 years, although enrollment should be quicker if more sites are added. Other factors, such as competing trials, can also impact enrollment rates.
It took about 1.5 years after enrollment was completed in the phase II study for enough deaths to occur to calculate median OS. Assuming patients in the phase III trial will survive a similar amount of time, it could take as long as 3 years after enrollment is completed before results are available, since there could be twice the number of randomized patients. For comparison, Roche's (OTCQX:RHHBY) AVAglio trial enrolled 921 newly diagnosed GBM patients at 132 sites throughout the world and took about 4 years to complete.
Keep in mind that the estimates previously mentioned are very rough. As more information becomes available, these estimates will change. For example, if less patients need to be randomized and more sites are added, the phase III study could be completed in less time. Once the end-of-phase II FDA meeting has taken place, there should be some more clarity as to what the phase III trial will actually look like.